• Re: [OT] Dead DEC Punch card times (was Re: Taking Speedy too seriously)

    From John Ames@commodorejohn@gmail.com to comp.os.linux.misc on Thu Nov 5 09:51:48 2026
    From Newsgroup: comp.os.linux.misc

    On Tue, 3 Nov 2026 22:48:27 -0500
    c186282 <c186282@nnada.net> wrote:

    Anyway though, the Z-80 days are LONG gone.

    Not *long* gone, it was only discontinued in 2024 - certainly an all-
    timer for single-chip longevity. Still gets a lot of play in the retro/ homebrew scene, too.

    Kind of a pity ! Good stuff for minimal CPU.

    I'm partial to the 6502 myself, but it certainly wasn't bad. As with
    most 8-bitters, it wasn't very orthogonal, but it had a surprisingly
    roomy register file; the ability to quick-switch to an alternate set
    for fast interrupt response was particularly nice.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From rbowman@bowman@montana.com to comp.os.linux.misc on Thu Nov 5 19:11:57 2026
    From Newsgroup: comp.os.linux.misc

    On Thu, 5 Nov 2026 09:51:48 -0800, John Ames wrote:

    I'm partial to the 6502 myself, but it certainly wasn't bad. As with
    most 8-bitters, it wasn't very orthogonal, but it had a surprisingly
    roomy register file; the ability to quick-switch to an alternate set for
    fast interrupt response was particularly nice.

    https://www.youtube.com/watch?v=LnzuMJLZRdU

    Somewhere along the line I acquired De Jong's 'Programming & Interfacing the 6502 with Experiments'
    although I never worked with the processor. Too many irons on the fire or I'd play along with
    Eater's videos.

    https://archive.org/stream/programming-and-interfacing-6502-with-experiments/ProgrammingAndInterfacing6502WithExperiments_djvu.txt
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Charlie Gibbs@cgibbs@kltpzyxm.invalid to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Thu Nov 5 19:18:00 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-05, Peter Flass <Peter@Iron-Spring.com> wrote:

    On 11/4/26 12:51, Charlie Gibbs wrote:

    On 2026-11-04, Peter Flass <Peter@Iron-Spring.com> wrote:

    I kind of agree. As a former assembler programmer - at one time it was
    my first language - I'd like it to be true (although I'm out of the job
    market), but I can't see it. These days most HLL code is more optimized >>> than a lot of assembler. A good assembler programmer would not make a
    lot of these optimizations due to lack of readability.

    A mediocre assembly language programmer would not let lack of
    readability stand in the way. And once you get into self-modifying
    code there are all sorts of horrible things you can do.

    1. "standard" things like converting a branch to a nop or back.

    When reading about interfacing assembly language subroutines to
    RPG programs, I noted that indicators were stored in one byte as
    0x00 for off, and 0xF0 for on. I suspected (but never confirmed)
    that this would allow code generated by the RPG compiler to stuff
    the indicator byte into the second byte of a branch instruction,
    which on a 360-style architecture would change it to either a NOP
    or a branch.

    I know 1401 Autocoder people used to do horrible things like using
    part of an instruction as a constant.

    Somewhere in my archives I have a slim pamphlet - I think it was
    part of an official IBM series on Autocoder programming - which
    contained tips and techniques for writing self-modifying code.

    2. You can do anything as long as you comment it. I once had a sequence
    to test-store-conditional branch, for some reason I don't recall, bit I tagged it with comments saying "these three instructions must be kept together".

    Sounds like you were counting on the condition code from the test
    surviving through the store instruction so that the conditional branch
    would properly pick it up. Yes, definitely worth a cautionary note.
    --
    /~\ Charlie Gibbs | In this world there are
    \ / <cgibbs@kltpzyxm.invalid> | two kinds of people:
    X I'm really at ac.dekanfrus | 1. Those who can extrapolate
    / \ if you read it the right way. | from incomplete data.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Thomas Koenig@tkoenig@netcologne.de to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Thu Nov 5 19:57:52 2026
    From Newsgroup: comp.os.linux.misc

    Charlie Gibbs <cgibbs@kltpzyxm.invalid> schrieb:
    On 2026-11-04, John Levine <johnl@taugh.com> wrote:

    According to c186282 <c186282@nnada.net>:

    I can assure you that PDP-11s were minicomputers. I spent quite a while >>>> programming them back in the 1970s.

    AT THE TIME, characterizing the PDP-11 as
    a "mini" is not wrong. I view a "mini" as
    a 'cut-down' mainframe.

    A mini was never a cut-down mainframe. Mainframes
    had channels, minis had simple I/O and fast interrupts.
    A 360/30 was considerably slower than a PDP-8, but its
    I/O architecture clearly made it a mainframe.

    I agree. I've always believed that the difference between
    a mainframe and a mini had nothing to do with the size of
    the box, but everything to do with architecture, particularly
    I/O. Mainframes were record-oriented, while minis were
    byte-oriented.

    The PDP-8 was anything but byte-oriented, it was a 12-bit machine.

    Plus, DECTape was block-oriented.
    --
    This USENET posting was made without artificial intelligence,
    artificial impertinence, artificial arrogance, artificial stupidity,
    artificial flavorings or artificial colorants.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From John Ames@commodorejohn@gmail.com to comp.os.linux.misc on Thu Nov 5 12:12:55 2026
    From Newsgroup: comp.os.linux.misc

    On 5 Nov 2026 19:11:57 GMT
    rbowman <bowman@montana.com> wrote:

    I'm partial to the 6502 myself, but it certainly wasn't bad. As with
    most 8-bitters, it wasn't very orthogonal, but it had a surprisingly
    roomy register file; the ability to quick-switch to an alternate
    set for fast interrupt response was particularly nice.

    https://www.youtube.com/watch?v=LnzuMJLZRdU

    Somewhere along the line I acquired De Jong's 'Programming &
    Interfacing the 6502 with Experiments' although I never worked with
    the processor. Too many irons on the fire or I'd play along with
    Eater's videos.

    Yeah, I've heard his name tossed around a lot in recent years. Reminds
    me, I need to get back to putting together the 65C02 board for RC2014
    that I picked up a while ago...

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From c186282@c186282@nnada.net to comp.os.linux.misc on Thu Nov 5 15:50:22 2026
    From Newsgroup: comp.os.linux.misc

    On 11/5/26 15:12, John Ames wrote:
    On 5 Nov 2026 19:11:57 GMT
    rbowman <bowman@montana.com> wrote:

    I'm partial to the 6502 myself, but it certainly wasn't bad. As with
    most 8-bitters, it wasn't very orthogonal, but it had a surprisingly
    roomy register file; the ability to quick-switch to an alternate
    set for fast interrupt response was particularly nice.

    https://www.youtube.com/watch?v=LnzuMJLZRdU

    Somewhere along the line I acquired De Jong's 'Programming &
    Interfacing the 6502 with Experiments' although I never worked with
    the processor. Too many irons on the fire or I'd play along with
    Eater's videos.

    Yeah, I've heard his name tossed around a lot in recent years. Reminds
    me, I need to get back to putting together the 65C02 board for RC2014
    that I picked up a while ago...

    Nothing wrong with the 6502 at all. Meant
    as an 'improved' version of the m6800 by
    some of the people who made the 6800. The
    big trick was that it was much CHEAPER to
    make than the 6800 series. Not sure we'd
    have had the first wave of PCs without
    the affordable chip.

    Alas I don't remember any 6502s that were
    clocked much over 1mhz - while the Z-80s
    could run at four. MIGHT be that the Z-80
    required more cycles per instruction though,
    reducing that advantage - somebody here
    knows ..... I think the Z-80 needed twice
    as many clocks.

    Anyway, 16/32 bit CPUs were already in the
    pipeline and users needed higher performance
    and more mem so they could do higher-rez graphics.
    The 8-bit chips may have started-off things but
    couldn't cope with WHAT they'd started.

    The m6809 was a nice chip too, but again
    Motorola never got the PRICE down fast
    enough so it got lost in the crowd.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From The Natural Philosopher@tnp@invalid.invalid to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Thu Nov 5 20:58:56 2026
    From Newsgroup: comp.os.linux.misc

    On 04/11/2026 14:36, Peter Flass wrote:
    On 11/3/26 13:27, The Natural Philosopher wrote:
    On 03/11/2026 15:13, Scott Lurndal wrote:
    Actually, the numbers decline every year.-a-a The vast majority are
    legacy installations (particularly for the Burroughs/Univac/Unisys
    mainframes) and there are very few, if any, new customers adopting
    mainframes of any flavor.

    "Global mainframe market size and IBM Z mainframe revenue trends
    highlight a cyclical pattern dominated by multi-year refresh waves (such
    as the z16 and z17 cycles):

    Global Mainframe Market Size:

    -a-a-a-a-a-a-a-a Valued at approximately $5.33 billion to $5.71 billion in 2025.
    -a-a-a-a-a-a-a-a Projected to grow to $5.65 billion to $6.11 billion in 2026.
    -a-a-a-a-a-a-a-a Forecasted to reach $7.54 billion to $10.53 billion by
    2031rCo2034, driven by a compound annual growth rate (CAGR) of roughly
    5.95% to 7.05%.

    IBM Z Mainframe Performance & Fluctuations:
    -a-a-a-a-a-a-a-a Late 2025 (z17 Launch / Q4): IBM reported its strongest Q4 >> mainframe revenue in 20 years, with IBM Z sales surging 67%
    year-over-year as enterprises adopted the z17 architecture.
    -a-a-a-a-a-a-a-a Mid-2026 (Q2): IBM Z mainframe revenue experienced a short-term
    42% year-over-year decline in fiscal Q2 2026 due to delayed large
    capital expenditure (capex) deals and client shifts toward
    supply-constrained servers, though management noted these were deferred
    purchases rather than lost contracts (with roughly one-third closing
    shortly after the quarter)."

    So, as usual, the facts are not correctly represented by the resident
    armchair experts


    My only question is how many mainframe shipments represent NEW
    customers, as opposed to additional horsepower for current customers.

    My only question is how many computer sales are to new users, rather
    than upgrades
    My only question is how many car sales are to new users, rather than
    upgrades
    My only question is how many smartphone sales are to new users, rather
    than upgrades

    Completely specious weasel

    I suppose, given that the trend today is to consolidation of everything, that eventually everything will be owned by BigCorp, who will then be
    the only mainframe customer.

    Oh dear.
    --
    If you tell a lie big enough and keep repeating it, people will
    eventually come to believe it. The lie can be maintained only for such
    time as the State can shield the people from the political, economic
    and/or military consequences of the lie. It thus becomes vitally
    important for the State to use all of its powers to repress dissent, for
    the truth is the mortal enemy of the lie, and thus by extension, the
    truth is the greatest enemy of the State.

    Joseph Goebbels




    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From The Natural Philosopher@tnp@invalid.invalid to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Thu Nov 5 21:00:08 2026
    From Newsgroup: comp.os.linux.misc

    On 04/11/2026 19:51, Charlie Gibbs wrote:
    A mediocre assembly language programmer would not let lack of
    readability stand in the way. And once you get into self-modifying
    code there are all sorts of horrible things you can do.

    IBM macro assembler is about as machine code as COBOL
    --
    If you tell a lie big enough and keep repeating it, people will
    eventually come to believe it. The lie can be maintained only for such
    time as the State can shield the people from the political, economic
    and/or military consequences of the lie. It thus becomes vitally
    important for the State to use all of its powers to repress dissent, for
    the truth is the mortal enemy of the lie, and thus by extension, the
    truth is the greatest enemy of the State.

    Joseph Goebbels




    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From The Natural Philosopher@tnp@invalid.invalid to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Thu Nov 5 21:01:41 2026
    From Newsgroup: comp.os.linux.misc

    On 04/11/2026 14:54, Peter Flass wrote:
    On 11/3/26 15:15, Pancho wrote:

    If you want heavy usage, I would guess bank risk calcs, weather calcs,
    lots of science applications, running on compute grids, (or whatever
    the modern word is), blow this mainframe high i/o away.

    I think back in the day the quants preferred Sun.

    Not for transactuon processing
    --
    In theory, there is no difference between theory and practice.
    In practice, there is.
    -- Yogi Berra

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From The Natural Philosopher@tnp@invalid.invalid to comp.os.linux.misc on Thu Nov 5 21:02:31 2026
    From Newsgroup: comp.os.linux.misc

    On 05/11/2026 06:27, rbowman wrote:
    On Thu, 5 Nov 2026 03:34:03 -0000 (UTC), Lawrence DrCOOliveiro wrote:

    On Wed, 4 Nov 2026 19:04:22 -0700, Peter Flass wrote:

    x86 is RISC under the hood.

    ThatrCOs nonsense. ThatrCOs Intel/AMD PR talking.

    What do you call micro-operations?

    Micro operations
    --
    In theory, there is no difference between theory and practice.
    In practice, there is.
    -- Yogi Berra

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From The Natural Philosopher@tnp@invalid.invalid to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Thu Nov 5 21:05:26 2026
    From Newsgroup: comp.os.linux.misc

    On 04/11/2026 19:33, Johann 'Myrkraverk' Oskarsson wrote:
    Do I have something to say on this subject?-a No, not really; I just find
    it amusing when trolls accuse other people of being trolls, and then, presumably, killfile the person.-a Consider it "trash taking itself out."


    Best wishes, and happy philosophering of natural phenomena!

    Thank you.
    --
    rCLIt is hard to imagine a more stupid decision or more dangerous way of making decisions than by putting those decisions in the hands of people
    who pay no price for being wrong.rCY

    Thomas Sowell

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From The Natural Philosopher@tnp@invalid.invalid to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Thu Nov 5 21:09:15 2026
    From Newsgroup: comp.os.linux.misc

    On 04/11/2026 22:22, Scott Lurndal wrote:
    Lawrence =?iso-8859-13?q?D=FFOliveiro?= <ldo@nz.invalid> writes:
    On Wed, 4 Nov 2026 10:53:41 +0100, Carlos E. R. wrote:

    On 2026-11-04 03:23, Janis Papanagnou wrote:

    On 2026-11-03 12:45, Carlos E. R. wrote:

    I know at least one guy working on IBM mainframes, and he tells me
    they are in high need of good assembler programmers.

    Wow! - I'd have thought that assembler programmers should meanwhile
    be at "best" a niche, and systems programming with "C" (or
    similar).

    They need speed that compilers can not achieve.

    The worldrCOs fastest supercomputers run Linux. And that is mostly written >> in C, with very little assembly language.

    Again with the incorrect characterizations.

    The fastest supercomputers are running massive amounts of
    modern (and some ancient) Fortran code.

    Strictly he said 'running Linux, and *that* is mostly written mainly in C' Fortran is perfectly good language for scientific modelling, agreed
    --
    Microsoft : the best reason to go to Linux that ever existed.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From The Natural Philosopher@tnp@invalid.invalid to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Thu Nov 5 21:12:55 2026
    From Newsgroup: comp.os.linux.misc

    On 04/11/2026 23:14, Carlos E. R. wrote:
    On 2026-11-04 22:47, Lawrence DrCOOliveiro wrote:
    On Wed, 4 Nov 2026 10:53:41 +0100, Carlos E. R. wrote:

    On 2026-11-04 03:23, Janis Papanagnou wrote:

    On 2026-11-03 12:45, Carlos E. R. wrote:

    I know at least one guy working on IBM mainframes, and he tells me
    they are in high need of good assembler programmers.

    Wow! - I'd have thought that assembler programmers should meanwhile
    be at "best" a niche, and systems programming with "C" (or
    similar).

    They need speed that compilers can not achieve.

    The worldrCOs fastest supercomputers run Linux. And that is mostly written >> in C, with very little assembly language.

    Well, I know for a fact that this machine doesn't run any standard OS. I asked him. Also, the disks do not have a filesystem, either (I also
    asked). Of course most of the programmers work in some high level
    language, but there are certain critical operations that are optimized
    in assembler. I don't remember which, and obviously I'm not going to
    quote him breaking confidence. But if anyone is a good programmer in
    those machines, in assembler, they are hiring.

    My mathematical friend needs to use assembler routines to access some of
    the vector processing that the chipset, but not available C libraries, supports.

    IBM Macro assembler is practically a high level language, given the
    scale of macro usage in legacy systems
    --
    Microsoft : the best reason to go to Linux that ever existed.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From The Natural Philosopher@tnp@invalid.invalid to comp.os.linux.misc on Thu Nov 5 21:19:12 2026
    From Newsgroup: comp.os.linux.misc

    On 05/11/2026 20:50, c186282 wrote:
    Alas I don't remember any 6502s that were
    -a clocked much over 1mhz - while the Z-80s
    -a could run at four. MIGHT be that the Z-80
    -a required more cycles per instruction though,
    -a reducing that advantage - somebody here
    -a knows ..... I think the Z-80 needed twice
    -a as many clocks.

    That was the case The 6502 was in general a faster chip.

    In many ways the Z80/8080 owed its success to Gary Kildall's CP/M

    And it proved a more upgradeable architecture than the 6502.

    But ultimately IBM and Microsoft made the rather shit 8086 the future,
    rather than the Z8000, or 68000.
    --
    Truth welcomes investigation because truth knows investigation will lead
    to converts. It is deception that uses all the other techniques.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From John Ames@commodorejohn@gmail.com to comp.os.linux.misc on Thu Nov 5 13:19:40 2026
    From Newsgroup: comp.os.linux.misc

    On Thu, 5 Nov 2026 15:50:22 -0500
    c186282 <c186282@nnada.net> wrote:

    Alas I don't remember any 6502s that were clocked much over 1mhz -
    while the Z-80s could run at four. MIGHT be that the Z-80 required
    more cycles per instruction though, reducing that advantage -
    somebody here knows ..... I think the Z-80 needed twice as many
    clocks.

    It's a perennial debate between the Commodore and Spectrum partisans,
    yep. Z80 systems frequently ran at higher clock rates, but equivalent instructions were around twice as long (IIRC part of it was that its
    ALU was 4-bit internally.) On the other hand, there's less "register
    pressure" on the Z80. On the other other hand, most of the registers
    are just dance partners for the accumulator and/or pointers for load/
    store, etc. etc., da capo ad nauseam.

    (The 6502 did have versions up to 3-4 MHz even in its original form,
    IIRC, but of course that doesn't mean vendors of home computers back in
    the day actually *ran* things that fast. It got a 16-bit upgrade in the
    form of the 65C816, as well, but by that time personal computers were
    already moving towards 32-bit architectures...)

    Suffice to say, no matter which you're working with, creative thinking
    and a comprehensive knowledge of the instruction set go a long way.

    The m6809 was a nice chip too, but again Motorola never got the PRICE
    down fast enough so it got lost in the crowd.

    That was a nice one alright, rather beloved among Forth people.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Carlos E. R.@robin_listas@es.invalid to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Thu Nov 5 22:22:37 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-05 22:12, The Natural Philosopher wrote:
    On 04/11/2026 23:14, Carlos E. R. wrote:
    On 2026-11-04 22:47, Lawrence DrCOOliveiro wrote:
    On Wed, 4 Nov 2026 10:53:41 +0100, Carlos E. R. wrote:

    On 2026-11-04 03:23, Janis Papanagnou wrote:

    On 2026-11-03 12:45, Carlos E. R. wrote:

    I know at least one guy working on IBM mainframes, and he tells me >>>>>> they are in high need of good assembler programmers.

    Wow! - I'd have thought that assembler programmers should meanwhile
    be at "best" a niche, and systems programming with "C" (or
    similar).

    They need speed that compilers can not achieve.

    The worldrCOs fastest supercomputers run Linux. And that is mostly written >>> in C, with very little assembly language.

    Well, I know for a fact that this machine doesn't run any standard OS.
    I asked him. Also, the disks do not have a filesystem, either (I also
    asked). Of course most of the programmers work in some high level
    language, but there are certain critical operations that are optimized
    in assembler. I don't remember which, and obviously I'm not going to
    quote him breaking confidence. But if anyone is a good programmer in
    those machines, in assembler, they are hiring.

    My mathematical friend needs to use assembler routines to access some of
    the vector processing that the chipset, but not available-a C libraries, supports.

    IBM Macro-a assembler is practically a high level language, given the
    scale of macro usage in legacy systems

    It fits what I know.
    --
    Cheers,
    Carlos E.R.
    ESEfc-Efc+, EUEfc-Efc|.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Carlos E. R.@robin_listas@es.invalid to comp.os.linux.misc on Thu Nov 5 22:24:10 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-05 22:19, The Natural Philosopher wrote:
    On 05/11/2026 20:50, c186282 wrote:
    Alas I don't remember any 6502s that were
    -a-a clocked much over 1mhz - while the Z-80s
    -a-a could run at four. MIGHT be that the Z-80
    -a-a required more cycles per instruction though,
    -a-a reducing that advantage - somebody here
    -a-a knows ..... I think the Z-80 needed twice
    -a-a as many clocks.

    That was the case The 6502 was in general a faster chip.

    In many ways the Z80/8080 owed its success to Gary Kildall's CP/M

    And it proved a more upgradeable architecture than the 6502.

    But ultimately IBM-a and Microsoft made the rather shit 8086 the future, rather than the Z8000, or 68000.


    The 68000 had a beautiful assembler. Or rather, a beautiful instruction set.
    --
    Cheers,
    Carlos E.R.
    ESEfc-Efc+, EUEfc-Efc|.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From The Natural Philosopher@tnp@invalid.invalid to comp.os.linux.misc on Thu Nov 5 21:29:08 2026
    From Newsgroup: comp.os.linux.misc

    On 05/11/2026 21:19, John Ames wrote:
    The m6809 was a nice chip too, but again Motorola never got the PRICE
    down fast enough so it got lost in the crowd.
    That was a nice one alright, rather beloved among Forth people.

    I programmed that chip. Had a dire Introl C cross compiler running on a
    PDP/11
    So bad that assembler was often easier.

    We had DOS PcS running wordstar to write the code, then uploaded the
    source to the PDP by means I cannot recall - some sort of serial
    protocol. then yelled 'ok to compile?' across the lab and if there were
    any spare CPU cycles, we compiled and linked the target....
    ...I cannot recall how we got it onto the 6809 PROMS.

    Weird project, nice team though.
    --
    Civilization exists by geological consent, subject to change without notice.
    rCo Will Durant

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From John Levine@johnl@taugh.com to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Thu Nov 5 21:57:20 2026
    From Newsgroup: comp.os.linux.misc

    According to Thomas Koenig <tkoenig@netcologne.de>:
    Charlie Gibbs <cgibbs@kltpzyxm.invalid> schrieb:
    I agree. I've always believed that the difference between
    a mainframe and a mini had nothing to do with the size of
    the box, but everything to do with architecture, particularly
    I/O. Mainframes were record-oriented, while minis were
    byte-oriented.

    The PDP-8 was anything but byte-oriented, it was a 12-bit machine.

    Well, OK, word or byte depending on the vintage of the mini, but definitely
    not a record at at time.

    When a 360/30 read a card, it did a Start I/O instruction that pointed
    at a channel program that had the address and length of the buffer,
    then went away and did something else until it interrupted after the
    entire card was read in.

    On a PDP-8/I (I have the manual here), to read a card it first did a
    RCSE instruction to start a card moving, then for each column a busy
    wait loop or interrupt, RCSF to check and skip if the column is ready,
    RCRA or RCRB to read in the column, store it with a DCA instruction,
    then repeat for each column until an RCSD instruction said the card
    was done. It had to fetch each column within 1.5ms after its data was
    ready, but a PDP-8 was fast enough that the column routine might take
    50us so there was plenty of time to do other stuft.

    Plus, DECTape was block-oriented.

    It was but it was also pretty slow. The older TU55/TC01 DECtape used data break
    for I/O but for the later Omnibus PDP-8/e and 8/m had a cheaper interface that just read or wrote a word at a time in a tight loop, no data break or DMA.
    --
    Regards,
    John Levine, johnl@taugh.com, Primary Perpetrator of "The Internet for Dummies",
    Please consider the environment before reading this e-mail. https://jl.ly
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From John Ames@commodorejohn@gmail.com to comp.os.linux.misc on Thu Nov 5 13:57:18 2026
    From Newsgroup: comp.os.linux.misc

    On Thu, 5 Nov 2026 21:29:08 +0000
    The Natural Philosopher <tnp@invalid.invalid> wrote:

    The m6809 was a nice chip too, but again Motorola never got the
    PRICE down fast enough so it got lost in the crowd.

    That was a nice one alright, rather beloved among Forth people.

    I programmed that chip. Had a dire Introl C cross compiler running on
    a PDP/11
    So bad that assembler was often easier.

    We had DOS PcS running wordstar to write the code, then uploaded the
    source to the PDP by means I cannot recall - some sort of serial
    protocol. then yelled 'ok to compile?' across the lab and if there
    were any spare CPU cycles, we compiled and linked the target....
    ...I cannot recall how we got it onto the 6809 PROMS.

    Weird project, nice team though.

    I still have a Motorola trainer board for the thing stashed away, which
    I picked up from somewhere. Never have gotten around to doing anything
    with it.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Charlie Gibbs@cgibbs@kltpzyxm.invalid to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Thu Nov 5 21:57:51 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-05, The Natural Philosopher <tnp@invalid.invalid> wrote:

    On 04/11/2026 19:51, Charlie Gibbs wrote:

    A mediocre assembly language programmer would not let lack of
    readability stand in the way. And once you get into self-modifying
    code there are all sorts of horrible things you can do.

    IBM macro assembler is about as machine code as COBOL

    It can be. Or not. If the 360 assembler didn't support
    extended branch mnemonics, you could do them with macros -
    one line in, one line out.

    I do admit that the other extreme can become hairy.
    I once saw a recursive macro that solved the Towers
    of Hanoi. It generated no code; the solution was
    in the expanded source listing.
    --
    /~\ Charlie Gibbs | In this world there are
    \ / <cgibbs@kltpzyxm.invalid> | two kinds of people:
    X I'm really at ac.dekanfrus | 1. Those who can extrapolate
    / \ if you read it the right way. | from incomplete data.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Charlie Gibbs@cgibbs@kltpzyxm.invalid to comp.os.linux.misc on Thu Nov 5 21:57:52 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-05, Carlos E. R. <robin_listas@es.invalid> wrote:

    The 68000 had a beautiful assembler. Or rather, a beautiful instruction set.

    +1

    It did have a beautiful assembler too, though. I wrote one for
    the Amiga. I started from an article in Dr. Dobb's, then added
    things like branch optimization, a macro processor, and a cross-
    reference listing. Plus, of course, Amiga-format object code output.
    Matt Dillon combined it with his C compiler to create DICE, Dillon's
    Integrated C Environment. I tested the macro processor with a port
    of that recursive Towers of Hanoi macro that I mentioned elsewhere.
    --
    /~\ Charlie Gibbs | In this world there are
    \ / <cgibbs@kltpzyxm.invalid> | two kinds of people:
    X I'm really at ac.dekanfrus | 1. Those who can extrapolate
    / \ if you read it the right way. | from incomplete data.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Charlie Gibbs@cgibbs@kltpzyxm.invalid to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Thu Nov 5 21:57:53 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-05, Thomas Koenig <tkoenig@netcologne.de> wrote:

    Charlie Gibbs <cgibbs@kltpzyxm.invalid> schrieb:

    On 2026-11-04, John Levine <johnl@taugh.com> wrote:

    According to c186282 <c186282@nnada.net>:

    I can assure you that PDP-11s were minicomputers. I spent quite a while >>>>> programming them back in the 1970s.

    AT THE TIME, characterizing the PDP-11 as
    a "mini" is not wrong. I view a "mini" as
    a 'cut-down' mainframe.

    A mini was never a cut-down mainframe. Mainframes
    had channels, minis had simple I/O and fast interrupts.
    A 360/30 was considerably slower than a PDP-8, but its
    I/O architecture clearly made it a mainframe.

    I agree. I've always believed that the difference between
    a mainframe and a mini had nothing to do with the size of
    the box, but everything to do with architecture, particularly
    I/O. Mainframes were record-oriented, while minis were
    byte-oriented.

    The PDP-8 was anything but byte-oriented, it was a 12-bit machine.

    Plus, DECTape was block-oriented.

    You've got me there. It's amazing how slippery definitions become,
    even when starting from "I know one when I see one." I could say
    that minis tend to process a character at a time (e.g. a PDP-11
    driving a VT50) while mainframes tend to process a record at a
    time (e.g. a 360 filling a 2260's screen in a single operation).
    But exceptions (e.g. your DECTape) creep in...
    --
    /~\ Charlie Gibbs | In this world there are
    \ / <cgibbs@kltpzyxm.invalid> | two kinds of people:
    X I'm really at ac.dekanfrus | 1. Those who can extrapolate
    / \ if you read it the right way. | from incomplete data.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From c186282@c186282@nnada.net to comp.os.linux.misc on Thu Nov 5 17:15:38 2026
    From Newsgroup: comp.os.linux.misc

    On 11/5/26 16:19, John Ames wrote:
    On Thu, 5 Nov 2026 15:50:22 -0500
    c186282 <c186282@nnada.net> wrote:

    Alas I don't remember any 6502s that were clocked much over 1mhz -
    while the Z-80s could run at four. MIGHT be that the Z-80 required
    more cycles per instruction though, reducing that advantage -
    somebody here knows ..... I think the Z-80 needed twice as many
    clocks.

    It's a perennial debate between the Commodore and Spectrum partisans,
    yep. Z80 systems frequently ran at higher clock rates, but equivalent instructions were around twice as long (IIRC part of it was that its
    ALU was 4-bit internally.) On the other hand, there's less "register pressure" on the Z80. On the other other hand, most of the registers
    are just dance partners for the accumulator and/or pointers for load/
    store, etc. etc., da capo ad nauseam.

    (The 6502 did have versions up to 3-4 MHz even in its original form,
    IIRC, but of course that doesn't mean vendors of home computers back in
    the day actually *ran* things that fast. It got a 16-bit upgrade in the
    form of the 65C816, as well, but by that time personal computers were
    already moving towards 32-bit architectures...)

    Suffice to say, no matter which you're working with, creative thinking
    and a comprehensive knowledge of the instruction set go a long way.

    The m6809 was a nice chip too, but again Motorola never got the PRICE
    down fast enough so it got lost in the crowd.

    That was a nice one alright, rather beloved among Forth people.

    They were all GOOD CHIPS - and what you could do with
    one you could do with the others. Z80s got kind of
    stuck in the 'business' realm, but COULD serve well
    for home PCs too.

    The 65HC was alas 'too late'. By then the m68000 and I86/88
    were arriving. The 68000s were a joy to work with. Read
    somewhere that something about the design precluded moving
    up the speed ladder as much or as fast as Intel. Were
    also more expensive.

    The best thing to come out of the 6809 was OS-9 ...
    RTOS with a sort of UNIX-ish feel. Used to run it
    on a CoCo. I think you can still buy OS-9, but
    not sure there's any 6809-ish hardware to run
    it on anymore outside of E-Bay. Emulators running
    on Linux might work.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From c186282@c186282@nnada.net to comp.os.linux.misc on Thu Nov 5 17:18:12 2026
    From Newsgroup: comp.os.linux.misc

    On 11/5/26 16:24, Carlos E. R. wrote:
    On 2026-11-05 22:19, The Natural Philosopher wrote:
    On 05/11/2026 20:50, c186282 wrote:
    Alas I don't remember any 6502s that were
    -a-a clocked much over 1mhz - while the Z-80s
    -a-a could run at four. MIGHT be that the Z-80
    -a-a required more cycles per instruction though,
    -a-a reducing that advantage - somebody here
    -a-a knows ..... I think the Z-80 needed twice
    -a-a as many clocks.

    That was the case The 6502 was in general a faster chip.

    In many ways the Z80/8080 owed its success to Gary Kildall's CP/M

    And it proved a more upgradeable architecture than the 6502.

    But ultimately IBM-a and Microsoft made the rather shit 8086 the
    future, rather than the Z8000, or 68000.


    The 68000 had a beautiful assembler. Or rather, a beautiful instruction
    set.

    Still my favorite chip ... brilliant.

    But they couldn't match the volume/price of
    the Intel chips.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From John Ames@commodorejohn@gmail.com to comp.os.linux.misc on Thu Nov 5 14:29:04 2026
    From Newsgroup: comp.os.linux.misc

    On Thu, 5 Nov 2026 17:15:38 -0500
    c186282 <c186282@nnada.net> wrote:

    The 68000s were a joy to work with. Read somewhere that something
    about the design precluded moving up the speed ladder as much or as
    fast as Intel. Were also more expensive.

    The 68k was the final flourish of assembler-friendly CISC design - very pleasant to work with, but its plethora of addressing modes, IIRC, was
    one of the main things that kept it from keeping up in the Great Mega-
    hertz Race. The Coldfire spinoff architecture dropped some of the more
    exotic stuff and was able to rally for a bit, but Motorola had already
    moved on to PowerPC for PC-class processors by that point.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Charlie Gibbs@cgibbs@kltpzyxm.invalid to comp.os.linux.misc on Thu Nov 5 22:42:07 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-05, c186282 <c186282@nnada.net> wrote:

    The 65HC was alas 'too late'. By then the m68000 and I86/88
    were arriving. The 68000s were a joy to work with. Read
    somewhere that something about the design precluded moving
    up the speed ladder as much or as fast as Intel. Were
    also more expensive.

    Alas also too late. Especially the 68008, since IBM wanted
    to cheap out by using 8-bit support chips.

    It's yet another example of how it's better to be first than best. :-(
    --
    /~\ Charlie Gibbs | In this world there are
    \ / <cgibbs@kltpzyxm.invalid> | two kinds of people:
    X I'm really at ac.dekanfrus | 1. Those who can extrapolate
    / \ if you read it the right way. | from incomplete data.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Peter Flass@Peter@Iron-Spring.com to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Thu Nov 5 15:59:05 2026
    From Newsgroup: comp.os.linux.misc

    On 11/5/26 14:12, The Natural Philosopher wrote:
    On 04/11/2026 23:14, Carlos E. R. wrote:
    On 2026-11-04 22:47, Lawrence DrCOOliveiro wrote:
    On Wed, 4 Nov 2026 10:53:41 +0100, Carlos E. R. wrote:

    On 2026-11-04 03:23, Janis Papanagnou wrote:

    On 2026-11-03 12:45, Carlos E. R. wrote:

    I know at least one guy working on IBM mainframes, and he tells me >>>>>> they are in high need of good assembler programmers.

    Wow! - I'd have thought that assembler programmers should meanwhile
    be at "best" a niche, and systems programming with "C" (or
    similar).

    They need speed that compilers can not achieve.

    The worldrCOs fastest supercomputers run Linux. And that is mostly written >>> in C, with very little assembly language.

    Well, I know for a fact that this machine doesn't run any standard OS.
    I asked him. Also, the disks do not have a filesystem, either (I also
    asked). Of course most of the programmers work in some high level
    language, but there are certain critical operations that are optimized
    in assembler. I don't remember which, and obviously I'm not going to
    quote him breaking confidence. But if anyone is a good programmer in
    those machines, in assembler, they are hiring.

    My mathematical friend needs to use assembler routines to access some of
    the vector processing that the chipset, but not available-a C libraries, supports.

    IBM Macro-a assembler is practically a high level language, given the
    scale of macro usage in legacy systems


    It can be. Obviously there are probably a lot of macros in any assembler program, but it also gets down to the nitty-gritty of registers,
    condition codes, etc., and that is probably at least 80% of the lines of
    code.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Peter Flass@Peter@Iron-Spring.com to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Thu Nov 5 16:01:21 2026
    From Newsgroup: comp.os.linux.misc

    On 11/5/26 14:57, Charlie Gibbs wrote:
    On 2026-11-05, Thomas Koenig <tkoenig@netcologne.de> wrote:

    Charlie Gibbs <cgibbs@kltpzyxm.invalid> schrieb:

    On 2026-11-04, John Levine <johnl@taugh.com> wrote:

    According to c186282 <c186282@nnada.net>:

    I can assure you that PDP-11s were minicomputers. I spent quite a while >>>>>> programming them back in the 1970s.

    AT THE TIME, characterizing the PDP-11 as
    a "mini" is not wrong. I view a "mini" as
    a 'cut-down' mainframe.

    A mini was never a cut-down mainframe. Mainframes
    had channels, minis had simple I/O and fast interrupts.
    A 360/30 was considerably slower than a PDP-8, but its
    I/O architecture clearly made it a mainframe.

    I agree. I've always believed that the difference between
    a mainframe and a mini had nothing to do with the size of
    the box, but everything to do with architecture, particularly
    I/O. Mainframes were record-oriented, while minis were
    byte-oriented.

    The PDP-8 was anything but byte-oriented, it was a 12-bit machine.

    Plus, DECTape was block-oriented.

    You've got me there. It's amazing how slippery definitions become,
    even when starting from "I know one when I see one." I could say
    that minis tend to process a character at a time (e.g. a PDP-11
    driving a VT50) while mainframes tend to process a record at a
    time (e.g. a 360 filling a 2260's screen in a single operation).
    But exceptions (e.g. your DECTape) creep in...


    I usually just felt like mini=16 bits, mainframe=32 bits or more.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lawrence =?iso-8859-13?q?D=FFOliveiro?=@ldo@nz.invalid to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Thu Nov 5 23:04:23 2026
    From Newsgroup: comp.os.linux.misc

    On Thu, 05 Nov 2026 21:57:51 GMT, Charlie Gibbs wrote:

    I do admit that the other extreme can become hairy. I once saw a
    recursive macro that solved the Towers of Hanoi. It generated no
    code; the solution was in the expanded source listing.

    String-based macro languages in general can be pretty hairy.
    Towers of Hanoi wouldnrCOt even be that hard.

    ----
    dnl
    dnl Example m4 implementation of the GPM example output shown on page 64
    dnl of Bryan Higman's "A Comparative Study of Programming Languages", 1st ed. dnl
    dnl Created 2008 September 10 by Lawrence D'Oliveiro.
    dnl
    define(`repeat', `ifelse($2, 0, `', $2, 1, `$1', `$1`'repeat(`$1', eval($2 - 1))')')dnl
    dnl repeat(str, n) outputs n consecutive repetitions of str
    define(`n', `regexp(` x one two three four five six seven eight nine ten', `^\'repeat(` [^ ]+', $1)` \([^ ]+\)', `\1')')dnl
    dnl n(i) returns the word representation of i. Need to use the "repeat" macro dnl because m4's regexps don't support \{n\} repeat coefficients.
    define(`cap', `regexp(`$1', `^\(.\)\(.*\)$', `translit(`\1', `a-z', `A-Z')\2')')dnl
    dnl cap(s) returns s with the first letter capitalized.
    dnl
    define(`nrverses', 10)dnl
    define(`nmen', `cap(n($1)) ifelse($1, 1, man, men)')dnl
    dnl nmen(i) outputs the phrase "<i> man/men"
    define(`tomow', ` went to mow')dnl
    define(`mow1', `tomow a meadow')dnl
    define(`firstline', `nmen($1)`'tomow,mow1')dnl
    dnl firstline(i) outputs the first line of verse i.
    define(`restlines', `ifelse($1, 1, `nmen(1) and his dog,mow1', `nmen($1) restlines(eval($1 - 1))')')dnl
    dnl restlines(i) outputs the remaining lines of verse i
    define(`verse', `firstline($1)
    restlines($1)')dnl
    dnl verse(i) outputs the complete verse i
    define(`verses', `verse($1)`'ifelse($1, nrverses, `', `

    verses(eval($1 + 1))')')dnl
    dnl verses(i) outputs the verses from i up to the last one.
    dnl here we go...
    verses(1)
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lawrence =?iso-8859-13?q?D=FFOliveiro?=@ldo@nz.invalid to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Thu Nov 5 23:06:13 2026
    From Newsgroup: comp.os.linux.misc

    On Thu, 5 Nov 2026 16:01:21 -0700, Peter Flass wrote:

    I usually just felt like mini=16 bits, mainframe=32 bits or more.

    If yourCOd said rCLmini=ren18 bitsrCY, that would have been consistent with
    the state of the market up to about the mid-1970s.

    And then the 32-bit rCLsuperminisrCY appeared ...
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lawrence =?iso-8859-13?q?D=FFOliveiro?=@ldo@nz.invalid to comp.os.linux.misc,alt.folklore.computers on Thu Nov 5 23:17:56 2026
    From Newsgroup: comp.os.linux.misc

    On Thu, 5 Nov 2026 22:24:10 +0100, Carlos E. R. wrote:

    The 68000 had a beautiful assembler. Or rather, a beautiful
    instruction set.

    Definitely inspired by the PDP-11. Pity it was big-endian. Also, the
    division into rCLaddressrCY versus rCLdatarCY registers did seem to be a step backwards.

    HererCOs one trivia question: the decrement-and-branch instruction would
    fall out of the loop when the count reached -1, not 0 as the
    equivalent instruction in nearly every other computer architecture
    did. Why?
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From c186282@c186282@nnada.net to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Thu Nov 5 19:15:17 2026
    From Newsgroup: comp.os.linux.misc

    On 11/5/26 18:04, Lawrence DrCOOliveiro wrote:
    On Thu, 05 Nov 2026 21:57:51 GMT, Charlie Gibbs wrote:

    I do admit that the other extreme can become hairy. I once saw a
    recursive macro that solved the Towers of Hanoi. It generated no
    code; the solution was in the expanded source listing.

    String-based macro languages in general can be pretty hairy.
    Towers of Hanoi wouldnrCOt even be that hard.

    ----
    dnl
    dnl Example m4 implementation of the GPM example output shown on page 64
    dnl of Bryan Higman's "A Comparative Study of Programming Languages", 1st ed. dnl
    dnl Created 2008 September 10 by Lawrence D'Oliveiro.
    dnl
    define(`repeat', `ifelse($2, 0, `', $2, 1, `$1', `$1`'repeat(`$1', eval($2 - 1))')')dnl
    dnl repeat(str, n) outputs n consecutive repetitions of str
    define(`n', `regexp(` x one two three four five six seven eight nine ten', `^\'repeat(` [^ ]+', $1)` \([^ ]+\)', `\1')')dnl
    dnl n(i) returns the word representation of i. Need to use the "repeat" macro dnl because m4's regexps don't support \{n\} repeat coefficients. define(`cap', `regexp(`$1', `^\(.\)\(.*\)$', `translit(`\1', `a-z', `A-Z')\2')')dnl
    dnl cap(s) returns s with the first letter capitalized.
    dnl
    define(`nrverses', 10)dnl
    define(`nmen', `cap(n($1)) ifelse($1, 1, man, men)')dnl
    dnl nmen(i) outputs the phrase "<i> man/men"
    define(`tomow', ` went to mow')dnl
    define(`mow1', `tomow a meadow')dnl
    define(`firstline', `nmen($1)`'tomow,mow1')dnl
    dnl firstline(i) outputs the first line of verse i.
    define(`restlines', `ifelse($1, 1, `nmen(1) and his dog,mow1', `nmen($1) restlines(eval($1 - 1))')')dnl
    dnl restlines(i) outputs the remaining lines of verse i
    define(`verse', `firstline($1)
    restlines($1)')dnl
    dnl verse(i) outputs the complete verse i
    define(`verses', `verse($1)`'ifelse($1, nrverses, `', `

    verses(eval($1 + 1))')')dnl
    dnl verses(i) outputs the verses from i up to the last one.
    dnl here we go...
    verses(1)


    Well, that's totally transparent :-)

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From rbowman@bowman@montana.com to comp.os.linux.misc on Fri Nov 6 00:38:59 2026
    From Newsgroup: comp.os.linux.misc

    On Thu, 5 Nov 2026 22:24:10 +0100, Carlos E. R. wrote:

    On 2026-11-05 22:19, The Natural Philosopher wrote:
    On 05/11/2026 20:50, c186282 wrote:
    Alas I don't remember any 6502s that were
    -a-a clocked much over 1mhz - while the Z-80s could run at four. MIGHT
    -a-a be that the Z-80 required more cycles per instruction though,
    -a-a reducing that advantage - somebody here knows ..... I think the
    -a-a Z-80 needed twice as many clocks.

    That was the case The 6502 was in general a faster chip.

    In many ways the Z80/8080 owed its success to Gary Kildall's CP/M

    And it proved a more upgradeable architecture than the 6502.

    But ultimately IBM-a and Microsoft made the rather shit 8086 the future,
    rather than the Z8000, or 68000.


    The 68000 had a beautiful assembler. Or rather, a beautiful instruction
    set.

    I had a 68000 development board it was a nice instruction set. Some Z80 designs were already doing bank switching; the 8086 pasted it into the
    CPU.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From rbowman@bowman@montana.com to comp.os.linux.misc on Fri Nov 6 00:51:46 2026
    From Newsgroup: comp.os.linux.misc

    On Thu, 5 Nov 2026 21:19:12 +0000, The Natural Philosopher wrote:

    But ultimately IBM and Microsoft made the rather shit 8086 the future, rather than the Z8000, or 68000.

    Exxon had bought Zilog and IBM was in a pissing contest with Exxon so that
    was out. Motorola couldn't come up with 68008 parts that would work with
    the cheap 8 bit peripherals. IBM had used the 8085 in the System 23
    successor to the 5100 series so had some familiarity with Intel.

    Intel saw the 8088/8086 as an interim solution while they developed the
    iAPX 432. The 432 was an utter failure so here we are today.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From rbowman@bowman@montana.com to comp.os.linux.misc on Fri Nov 6 00:59:31 2026
    From Newsgroup: comp.os.linux.misc

    On Thu, 5 Nov 2026 13:19:40 -0800, John Ames wrote:

    On Thu, 5 Nov 2026 15:50:22 -0500 c186282 <c186282@nnada.net> wrote:

    Alas I don't remember any 6502s that were clocked much over 1mhz -
    while the Z-80s could run at four. MIGHT be that the Z-80 required more
    cycles per instruction though, reducing that advantage - somebody here
    knows ..... I think the Z-80 needed twice as many clocks.

    It's a perennial debate between the Commodore and Spectrum partisans,
    yep.
    Z80 systems frequently ran at higher clock rates, but equivalent
    instructions were around twice as long (IIRC part of it was that its ALU
    was 4-bit internally.) On the other hand, there's less "register
    pressure"
    on the Z80. On the other other hand, most of the registers are just
    dance partners for the accumulator and/or pointers for load/ store, etc. etc., da capo ad nauseam.

    Quite a few Z80 systems ran at 3.57954 MHz. Love those cheap NTSC color
    burst crystals.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Peter Flass@Peter@Iron-Spring.com to comp.os.linux.misc,alt.folklore.computers on Thu Nov 5 19:37:43 2026
    From Newsgroup: comp.os.linux.misc

    On 11/5/26 16:17, Lawrence DrCOOliveiro wrote:
    On Thu, 5 Nov 2026 22:24:10 +0100, Carlos E. R. wrote:

    The 68000 had a beautiful assembler. Or rather, a beautiful
    instruction set.

    Definitely inspired by the PDP-11. Pity it was big-endian. Also, the
    division into rCLaddressrCY versus rCLdatarCY registers did seem to be a step backwards.

    HererCOs one trivia question: the decrement-and-branch instruction would
    fall out of the loop when the count reached -1, not 0 as the
    equivalent instruction in nearly every other computer architecture
    did. Why?

    Big-endian is the only true way.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lars Poulsen@lars@beagle-ears.com to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Thu Nov 5 18:43:16 2026
    From Newsgroup: comp.os.linux.misc

    On 11/4/2026 11:51, Charlie Gibbs wrote:
    On the other side of the coin, someone at Univac decided that
    the key to productivity was not so much a quick response time
    as a consistent one - which led to their system inserting
    delays if a response was ready too soon. :-p

    Back in my University Computer Center days around 1972, I was hacking
    a local email feature into the EXEC-8 editor, and decided that less
    variation in response times would be a good thing, so I added a 1/2
    second wait before each editor command prompt. I got in serious trouble
    for this, because the process scheduler assumed that 500 ms was "a long
    time", and a program that was going to wait for a long time should be
    swapped out of memory.
    --
    Lars Poulsen - an old geek in Santa Barbara, California
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From John Levine@johnl@taugh.com to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Fri Nov 6 02:50:35 2026
    From Newsgroup: comp.os.linux.misc

    According to Peter Flass <Peter@Iron-Spring.com>:
    The PDP-8 was anything but byte-oriented, it was a 12-bit machine.

    Plus, DECTape was block-oriented.

    You've got me there. It's amazing how slippery definitions become,
    even when starting from "I know one when I see one." I could say
    that minis tend to process a character at a time (e.g. a PDP-11
    driving a VT50) while mainframes tend to process a record at a
    time (e.g. a 360 filling a 2260's screen in a single operation).
    But exceptions (e.g. your DECTape) creep in...

    I usually just felt like mini=16 bits, mainframe=32 bits or more.

    Oh you young folks. The 12 bit PDP-8 and the 18 but PDP-1, 4, 7, 9, and 15 were defintiely minicomputers.

    On the other hand the Interdata 7/32 and 8/32 to which Unix was ported (twice) in
    the 1970s was a 32 bit supermini. Its instruction set was reminiscent of S/360 but the simple I/O made it a mini..
    --
    Regards,
    John Levine, johnl@taugh.com, Primary Perpetrator of "The Internet for Dummies",
    Please consider the environment before reading this e-mail. https://jl.ly
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From c186282@c186282@nnada.net to comp.os.linux.misc,alt.folklore.computers on Thu Nov 5 21:54:12 2026
    From Newsgroup: comp.os.linux.misc

    On 11/5/26 21:37, Peter Flass wrote:
    On 11/5/26 16:17, Lawrence DrCOOliveiro wrote:
    On Thu, 5 Nov 2026 22:24:10 +0100, Carlos E. R. wrote:

    The 68000 had a beautiful assembler. Or rather, a beautiful
    instruction set.

    Definitely inspired by the PDP-11. Pity it was big-endian. Also, the
    division into rCLaddressrCY versus rCLdatarCY registers did seem to be a step
    backwards.

    HererCOs one trivia question: the decrement-and-branch instruction would
    fall out of the loop when the count reached -1, not 0 as the
    equivalent instruction in nearly every other computer architecture
    did. Why?

    Big-endian is the only true way.

    Ah, "religion" ...

    Either way works fine.

    PREF 'big' just for obviousness, but it's
    not necessary.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lawrence =?iso-8859-13?q?D=FFOliveiro?=@ldo@nz.invalid to comp.os.linux.misc,alt.folklore.computers on Fri Nov 6 04:24:54 2026
    From Newsgroup: comp.os.linux.misc

    On Thu, 5 Nov 2026 19:37:43 -0700, Peter Flass wrote:

    On 11/5/26 16:17, Lawrence DrCOOliveiro wrote:

    On Thu, 5 Nov 2026 22:24:10 +0100, Carlos E. R. wrote:

    The 68000 had a beautiful assembler. Or rather, a beautiful
    instruction set.

    Definitely inspired by the PDP-11. Pity it was big-endian.

    Big-endian is the only true way.

    The 68000 is a good example of the problems with it.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From c186282@c186282@nnada.net to comp.os.linux.misc,alt.folklore.computers on Thu Nov 5 23:31:02 2026
    From Newsgroup: comp.os.linux.misc

    On 11/5/26 23:24, Lawrence DrCOOliveiro wrote:
    On Thu, 5 Nov 2026 19:37:43 -0700, Peter Flass wrote:

    On 11/5/26 16:17, Lawrence DrCOOliveiro wrote:

    On Thu, 5 Nov 2026 22:24:10 +0100, Carlos E. R. wrote:

    The 68000 had a beautiful assembler. Or rather, a beautiful
    instruction set.

    Definitely inspired by the PDP-11. Pity it was big-endian.

    Big-endian is the only true way.

    The 68000 is a good example of the problems with it.

    Umm .. not really IMHO. In theory it's
    harder to set up Big for a CPU, but the
    comprehension issues still favor Big.

    The "AI"s ... they may eventually come to
    favor Small for their own designs.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lawrence =?iso-8859-13?q?D=FFOliveiro?=@ldo@nz.invalid to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Fri Nov 6 04:34:00 2026
    From Newsgroup: comp.os.linux.misc

    On Wed, 04 Nov 2026 19:51:29 GMT, Charlie Gibbs wrote:

    I remember reading about an experiment done in an IBM shop where
    they shoveled in enough hardware and tweaked the protocols so that
    terminal users got sub-second response time. Apparently it increased throughput even more than one would expect by just scaling the
    response time - users could develop a rhythm and not have their
    minds wander while waiting for a response.

    I remember this, too: it was one of the first experiments into
    human-factors aspects of computer systems. And from IBM, too: the
    maker of systems that were not exactly famous for a lot of thought
    going into their human-factors aspects ...

    On the other side of the coin, someone at Univac decided that the
    key to productivity was not so much a quick response time as a
    consistent one - which led to their system inserting delays if a
    response was ready too soon. :-p

    And hererCOs another tale: this was probably from the very earliest days
    of computers, but the story I remember reading involves users entering
    a query, and how long it took to get back a response. Apparently, if
    the response came back *too* quickly, the users felt that the computer
    somehow wasnrCOt treating their question with the importance it
    deserved.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Robert Riches@spamtrap42@jacob21819.net to comp.os.linux.misc on Fri Nov 6 05:03:59 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-06, Peter Flass <Peter@Iron-Spring.com> wrote:
    On 11/5/26 16:17, Lawrence DrCOOliveiro wrote:
    On Thu, 5 Nov 2026 22:24:10 +0100, Carlos E. R. wrote:

    The 68000 had a beautiful assembler. Or rather, a beautiful
    instruction set.

    Definitely inspired by the PDP-11. Pity it was big-endian. Also, the
    division into rCLaddressrCY versus rCLdatarCY registers did seem to be a step
    backwards.

    HererCOs one trivia question: the decrement-and-branch instruction would
    fall out of the loop when the count reached -1, not 0 as the
    equivalent instruction in nearly every other computer architecture
    did. Why?

    Big-endian is the only true way.

    It is if you insist on string comparison and integer comparison
    being identical or being able to read digits left to right and
    put them in increasing memory addresses.

    In 2001, while working on my MS, the OOP class required a term
    project based on Squeak, an implementation of Smalltalk. To save
    travel time, I wanted to do the project on my Alpha machine at
    home rather than on the hardware at the grad school. However,
    Squeak was not yet ported to the Alpha.

    The C code level of the Squeak program treated everything as if
    it were a 32-bit int. The same memory address would be used to
    store or load a pointer and load or store a 32-bit int. Because
    Alpha was little endian, it was relatively easy to port Squeak to
    run on the Alpha, and I completed the term project.

    That could not have been done on a big endian machine, because
    when the Squeak C code stored a pointer to memory and then used
    the same memory address to load it as a 32-bit int, a big-endian
    machine would have returned the most significant half, not the
    least significant half.

    One of the related things I had to do was to set a 'personality'
    flag so that the Squeak program would run as if it were
    constrained to a 32-bit address space, which kept the most
    significant half of a 64-bit pointer stuck as zero.
    --
    Robert Riches
    spamtrap42@jacob21819.net
    (Yes, that is one of my email addresses.)
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Thomas Koenig@tkoenig@netcologne.de to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Fri Nov 6 06:35:23 2026
    From Newsgroup: comp.os.linux.misc

    The Natural Philosopher <tnp@invalid.invalid> schrieb:

    IBM Macro assembler is practically a high level language, given the
    scale of macro usage in legacy systems

    High-level languages abstract away the underlying machine, which
    I presume the Macro assembler does not.

    Plus, does it do register allocation?

    (OK, then Wirth's PL/360 creeps in, which does not do so either,
    but then I would argue that that is a glorified assembler - it also
    does not do operator precenence).
    --
    This USENET posting was made without artificial intelligence,
    artificial impertinence, artificial arrogance, artificial stupidity,
    artificial flavorings or artificial colorants.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Thomas Koenig@tkoenig@netcologne.de to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Fri Nov 6 06:38:00 2026
    From Newsgroup: comp.os.linux.misc

    Charlie Gibbs <cgibbs@kltpzyxm.invalid> schrieb:
    On 2026-11-04, Johann 'Myrkraverk' Oskarsson <johann@myrkraverk.invalid> wrote:

    And if you think the fruit vendor is better, you can rest easy, because
    at one time they refused to update my card number. Actually refused.
    After reaching the /right/ support person, that person could not help
    figure out why the website told me the card number had been updated, but
    then the TV application always tried to use the old card number.

    Ah, you've been around the mulberry bush with Netflix too, I presume.
    We recently found we couldn't log in to Netflix because they didn't
    like our card number - and quoted the last four digits and date of
    a card that expired seven years ago. I wonder how they had managed
    to bill us the previous month.

    I called their customer support line, and the droid became downright belligerent, claiming that it wasn't their fault but our bank's.
    I wound up hanging up on him, and we cancelled our Netflix account.

    So whoever was at fault for the issue did you a huge favor.
    --
    This USENET posting was made without artificial intelligence,
    artificial impertinence, artificial arrogance, artificial stupidity,
    artificial flavorings or artificial colorants.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Thomas Koenig@tkoenig@netcologne.de to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Fri Nov 6 06:40:43 2026
    From Newsgroup: comp.os.linux.misc

    Peter Flass <Peter@Iron-Spring.com> schrieb:
    On 11/3/26 20:30, c186282 wrote:

    -a NOTICED - my principal bank - a LOT of transactions
    -a come up with a "pending" label. Might take DAYS. This
    -a is not some 3rd-tier bank either.

    -a Two or more larger credit/debit card charges, not
    -a sure HOW they 'compete' in that system. If too
    -a near each other one MIGHT get bumped. Problem !

    -a In short, even 'online/instant' banking may not
    -a be THAT instant.

    Some of this is communication between banks, and some is your bank
    milking the float for as long as is legally allowed.

    In Europe, you can do "immediate" bank transfers for free. Takes a
    few seconds. Banks are required by law to provide this, but they
    differ in the amount that they allow.

    In principle, this is good. But it also allows scammers to clean out
    accounts more easily, because otherwise you could cancel an order
    if you noticed anything is amiss.
    --
    This USENET posting was made without artificial intelligence,
    artificial impertinence, artificial arrogance, artificial stupidity,
    artificial flavorings or artificial colorants.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From rbowman@bowman@montana.com to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Fri Nov 6 06:44:31 2026
    From Newsgroup: comp.os.linux.misc

    On Thu, 5 Nov 2026 18:43:16 -0800, Lars Poulsen wrote:

    I got in serious trouble for this,
    because the process scheduler assumed that 500 ms was "a long time", and
    a program that was going to wait for a long time should be swapped out
    of memory.

    It is a long time :) For example if you click on a button and there is no perceptible change for 500 ms the users will complain about the program
    being slow.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Janis Papanagnou@janis_papanagnou+ng@hotmail.com to comp.os.linux.misc on Fri Nov 6 08:51:28 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-06 07:44, rbowman wrote:
    On Thu, 5 Nov 2026 18:43:16 -0800, Lars Poulsen wrote:

    I got in serious trouble for this,
    because the process scheduler assumed that 500 ms was "a long time", and
    a program that was going to wait for a long time should be swapped out
    of memory.

    It is a long time :) For example if you click on a button and there is no perceptible change for 500 ms the users will complain about the program
    being slow.

    There's actually standards and best practices for that investigated in
    surveys; accepted upper limits for interactive activities typically
    vary in about the range of 100-300 ms. Anything above that is usually perceived as extremely disturbing a continuous activity. (These limits
    had been our landmarks, what we needed to achieve and guarantee in our
    client software development 30+ years ago. - Of course we strived for
    minimum latency.)

    Janis

    PS: It's beyond me why one would artificially add 500-1000 ms delays
    (as had been reported here), and who would actually accept that. But
    maybe that was just a joke (that I missed), or a hilarious myth.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Richard Kettlewell@invalid@invalid.invalid to comp.os.linux.misc,alt.folklore.computers on Fri Nov 6 08:46:25 2026
    From Newsgroup: comp.os.linux.misc

    Lawrence DrCOOliveiro <ldo@nz.invalid> writes:
    Carlos E. R. wrote:
    The 68000 had a beautiful assembler. Or rather, a beautiful
    instruction set.

    Definitely inspired by the PDP-11. Pity it was big-endian. Also, the
    division into rCLaddressrCY versus rCLdatarCY registers did seem to be a step backwards.

    HererCOs one trivia question: the decrement-and-branch instruction would
    fall out of the loop when the count reached -1, not 0 as the
    equivalent instruction in nearly every other computer architecture
    did. Why?

    At the surface level, because DBcc is a rather bizarrely shaped
    instruction, with the condition code applying to the whole instruction
    rather; operating with inverse sense; and branching unless the decrement results in a value of -1 (i.e. 0xFFFF). So thererCOs _two_ conditional
    steps (unless the condition is a constant, e.g. the common use of
    DBF). Page 4-90 in the PRM[1] shows the definition.

    So thererCOs several things to explain. I donrCOt have access to the
    thinking of the designers, but I think the following are plausible explanations:

    1) The whole-instruction conditionality allows for early exits from
    loops (depending of course how they are structured). So you get two
    checks for a single instruction.

    2) Using 0xFFFF for the sentinel value comes, microarchitecturally, rCyfor
    freerCO because it just means checking the carry out from the
    decrement, which is 1 in precisely the situation where the decrement
    reaches -1. Using 0 as the sentintel would mean testing the whole
    result value.

    3) The inverse sense of the conditional. The PRM explains this one in
    terms of rCyUNTILrCO loops, which were perhaps more popular in the
    programming languages of the time than they are today.

    [1] https://www.nxp.com/docs/en/reference-manual/M68000PRM.pdf
    --
    https://www.greenend.org.uk/rjk/
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Carlos E. R.@robin_listas@es.invalid to comp.os.linux.misc,alt.folklore.computers on Fri Nov 6 14:30:35 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-06 00:17, Lawrence DrCOOliveiro wrote:
    On Thu, 5 Nov 2026 22:24:10 +0100, Carlos E. R. wrote:

    The 68000 had a beautiful assembler. Or rather, a beautiful
    instruction set.

    Definitely inspired by the PDP-11. Pity it was big-endian. Also, the
    division into rCLaddressrCY versus rCLdatarCY registers did seem to be a step backwards.

    HererCOs one trivia question: the decrement-and-branch instruction would
    fall out of the loop when the count reached -1, not 0 as the
    equivalent instruction in nearly every other computer architecture
    did. Why?

    Mmm. I don't remember. A flag went up?
    --
    Cheers,
    Carlos E.R.
    ESEfc-Efc+, EUEfc-Efc|.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Peter Flass@Peter@Iron-Spring.com to comp.os.linux.misc,alt.folklore.computers on Fri Nov 6 07:41:01 2026
    From Newsgroup: comp.os.linux.misc

    On 11/6/26 01:46, Richard Kettlewell wrote:
    Lawrence DrCOOliveiro <ldo@nz.invalid> writes:
    Carlos E. R. wrote:
    The 68000 had a beautiful assembler. Or rather, a beautiful
    instruction set.

    Definitely inspired by the PDP-11. Pity it was big-endian. Also, the
    division into rCLaddressrCY versus rCLdatarCY registers did seem to be a step
    backwards.

    HererCOs one trivia question: the decrement-and-branch instruction would
    fall out of the loop when the count reached -1, not 0 as the
    equivalent instruction in nearly every other computer architecture
    did. Why?

    At the surface level, because DBcc is a rather bizarrely shaped
    instruction, with the condition code applying to the whole instruction rather; operating with inverse sense; and branching unless the decrement results in a value of -1 (i.e. 0xFFFF). So thererCOs _two_ conditional
    steps (unless the condition is a constant, e.g. the common use of
    DBF). Page 4-90 in the PRM[1] shows the definition.

    So thererCOs several things to explain. I donrCOt have access to the
    thinking of the designers, but I think the following are plausible explanations:

    1) The whole-instruction conditionality allows for early exits from
    loops (depending of course how they are structured). So you get two
    checks for a single instruction.

    2) Using 0xFFFF for the sentinel value comes, microarchitecturally, rCyfor
    freerCO because it just means checking the carry out from the
    decrement, which is 1 in precisely the situation where the decrement
    reaches -1. Using 0 as the sentintel would mean testing the whole
    result value.

    3) The inverse sense of the conditional. The PRM explains this one in
    terms of rCyUNTILrCO loops, which were perhaps more popular in the
    programming languages of the time than they are today.

    [1] https://www.nxp.com/docs/en/reference-manual/M68000PRM.pdf



    As I recall for the IBM 1130 conditional branch instruction you have to
    code for which conditions the branch is NOT to be taken.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From scott@scott@slp53.sl.home (Scott Lurndal) to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Fri Nov 6 15:45:23 2026
    From Newsgroup: comp.os.linux.misc

    The Natural Philosopher <tnp@invalid.invalid> writes:
    On 04/11/2026 14:54, Peter Flass wrote:
    On 11/3/26 15:15, Pancho wrote:

    If you want heavy usage, I would guess bank risk calcs, weather calcs,
    lots of science applications, running on compute grids, (or whatever
    the modern word is), blow this mainframe high i/o away.

    I think back in the day the quants preferred Sun.

    Not for transactuon processing

    You clearly don't know what a "quant" is, in this context.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Charlie Gibbs@cgibbs@kltpzyxm.invalid to comp.os.linux.misc on Fri Nov 6 20:51:46 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-06, rbowman <bowman@montana.com> wrote:

    On Thu, 5 Nov 2026 13:19:40 -0800, John Ames wrote:

    On Thu, 5 Nov 2026 15:50:22 -0500 c186282 <c186282@nnada.net> wrote:

    Alas I don't remember any 6502s that were clocked much over 1mhz -
    while the Z-80s could run at four. MIGHT be that the Z-80 required more
    cycles per instruction though, reducing that advantage - somebody here
    knows ..... I think the Z-80 needed twice as many clocks.

    It's a perennial debate between the Commodore and Spectrum partisans,
    yep.
    Z80 systems frequently ran at higher clock rates, but equivalent
    instructions were around twice as long (IIRC part of it was that its ALU
    was 4-bit internally.) On the other hand, there's less "register
    pressure"
    on the Z80. On the other other hand, most of the registers are just
    dance partners for the accumulator and/or pointers for load/ store, etc.
    etc., da capo ad nauseam.

    Quite a few Z80 systems ran at 3.57954 MHz. Love those cheap NTSC color burst crystals.

    Amigas ran at 7.16 MHz. They showed up at a lot of cable TV
    stations generating those time/date/weather screens. And the
    Video Toaster did cutting-edge video effects. Todd Rundgren
    created his "Change Myself" video using a bank of ten
    Toaster-equipped Amigas.
    --
    /~\ Charlie Gibbs | In this world there are
    \ / <cgibbs@kltpzyxm.invalid> | two kinds of people:
    X I'm really at ac.dekanfrus | 1. Those who can extrapolate
    / \ if you read it the right way. | from incomplete data.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Charlie Gibbs@cgibbs@kltpzyxm.invalid to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Fri Nov 6 20:51:49 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-06, Lawrence DrCOOliveiro <ldo@nz.invalid> wrote:

    And hererCOs another tale: this was probably from the very earliest days
    of computers, but the story I remember reading involves users entering
    a query, and how long it took to get back a response. Apparently, if
    the response came back *too* quickly, the users felt that the computer somehow wasnrCOt treating their question with the importance it
    deserved.

    I wrote a program that searches a large collection of files for
    various data. I noticed the same thing: it ran so fast that
    people weren't taking it seriously. It didn't help that the
    search parameter screen was too easy to use, either.

    I guess that in this modern age, where complexity is equated
    with sophistication, any sign of simplicity or efficiency
    is suspect.
    --
    /~\ Charlie Gibbs | In this world there are
    \ / <cgibbs@kltpzyxm.invalid> | two kinds of people:
    X I'm really at ac.dekanfrus | 1. Those who can extrapolate
    / \ if you read it the right way. | from incomplete data.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Charlie Gibbs@cgibbs@kltpzyxm.invalid to comp.os.linux.misc,alt.folklore.computers on Fri Nov 6 20:51:47 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-06, c186282 <c186282@nnada.net> wrote:

    On 11/5/26 21:37, Peter Flass wrote:

    On 11/5/26 16:17, Lawrence DrCOOliveiro wrote:

    On Thu, 5 Nov 2026 22:24:10 +0100, Carlos E. R. wrote:

    The 68000 had a beautiful assembler. Or rather, a beautiful
    instruction set.

    Definitely inspired by the PDP-11. Pity it was big-endian. Also, the
    division into rCLaddressrCY versus rCLdatarCY registers did seem to be a step
    backwards.

    HererCOs one trivia question: the decrement-and-branch instruction would >>> fall out of the loop when the count reached -1, not 0 as the
    equivalent instruction in nearly every other computer architecture
    did. Why?

    Big-endian is the only true way.

    Ah, "religion" ...

    Either way works fine.

    PREF 'big' just for obviousness, but it's
    not necessary.

    You're reading me now from the right to the left
    And it's coming out just the way you might expect
    Chinatown calculation, that's cool with no regrets
    (Chop suey, ch-chop suey...)
    -- Doug and the Slugs
    --
    /~\ Charlie Gibbs | In this world there are
    \ / <cgibbs@kltpzyxm.invalid> | two kinds of people:
    X I'm really at ac.dekanfrus | 1. Those who can extrapolate
    / \ if you read it the right way. | from incomplete data.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Charlie Gibbs@cgibbs@kltpzyxm.invalid to comp.os.linux.misc on Fri Nov 6 20:51:49 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-06, rbowman <bowman@montana.com> wrote:

    On Thu, 5 Nov 2026 18:43:16 -0800, Lars Poulsen wrote:

    I got in serious trouble for this,
    because the process scheduler assumed that 500 ms was "a long time", and
    a program that was going to wait for a long time should be swapped out
    of memory.

    It is a long time :) For example if you click on a button and there is no perceptible change for 500 ms the users will complain about the program being slow.

    And then he'll click on the button again, and wind up with some
    undesired double-click action when the program does get around
    to dealing with it.
    --
    /~\ Charlie Gibbs | In this world there are
    \ / <cgibbs@kltpzyxm.invalid> | two kinds of people:
    X I'm really at ac.dekanfrus | 1. Those who can extrapolate
    / \ if you read it the right way. | from incomplete data.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lawrence =?iso-8859-13?q?D=FFOliveiro?=@ldo@nz.invalid to comp.os.linux.misc,alt.folklore.computers on Fri Nov 6 20:53:43 2026
    From Newsgroup: comp.os.linux.misc

    On Fri, 6 Nov 2026 14:30:35 +0100, Carlos E. R. wrote:

    On 2026-11-06 00:17, Lawrence DrCOOliveiro wrote:

    HererCOs one trivia question: the decrement-and-branch instruction
    would fall out of the loop when the count reached -1, not 0 as the
    equivalent instruction in nearly every other computer architecture
    did. Why?

    Mmm. I don't remember. A flag went up?

    It was quite a clever feature, which I think ended up widely
    misunderstood. Otherwise more architectures would have copied it.

    The reason is, so that if the loop count was zero, you could just
    fall out of the loop automatically, without having to do an explicit
    test for that case.

    Example skeleton loop: (remember, the loop count was only 16 bits):

    move.w -2loop-count-+, dN ; dN = some convenient d-register
    bra loop_bottom
    loop_top:
    ... body of loop ...
    loop_bottom:
    dbra dN, loop_top
    ... here on loop end ...

    Notice this will loop exactly -2loop-count-+ times, even if -2loop-count-+
    is zero.

    If you want a 32-bit loop count, it only took a few more instructions:

    move.l -2loop-count-+, dN
    bra loop_bottom
    loop_top1:
    swap dN
    loop_top2:
    ... body of loop ...
    loop_bottom:
    dbra dN, loop_top2
    swap dN
    dbra dN, loop_top1
    ... here on loop end ...

    (This is all from memory, itrCOs been a long while since I wrote any 68000-family assembler.)
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From rbowman@bowman@montana.com to comp.os.linux.misc on Fri Nov 6 21:10:31 2026
    From Newsgroup: comp.os.linux.misc

    On Fri, 6 Nov 2026 08:51:28 +0100, Janis Papanagnou wrote:

    There's actually standards and best practices for that investigated in surveys; accepted upper limits for interactive activities typically vary
    in about the range of 100-300 ms. Anything above that is usually
    perceived as extremely disturbing a continuous activity. (These limits
    had been our landmarks, what we needed to achieve and guarantee in our
    client software development 30+ years ago. - Of course we strived for
    minimum latency.)

    Many people don't have a good feeling for time units under one second
    although they constantly are doing performing actions in milliseconds. A
    shot timer is used in pistol competitions. It's nothing complicated, a microphone and circuitry capable of measuring time. You could cook one up
    with an Arduino. A competent marksman, not a first rank shooter, is
    placing shots on target in the 200 ms range. In that time period you are aware of the ejected brass flying off, the front sight falling back into alignment with the rear sight and target, and squeezing the trigger.

    Catching the egg you accidentally knocked off the counter probably takes a
    bit longer since you weren't anticipating it but most people, other than
    those sitting at a traffic light when it turns green, react in well under
    a second.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Kaz Kylheku@046-301-5902@kylheku.com to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Fri Nov 6 21:29:00 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-06, Lars Poulsen <lars@beagle-ears.com> wrote:
    On 11/4/2026 11:51, Charlie Gibbs wrote:
    On the other side of the coin, someone at Univac decided that
    the key to productivity was not so much a quick response time
    as a consistent one - which led to their system inserting
    delays if a response was ready too soon. :-p

    Back in my University Computer Center days around 1972, I was hacking
    a local email feature into the EXEC-8 editor, and decided that less
    variation in response times would be a good thing, so I added a 1/2
    second wait before each editor command prompt. I got in serious trouble
    for this, because the process scheduler assumed that 500 ms was "a long time", and a program that was going to wait for a long time should be swapped out of memory.

    You should check whether Microsoft has tried to patent fake bullshit
    system response pauses. This is clear prior art!
    --
    TXR Programming Language: http://nongnu.org/txr
    Cygnal: Cygwin Native Application Library: http://kylheku.com/cygnal
    Mastodon: @Kazinator@mstdn.ca
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Kaz Kylheku@046-301-5902@kylheku.com to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Fri Nov 6 21:31:52 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-06, Charlie Gibbs <cgibbs@kltpzyxm.invalid> wrote:
    On 2026-11-06, Lawrence DrCOOliveiro <ldo@nz.invalid> wrote:

    And hererCOs another tale: this was probably from the very earliest days
    of computers, but the story I remember reading involves users entering
    a query, and how long it took to get back a response. Apparently, if
    the response came back *too* quickly, the users felt that the computer
    somehow wasnrCOt treating their question with the importance it
    deserved.

    I wrote a program that searches a large collection of files for
    various data. I noticed the same thing: it ran so fast that
    people weren't taking it seriously. It didn't help that the
    search parameter screen was too easy to use, either.

    That's much the same like people not taking a pair of headphones
    seriously if they are light; so Beats famously added hidden metal
    weights to them to create a more premium feel.
    --
    TXR Programming Language: http://nongnu.org/txr
    Cygnal: Cygwin Native Application Library: http://kylheku.com/cygnal
    Mastodon: @Kazinator@mstdn.ca
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Johann 'Myrkraverk' Oskarsson@johann@myrkraverk.invalid to comp.os.linux.misc,alt.folklore.computers on Sat Nov 7 06:13:40 2026
    From Newsgroup: comp.os.linux.misc

    On 11/7/2026 4:53 AM, Lawrence DrCOOliveiro wrote:
    (This is all from memory, itrCOs been a long while since I wrote any 68000-family assembler.)

    I have /68000 Family Assembly Language/ (1994) by Alan Clements on my
    shelf, but I'm not in a mood to second guess your work.


    Happy 68K assembly!
    --
    Johann | email: invalid -> com | http://www.myrkraverk.com/blog/
    I'm not from the Internet, I just work there. | via Easynews.com https://bsky.app/profile/myrkraverk.bsky.social | for ( ;; ) _:;
    Federated at https://fed.brid.gy/bsky/myrkraverk.bsky.social
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Johann 'Myrkraverk' Oskarsson@johann@myrkraverk.invalid to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Sat Nov 7 06:25:12 2026
    From Newsgroup: comp.os.linux.misc

    On 11/7/2026 4:51 AM, Charlie Gibbs wrote:
    On 2026-11-06, Lawrence DrCOOliveiro <ldo@nz.invalid> wrote:

    And hererCOs another tale: this was probably from the very earliest days
    of computers, but the story I remember reading involves users entering
    a query, and how long it took to get back a response. Apparently, if
    the response came back *too* quickly, the users felt that the computer
    somehow wasnrCOt treating their question with the importance it
    deserved.

    I wrote a program that searches a large collection of files for
    various data. I noticed the same thing: it ran so fast that
    people weren't taking it seriously. It didn't help that the
    search parameter screen was too easy to use, either.

    I guess that in this modern age, where complexity is equated
    with sophistication, any sign of simplicity or efficiency
    is suspect.


    I think it's more /instinctual/ than that. Humans, who dismiss sugges-
    tions without thought, normally do so as a status symbol: they are more important than you.

    It isn't about the idea or suggestion being stupid, because then there
    would be a followup explanation.

    Follow those experiences up with people who do shoddy work fast, and if
    the computer is fast too, it /didn't take the time to do things proper-
    ly/.

    The machine, as a servant, should never /feel/ more important than the
    user. Which unfortunately, seems to be more and more real in the age of
    "A.I." as business owners tend to prefer people who can crank out
    prompts rather than do the work /by hand/. The /machine/ is now more
    important than the employee who wields it. Anyway, that's a rant we
    don't need in alt.folklore.computers.


    Best wishes, and happy search algorithms!
    --
    Johann | email: invalid -> com | http://www.myrkraverk.com/blog/
    I'm not from the Internet, I just work there. | via Easynews.com https://bsky.app/profile/myrkraverk.bsky.social | for ( ;; ) _:;
    Federated at https://fed.brid.gy/bsky/myrkraverk.bsky.social
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Robert Riches@spamtrap42@jacob21819.net to comp.os.linux.misc on Fri Nov 6 22:29:32 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-06, Charlie Gibbs <cgibbs@kltpzyxm.invalid> wrote:
    On 2026-11-06, rbowman <bowman@montana.com> wrote:

    On Thu, 5 Nov 2026 13:19:40 -0800, John Ames wrote:

    On Thu, 5 Nov 2026 15:50:22 -0500 c186282 <c186282@nnada.net> wrote:

    Alas I don't remember any 6502s that were clocked much over 1mhz -
    while the Z-80s could run at four. MIGHT be that the Z-80 required more >>>> cycles per instruction though, reducing that advantage - somebody here >>>> knows ..... I think the Z-80 needed twice as many clocks.

    It's a perennial debate between the Commodore and Spectrum partisans,
    yep.
    Z80 systems frequently ran at higher clock rates, but equivalent
    instructions were around twice as long (IIRC part of it was that its ALU >>> was 4-bit internally.) On the other hand, there's less "register
    pressure"
    on the Z80. On the other other hand, most of the registers are just
    dance partners for the accumulator and/or pointers for load/ store, etc. >>> etc., da capo ad nauseam.

    Quite a few Z80 systems ran at 3.57954 MHz. Love those cheap NTSC color
    burst crystals.

    Amigas ran at 7.16 MHz. They showed up at a lot of cable TV
    stations generating those time/date/weather screens. And the
    Video Toaster did cutting-edge video effects. Todd Rundgren
    created his "Change Myself" video using a bank of ten
    Toaster-equipped Amigas.

    IIRC, the CGI in the TV series "Babylon 5" was created by Amigas.
    I'm not aware of how many or in what configuration.
    --
    Robert Riches
    spamtrap42@jacob21819.net
    (Yes, that is one of my email addresses.)
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Carlos E. R.@robin_listas@es.invalid to comp.os.linux.misc,alt.folklore.computers on Sat Nov 7 01:12:48 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-06 21:53, Lawrence DrCOOliveiro wrote:
    On Fri, 6 Nov 2026 14:30:35 +0100, Carlos E. R. wrote:

    On 2026-11-06 00:17, Lawrence DrCOOliveiro wrote:

    HererCOs one trivia question: the decrement-and-branch instruction
    would fall out of the loop when the count reached -1, not 0 as the
    equivalent instruction in nearly every other computer architecture
    did. Why?

    Mmm. I don't remember. A flag went up?

    It was quite a clever feature, which I think ended up widely
    misunderstood. Otherwise more architectures would have copied it.

    The reason is, so that if the loop count was zero, you could just
    fall out of the loop automatically, without having to do an explicit
    test for that case.

    Example skeleton loop: (remember, the loop count was only 16 bits):

    move.w -2loop-count-+, dN ; dN = some convenient d-register
    bra loop_bottom
    loop_top:
    ... body of loop ...
    loop_bottom:
    dbra dN, loop_top
    ... here on loop end ...

    Notice this will loop exactly -2loop-count-+ times, even if -2loop-count-+
    is zero.

    If you want a 32-bit loop count, it only took a few more instructions:

    move.l -2loop-count-+, dN
    bra loop_bottom
    loop_top1:
    swap dN
    loop_top2:
    ... body of loop ...
    loop_bottom:
    dbra dN, loop_top2
    swap dN
    dbra dN, loop_top1
    ... here on loop end ...

    (This is all from memory, itrCOs been a long while since I wrote any 68000-family assembler.)

    Last time I touched it, was 1992. :-(
    --
    Cheers,
    Carlos E.R.
    ESEfc-Efc+, EUEfc-Efc|.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Charlie Gibbs@cgibbs@kltpzyxm.invalid to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Sat Nov 7 01:22:40 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-06, Kaz Kylheku <046-301-5902@kylheku.com> wrote:

    On 2026-11-06, Charlie Gibbs <cgibbs@kltpzyxm.invalid> wrote:

    On 2026-11-06, Lawrence DrCOOliveiro <ldo@nz.invalid> wrote:

    And hererCOs another tale: this was probably from the very earliest days >>> of computers, but the story I remember reading involves users entering
    a query, and how long it took to get back a response. Apparently, if
    the response came back *too* quickly, the users felt that the computer
    somehow wasnrCOt treating their question with the importance it
    deserved.

    I wrote a program that searches a large collection of files for
    various data. I noticed the same thing: it ran so fast that
    people weren't taking it seriously. It didn't help that the
    search parameter screen was too easy to use, either.

    That's much the same like people not taking a pair of headphones
    seriously if they are light; so Beats famously added hidden metal
    weights to them to create a more premium feel.

    I've heard that car manufacturers put padding into the doors
    to give a satisfying thump rather than a clank.
    --
    /~\ Charlie Gibbs | In this world there are
    \ / <cgibbs@kltpzyxm.invalid> | two kinds of people:
    X I'm really at ac.dekanfrus | 1. Those who can extrapolate
    / \ if you read it the right way. | from incomplete data.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From rbowman@bowman@montana.com to comp.os.linux.misc on Sat Nov 7 02:32:27 2026
    From Newsgroup: comp.os.linux.misc

    On Fri, 06 Nov 2026 20:51:49 GMT, Charlie Gibbs wrote:

    On 2026-11-06, rbowman <bowman@montana.com> wrote:

    On Thu, 5 Nov 2026 18:43:16 -0800, Lars Poulsen wrote:

    I got in serious trouble for this,
    because the process scheduler assumed that 500 ms was "a long time",
    and a program that was going to wait for a long time should be swapped
    out of memory.

    It is a long time For example if you click on a button and there is no
    perceptible change for 500 ms the users will complain about the program
    being slow.

    And then he'll click on the button again, and wind up with some
    undesired double-click action when the program does get around to
    dealing with it.

    I've seen some web pages that explicitly say "Don't do that!" It isn't the button click itself but they know processing the request will take longer
    than a consumer oriented towards instant gratification will expect.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Janis Papanagnou@janis_papanagnou+ng@hotmail.com to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Sat Nov 7 05:00:51 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-07 02:22, Charlie Gibbs wrote:

    I've heard that car manufacturers put padding into the doors
    to give a satisfying thump rather than a clank.

    Unless the various sounds in cars are created artificially!

    (Not joking. - Inspect the automotive business job offers,
    specifically sound designers and sound engineers, and also
    inspect the information on the Net about where artificial
    sound is actually synthesized and applied to, including the
    doors.)

    Janis

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From c186282@c186282@nnada.net to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Fri Nov 6 23:09:53 2026
    From Newsgroup: comp.os.linux.misc

    On 11/6/26 01:35, Thomas Koenig wrote:
    The Natural Philosopher <tnp@invalid.invalid> schrieb:

    IBM Macro assembler is practically a high level language, given the
    scale of macro usage in legacy systems

    High-level languages abstract away the underlying machine, which
    I presume the Macro assembler does not.

    Plus, does it do register allocation?

    (OK, then Wirth's PL/360 creeps in, which does not do so either,
    but then I would argue that that is a glorified assembler - it also
    does not do operator precenence).

    I will agree that MASM isn't a good equiv to
    any higher-level language. Used it a fair bit
    in the bad old days, but never to its max cap.
    Served my particular needs however, back when
    ASM was the only way to get tight code with
    lowest-level access. 'C' eventually assumed
    that role for the most part, but not overnight.

    ASM was also *the* way for the older micro-controllers
    for a long time. However some of the optimized 'C'
    compilers can now rival ASM for speed and size and
    you don't have to hand-code bit-banging serial either :-)

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Charlie Gibbs@cgibbs@kltpzyxm.invalid to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Sat Nov 7 04:12:42 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-07, Janis Papanagnou <janis_papanagnou+ng@hotmail.com> wrote:

    On 2026-11-07 02:22, Charlie Gibbs wrote:

    I've heard that car manufacturers put padding into the doors
    to give a satisfying thump rather than a clank.

    Unless the various sounds in cars are created artificially!

    (Not joking. - Inspect the automotive business job offers,
    specifically sound designers and sound engineers, and also
    inspect the information on the Net about where artificial
    sound is actually synthesized and applied to, including the
    doors.)

    I did read recently that they have staffers charged with
    designing the weird noises electric cars make at low speeds.
    --
    /~\ Charlie Gibbs | In this world there are
    \ / <cgibbs@kltpzyxm.invalid> | two kinds of people:
    X I'm really at ac.dekanfrus | 1. Those who can extrapolate
    / \ if you read it the right way. | from incomplete data.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Janis Papanagnou@janis_papanagnou+ng@hotmail.com to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Sat Nov 7 05:39:08 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-07 05:12, Charlie Gibbs wrote:
    On 2026-11-07, Janis Papanagnou <janis_papanagnou+ng@hotmail.com> wrote:

    On 2026-11-07 02:22, Charlie Gibbs wrote:

    I've heard that car manufacturers put padding into the doors
    to give a satisfying thump rather than a clank.

    Unless the various sounds in cars are created artificially!

    (Not joking. - Inspect the automotive business job offers,
    specifically sound designers and sound engineers, and also
    inspect the information on the Net about where artificial
    sound is actually synthesized and applied to, including the
    doors.)

    I did read recently that they have staffers charged with
    designing the weird noises electric cars make at low speeds.

    Yes, that is one application case. Actually its a requirement
    that is even demanded by regulations hereabouts!

    But then there's also other things, like the mentioned door
    sounds, the voluminous sound of the engine, noise reduction
    (adding noise here and reducing it there; somewhat crazy!),
    feedback sounds (with Star Trek as general paragon :-), and
    some gimmicks.

    Janis

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Robert Riches@spamtrap42@jacob21819.net to comp.os.linux.misc on Sat Nov 7 04:50:19 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-07, Janis Papanagnou <janis_papanagnou+ng@hotmail.com> wrote:
    On 2026-11-07 05:12, Charlie Gibbs wrote:
    On 2026-11-07, Janis Papanagnou <janis_papanagnou+ng@hotmail.com> wrote:

    On 2026-11-07 02:22, Charlie Gibbs wrote:

    I've heard that car manufacturers put padding into the doors
    to give a satisfying thump rather than a clank.

    Unless the various sounds in cars are created artificially!

    (Not joking. - Inspect the automotive business job offers,
    specifically sound designers and sound engineers, and also
    inspect the information on the Net about where artificial
    sound is actually synthesized and applied to, including the
    doors.)

    I did read recently that they have staffers charged with
    designing the weird noises electric cars make at low speeds.

    Yes, that is one application case. Actually its a requirement
    that is even demanded by regulations hereabouts!

    But then there's also other things, like the mentioned door
    sounds, the voluminous sound of the engine, noise reduction
    (adding noise here and reducing it there; somewhat crazy!),
    feedback sounds (with Star Trek as general paragon :-), and
    some gimmicks.

    Janis

    I heard an apocryphal story several years or maybe a couple of
    decades ago that one car company hired the Yamaha audio company
    to design the exhaust system of one of their vehicles, and the
    result sounded EXTREMELY good.
    --
    Robert Riches
    spamtrap42@jacob21819.net
    (Yes, that is one of my email addresses.)
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From c186282@c186282@nnada.net to comp.os.linux.misc on Sat Nov 7 00:35:36 2026
    From Newsgroup: comp.os.linux.misc

    On 11/6/26 23:50, Robert Riches wrote:
    On 2026-11-07, Janis Papanagnou <janis_papanagnou+ng@hotmail.com> wrote:
    On 2026-11-07 05:12, Charlie Gibbs wrote:
    On 2026-11-07, Janis Papanagnou <janis_papanagnou+ng@hotmail.com> wrote: >>>
    On 2026-11-07 02:22, Charlie Gibbs wrote:

    I've heard that car manufacturers put padding into the doors
    to give a satisfying thump rather than a clank.

    Unless the various sounds in cars are created artificially!

    (Not joking. - Inspect the automotive business job offers,
    specifically sound designers and sound engineers, and also
    inspect the information on the Net about where artificial
    sound is actually synthesized and applied to, including the
    doors.)

    I did read recently that they have staffers charged with
    designing the weird noises electric cars make at low speeds.

    Yes, that is one application case. Actually its a requirement
    that is even demanded by regulations hereabouts!

    But then there's also other things, like the mentioned door
    sounds, the voluminous sound of the engine, noise reduction
    (adding noise here and reducing it there; somewhat crazy!),
    feedback sounds (with Star Trek as general paragon :-), and
    some gimmicks.

    Janis

    I heard an apocryphal story several years or maybe a couple of
    decades ago that one car company hired the Yamaha audio company
    to design the exhaust system of one of their vehicles, and the
    result sounded EXTREMELY good.

    MAY be true.

    IC engines CAN sound REALLY nice - if properly 'tuned'.

    Electrics ... TOO quiet, people use engine sounds as
    part of their instant eval to cross a street. No sound
    equals no safety. The blind should SUE.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From The Natural Philosopher@tnp@invalid.invalid to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Sat Nov 7 11:32:29 2026
    From Newsgroup: comp.os.linux.misc

    On 06/11/2026 15:45, Scott Lurndal wrote:
    The Natural Philosopher <tnp@invalid.invalid> writes:
    On 04/11/2026 14:54, Peter Flass wrote:
    On 11/3/26 15:15, Pancho wrote:

    If you want heavy usage, I would guess bank risk calcs, weather calcs, >>>> lots of science applications, running on compute grids, (or whatever
    the modern word is), blow this mainframe high i/o away.

    I think back in the day the quants preferred Sun.

    Not for transactuon processing

    You clearly don't know what a "quant" is, in this context.

    I do.

    More than you, probably
    --
    If I had all the money I've spent on drink...
    ..I'd spend it on drink.

    Sir Henry (at Rawlinson's End)

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From The Natural Philosopher@tnp@invalid.invalid to comp.os.linux.misc on Sat Nov 7 11:34:54 2026
    From Newsgroup: comp.os.linux.misc

    On 06/11/2026 21:10, rbowman wrote:
    On Fri, 6 Nov 2026 08:51:28 +0100, Janis Papanagnou wrote:

    There's actually standards and best practices for that investigated in
    surveys; accepted upper limits for interactive activities typically vary
    in about the range of 100-300 ms. Anything above that is usually
    perceived as extremely disturbing a continuous activity. (These limits
    had been our landmarks, what we needed to achieve and guarantee in our
    client software development 30+ years ago. - Of course we strived for
    minimum latency.)

    Many people don't have a good feeling for time units under one second although they constantly are doing performing actions in milliseconds. A
    shot timer is used in pistol competitions. It's nothing complicated, a microphone and circuitry capable of measuring time. You could cook one up with an Arduino. A competent marksman, not a first rank shooter, is
    placing shots on target in the 200 ms range. In that time period you are aware of the ejected brass flying off, the front sight falling back into alignment with the rear sight and target, and squeezing the trigger.

    Catching the egg you accidentally knocked off the counter probably takes a bit longer since you weren't anticipating it but most people, other than those sitting at a traffic light when it turns green, react in well under
    a second.

    the standard rate used by formula one to determine if a driver 'jumped
    the lights' is 200ms

    Less than that raises eyebrows...
    --
    rCLThe urge to save humanity is almost always only a false face for the
    urge to rule it.rCY
    rCo H. L. Mencken

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From The Natural Philosopher@tnp@invalid.invalid to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Sat Nov 7 11:37:39 2026
    From Newsgroup: comp.os.linux.misc

    On 07/11/2026 04:12, Charlie Gibbs wrote:
    I did read recently that they have staffers charged with
    designing the weird noises electric cars make at low speeds.

    My StupidPhone yelps like a puppy that has just been stepped on, on
    receipt of a text message.

    People look for the dog...
    --
    Karl Marx said religion is the opium of the people.
    But Marxism is the crack cocaine.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From The Natural Philosopher@tnp@invalid.invalid to comp.os.linux.misc on Sat Nov 7 11:39:47 2026
    From Newsgroup: comp.os.linux.misc

    On 07/11/2026 04:50, Robert Riches wrote:
    I heard an apocryphal story several years or maybe a couple of
    decades ago that one car company hired the Yamaha audio company
    to design the exhaust system of one of their vehicles, and the
    result sounded EXTREMELY good.

    I like the formula one engine programmed to play 'God Save the Queen' https://www.youtube.com/watch?v=XRXwWbo_mX0
    --
    Karl Marx said religion is the opium of the people.
    But Marxism is the crack cocaine.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From The Natural Philosopher@tnp@invalid.invalid to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Sat Nov 7 11:45:42 2026
    From Newsgroup: comp.os.linux.misc

    On 07/11/2026 04:09, c186282 wrote:
    ASM was also *the* way for the older micro-controllers
    -a for a long time. However some of the optimized 'C'
    -a compilers can now rival ASM for speed and size and
    -a you don't have to hand-code bit-banging serial either

    At some point I idly compiled some gcc C to assembler . I spent a
    pointless half hour trying to work out if I could have done it better.

    I couldn't..

    In terms of modern micros I wouldn't even bother unless I couldn't do it
    in 'C'.
    --
    rCLit should be clear by now to everyone that activist environmentalism
    (or environmental activism) is becoming a general ideology about humans,
    about their freedom, about the relationship between the individual and
    the state, and about the manipulation of people under the guise of a
    'noble' idea. It is not an honest pursuit of 'sustainable development,'
    a matter of elementary environmental protection, or a search for
    rational mechanisms designed to achieve a healthy environment. Yet
    things do occur that make you shake your head and remind yourself that
    you live neither in Joseph StalinrCOs Communist era, nor in the Orwellian utopia of 1984.rCY

    Vaclav Klaus

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Peter Flass@Peter@Iron-Spring.com to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Sat Nov 7 07:37:39 2026
    From Newsgroup: comp.os.linux.misc

    On 11/7/26 04:45, The Natural Philosopher wrote:
    On 07/11/2026 04:09, c186282 wrote:
    ASM was also *the* way for the older micro-controllers
    -a-a for a long time. However some of the optimized 'C'
    -a-a compilers can now rival ASM for speed and size and
    -a-a you don't have to hand-code bit-banging serial either

    At some point I idly compiled some gcc C to assembler . I spent a
    pointless half hour trying to work out if I could have done it better.

    I couldn't..

    In terms of modern micros I wouldn't even bother unless I couldn't do it
    in 'C'.


    But then you add all the cruft that GCC drags in and you're back where
    you started (size-wise, at least). When PCs were new I used a compiler
    called "Aztec C" that had almost no overhead in excess libraries,
    start-up code, etc., unless you asked for them.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Johann 'Myrkraverk' Oskarsson@johann@myrkraverk.invalid to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Sat Nov 7 22:49:09 2026
    From Newsgroup: comp.os.linux.misc

    On 11/7/2026 10:37 PM, Peter Flass wrote:
    On 11/7/26 04:45, The Natural Philosopher wrote:
    On 07/11/2026 04:09, c186282 wrote:
    ASM was also *the* way for the older micro-controllers
    -a-a for a long time. However some of the optimized 'C'
    -a-a compilers can now rival ASM for speed and size and
    -a-a you don't have to hand-code bit-banging serial either

    At some point I idly compiled some gcc C to assembler . I spent a
    pointless half hour trying to work out if I could have done it better.

    I couldn't..

    In terms of modern micros I wouldn't even bother unless I couldn't do
    it in 'C'.


    But then you add all the cruft that GCC drags in and you're back where
    you started (size-wise, at least). When PCs were new I used a compiler called "Aztec C" that had almost no overhead in excess libraries, start-
    up code, etc., unless you asked for them.

    And nonconsensual AVX code. I've heard some CPUs have a lower maximum
    Hz when AVX is enabled. And I have never heard of a benchmark that mea-
    sures the cutoff when doing things the old fashioned way, say REP MOVSB
    is faster or slower than the AVX method(s), due to this difference.

    So, why not tell me this cutoff? Are the GCC devs afraid of something? Shouldn't they be proud of their optimizations?

    Or was I lied to, about modern CPUs stepping down when AVX is in play?
    --
    Johann | email: invalid -> com | http://www.myrkraverk.com/blog/
    I'm not from the Internet, I just work there. | via Easynews.com https://bsky.app/profile/myrkraverk.bsky.social | for ( ;; ) _:;
    Federated at https://fed.brid.gy/bsky/myrkraverk.bsky.social
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Thomas Koenig@tkoenig@netcologne.de to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Sat Nov 7 14:54:06 2026
    From Newsgroup: comp.os.linux.misc

    Johann 'Myrkraverk' Oskarsson <johann@myrkraverk.invalid> schrieb:
    On 11/7/2026 10:37 PM, Peter Flass wrote:
    On 11/7/26 04:45, The Natural Philosopher wrote:
    On 07/11/2026 04:09, c186282 wrote:
    ASM was also *the* way for the older micro-controllers
    -a-a for a long time. However some of the optimized 'C'
    -a-a compilers can now rival ASM for speed and size and
    -a-a you don't have to hand-code bit-banging serial either

    At some point I idly compiled some gcc C to assembler . I spent a
    pointless half hour trying to work out if I could have done it better.

    I couldn't..

    In terms of modern micros I wouldn't even bother unless I couldn't do
    it in 'C'.


    But then you add all the cruft that GCC drags in and you're back where
    you started (size-wise, at least). When PCs were new I used a compiler
    called "Aztec C" that had almost no overhead in excess libraries, start-
    up code, etc., unless you asked for them.

    And nonconsensual AVX code.

    -fno-tree-optimize is your friend.

    I've heard some CPUs have a lower maximum
    Hz when AVX is enabled.

    AVX512, yes.
    --
    This USENET posting was made without artificial intelligence,
    artificial impertinence, artificial arrogance, artificial stupidity,
    artificial flavorings or artificial colorants.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Richard Kettlewell@invalid@invalid.invalid to comp.os.linux.misc on Sat Nov 7 16:04:46 2026
    From Newsgroup: comp.os.linux.misc

    Robert Riches <spamtrap42@jacob21819.net> writes:
    On 2026-11-06, Peter Flass <Peter@Iron-Spring.com> wrote:
    On 11/5/26 16:17, Lawrence DrCOOliveiro wrote:
    On Thu, 5 Nov 2026 22:24:10 +0100, Carlos E. R. wrote:

    The 68000 had a beautiful assembler. Or rather, a beautiful
    instruction set.

    Definitely inspired by the PDP-11. Pity it was big-endian. Also, the
    division into rCLaddressrCY versus rCLdatarCY registers did seem to be a step
    backwards.

    HererCOs one trivia question: the decrement-and-branch instruction would >>> fall out of the loop when the count reached -1, not 0 as the
    equivalent instruction in nearly every other computer architecture
    did. Why?

    Big-endian is the only true way.

    It is if you insist on string comparison and integer comparison
    being identical or being able to read digits left to right and
    put them in increasing memory addresses.

    General-purpose multiple precision integers commonly use a little-endian representation (i.e. least significant word first, but with native
    endianness for the individual words).

    ThererCOs an obvious performance justification for this: when the value of
    an integer grows and it needs more words, it avoids the need to shuffle everything up by one. (If the maximum size is known in advance, and not unreasonably large, then this justification goes away, of course.)

    Readability is a pretty good argument for big-endian representations, of course, if you spend a lof of time reading hex dumps. The original sin
    there was the failure to flip the conventional ordering of digits when
    Arabic numerals were adopted from the right-to-left Arabic script into left-to-right European scripts.
    --
    https://www.greenend.org.uk/rjk/
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Christian Weisgerber@naddy@mips.inka.de to comp.os.linux.misc,comp.unix.shell on Sat Nov 7 16:07:26 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-01, Carlos E. R. <robin_listas@es.invalid> wrote:

    There is the tool "pinfo", it can highlight somethings like the name of another man page, and jump to it.

    Still, no table of content. No jumping from one section while browsing
    the man page.

    "A more modern man page viewer for our terminals"
    https://github.com/plp13/qman

    Qman [i]s a modern, full-featured manual page viewer featuring
    hyperlinks, web browser like navigation, a table of contents for
    each page, incremental search, on-line help, and more. It also
    strives to be fast and tiny, so that it can be used everywhere. For
    this reason, it's been written in plain C and has only minimal
    dependencies.

    (I haven't tried it myself.)
    --
    Christian "naddy" Weisgerber naddy@mips.inka.de
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From c186282@c186282@nnada.net to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Sat Nov 7 13:31:06 2026
    From Newsgroup: comp.os.linux.misc

    On 11/7/26 09:37, Peter Flass wrote:
    On 11/7/26 04:45, The Natural Philosopher wrote:
    On 07/11/2026 04:09, c186282 wrote:
    ASM was also *the* way for the older micro-controllers
    -a-a for a long time. However some of the optimized 'C'
    -a-a compilers can now rival ASM for speed and size and
    -a-a you don't have to hand-code bit-banging serial either

    At some point I idly compiled some gcc C to assembler . I spent a
    pointless half hour trying to work out if I could have done it better.

    I couldn't..

    In terms of modern micros I wouldn't even bother unless I couldn't do
    it in 'C'.


    But then you add all the cruft that GCC drags in and you're back where
    you started (size-wise, at least). When PCs were new I used a compiler called "Aztec C" that had almost no overhead in excess libraries, start-
    up code, etc., unless you asked for them.

    Used Aztec ... indeed HAVE the compiler and EXTENSIVE
    docs stashed away somewhere. Will have to make a new
    CP/M-86 VM for it first. Not sure you'd get best effect
    on a simulated Z80. Have seen some 'kit' Z80 boards.

    There's also a DOS version - but it's apparently meant
    to run under XP.

    The docs are IMPRESSIVE - endless bit-level and flag
    stuff - exactly what you needed to know if trying to
    shoehorn something into a small system.

    A few versions of Turbo Pascal allowed code overlays
    in a fairly civil fashion, another way to squeeze a
    big pgm into a small machine. But, better have a HDD,
    pulling overlays from a floppy - you don't wanna
    go there :-) Did a couple of large pgms where overlays
    were needed.

    Anyway, if interested in Aztec, try :

    https://www.aztecmuseum.ca/compilers.htm



    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Johann 'Myrkraverk' Oskarsson@johann@myrkraverk.invalid to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Sun Nov 8 02:56:49 2026
    From Newsgroup: comp.os.linux.misc

    On 11/8/2026 2:31 AM, c186282 wrote:
    On 11/7/26 09:37, Peter Flass wrote:
    On 11/7/26 04:45, The Natural Philosopher wrote:
    On 07/11/2026 04:09, c186282 wrote:
    ASM was also *the* way for the older micro-controllers
    -a-a for a long time. However some of the optimized 'C'
    -a-a compilers can now rival ASM for speed and size and
    -a-a you don't have to hand-code bit-banging serial either

    At some point I idly compiled some gcc C to assembler . I spent a
    pointless half hour trying to work out if I could have done it better.

    I couldn't..

    In terms of modern micros I wouldn't even bother unless I couldn't do
    it in 'C'.


    But then you add all the cruft that GCC drags in and you're back where
    you started (size-wise, at least). When PCs were new I used a compiler
    called "Aztec C" that had almost no overhead in excess libraries,
    start- up code, etc., unless you asked for them.

    -a Used Aztec ... indeed HAVE the compiler and EXTENSIVE
    -a docs stashed away somewhere. Will have to make a new
    -a CP/M-86 VM for it first. Not sure you'd get best effect
    -a on a simulated Z80. Have seen some 'kit' Z80 boards.

    -a There's also a DOS version - but it's apparently meant
    -a to run under XP.

    -a The docs are IMPRESSIVE - endless bit-level and flag
    -a stuff - exactly what you needed to know if trying to
    -a shoehorn something into a small system.

    -a A few versions of Turbo Pascal allowed code overlays
    -a in a fairly civil fashion, another way to squeeze a
    -a big pgm into a small machine. But, better have a HDD,
    -a pulling overlays from a floppy - you don't wanna
    -a go there :-) Did a couple of large pgms where overlays
    -a were needed.

    -a Anyway, if interested in Aztec, try :

    https://www.aztecmuseum.ca/compilers.htm

    Thank you. It's mildly unclear if the compiler will run under pure DOS,
    and this XP package is merely a convenience -- alas I was not in the
    mood to find out.

    However, I am perusing the manual library, to see if I find something interesting.
    --
    Johann | email: invalid -> com | http://www.myrkraverk.com/blog/
    I'm not from the Internet, I just work there. | via Easynews.com https://bsky.app/profile/myrkraverk.bsky.social | for ( ;; ) _:;
    Federated at https://fed.brid.gy/bsky/myrkraverk.bsky.social
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lawrence =?iso-8859-13?q?D=FFOliveiro?=@ldo@nz.invalid to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Sat Nov 7 20:03:01 2026
    From Newsgroup: comp.os.linux.misc

    On Sat, 7 Nov 2026 07:37:39 -0700, Peter Flass wrote:

    But then you add all the cruft that GCC drags in ...

    You can build programs with GCC that donrCOt use glibc.

    Back when I got my first 64-bit machine, I tried to build the smallest
    program I could: it took a number as a command-line argument, and
    slept for that number of seconds. This gave me time to do a ps listing
    to see how much memory it was using. I built it against one of the
    alternative cut-down libc alternatives -- maybe musl or uclibc.

    ps reported that it was using just 20kB of memory.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From The Natural Philosopher@tnp@invalid.invalid to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Sat Nov 7 21:53:26 2026
    From Newsgroup: comp.os.linux.misc

    On 07/11/2026 14:37, Peter Flass wrote:
    On 11/7/26 04:45, The Natural Philosopher wrote:
    On 07/11/2026 04:09, c186282 wrote:
    ASM was also *the* way for the older micro-controllers
    -a-a for a long time. However some of the optimized 'C'
    -a-a compilers can now rival ASM for speed and size and
    -a-a you don't have to hand-code bit-banging serial either

    At some point I idly compiled some gcc C to assembler . I spent a
    pointless half hour trying to work out if I could have done it better.

    I couldn't..

    In terms of modern micros I wouldn't even bother unless I couldn't do
    it in 'C'.


    But then you add all the cruft that GCC drags in and you're back where
    you started (size-wise, at least).
    Hardly.


    When PCs were new I used a compiler
    called "Aztec C" that had almost no overhead in excess libraries,
    start-up code, etc., unless you asked for them.

    Well in general you DO ask for them
    --
    Climate Change: Socialism wearing a lab coat.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Robert Riches@spamtrap42@jacob21819.net to comp.os.linux.misc on Sat Nov 7 22:41:31 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-07, c186282 <c186282@nnada.net> wrote:
    On 11/6/26 23:50, Robert Riches wrote:
    On 2026-11-07, Janis Papanagnou <janis_papanagnou+ng@hotmail.com> wrote:
    On 2026-11-07 05:12, Charlie Gibbs wrote:
    On 2026-11-07, Janis Papanagnou <janis_papanagnou+ng@hotmail.com> wrote: >>>>
    On 2026-11-07 02:22, Charlie Gibbs wrote:

    I've heard that car manufacturers put padding into the doors
    to give a satisfying thump rather than a clank.

    Unless the various sounds in cars are created artificially!

    (Not joking. - Inspect the automotive business job offers,
    specifically sound designers and sound engineers, and also
    inspect the information on the Net about where artificial
    sound is actually synthesized and applied to, including the
    doors.)

    I did read recently that they have staffers charged with
    designing the weird noises electric cars make at low speeds.

    Yes, that is one application case. Actually its a requirement
    that is even demanded by regulations hereabouts!

    But then there's also other things, like the mentioned door
    sounds, the voluminous sound of the engine, noise reduction
    (adding noise here and reducing it there; somewhat crazy!),
    feedback sounds (with Star Trek as general paragon :-), and
    some gimmicks.

    Janis

    I heard an apocryphal story several years or maybe a couple of
    decades ago that one car company hired the Yamaha audio company
    to design the exhaust system of one of their vehicles, and the
    result sounded EXTREMELY good.

    MAY be true.

    IC engines CAN sound REALLY nice - if properly 'tuned'.

    Electrics ... TOO quiet, people use engine sounds as
    part of their instant eval to cross a street. No sound
    equals no safety. The blind should SUE.

    At least in the US, I understand there is a government-imposed
    requirement for electrics to make noise. When I first heard of
    that requirement, I thought about selling a retrofit kit that
    would contain a pack of playing cards and a few clothespins. :-)
    I have described the artificial noise they create as about
    halfway between a flying saucer and a garbage disposal.
    --
    Robert Riches
    spamtrap42@jacob21819.net
    (Yes, that is one of my email addresses.)
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Robert Riches@spamtrap42@jacob21819.net to comp.os.linux.misc on Sat Nov 7 22:48:01 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-07, The Natural Philosopher <tnp@invalid.invalid> wrote:
    On 07/11/2026 04:50, Robert Riches wrote:
    I heard an apocryphal story several years or maybe a couple of
    decades ago that one car company hired the Yamaha audio company
    to design the exhaust system of one of their vehicles, and the
    result sounded EXTREMELY good.

    I like the formula one engine programmed to play 'God Save the Queen' https://www.youtube.com/watch?v=XRXwWbo_mX0

    I recognized the tune, but I know it by a different title and
    different text. :-)

    In the early 1980s, for the Tektronix 4050 series of computers,
    Tek had written a program (undoubtedly in assembly rather than
    BASIC) that would play limited polyphonic music on a 1-bit
    speaker. Normally, the 1-bit speaker was limited to a single
    tone, but they apparently figured out a way to combine two tones.
    I suspect the two tones had to have a simpler (2/3 or maybe 4/5
    or so) frequency ratio between the tones. The 4052 and 4054 had
    a 20MHz bit-slice CPU, so it might have been able to do PWM of
    multiple tones, but the 4051 was a 6800 at around ~1MHz. If I
    remember correctly, the tune they did was Greensleeves.

    Perchance, is anyone acquainted with how Tek did polyphonic sound
    from a 1-bit speaker with a 1MHz CPU?
    --
    Robert Riches
    spamtrap42@jacob21819.net
    (Yes, that is one of my email addresses.)
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From c186282@c186282@nnada.net to comp.os.linux.misc on Sat Nov 7 19:56:37 2026
    From Newsgroup: comp.os.linux.misc

    On 11/7/26 17:41, Robert Riches wrote:
    On 2026-11-07, c186282 <c186282@nnada.net> wrote:
    On 11/6/26 23:50, Robert Riches wrote:
    On 2026-11-07, Janis Papanagnou <janis_papanagnou+ng@hotmail.com> wrote: >>>> On 2026-11-07 05:12, Charlie Gibbs wrote:
    On 2026-11-07, Janis Papanagnou <janis_papanagnou+ng@hotmail.com> wrote: >>>>>
    On 2026-11-07 02:22, Charlie Gibbs wrote:

    I've heard that car manufacturers put padding into the doors
    to give a satisfying thump rather than a clank.

    Unless the various sounds in cars are created artificially!

    (Not joking. - Inspect the automotive business job offers,
    specifically sound designers and sound engineers, and also
    inspect the information on the Net about where artificial
    sound is actually synthesized and applied to, including the
    doors.)

    I did read recently that they have staffers charged with
    designing the weird noises electric cars make at low speeds.

    Yes, that is one application case. Actually its a requirement
    that is even demanded by regulations hereabouts!

    But then there's also other things, like the mentioned door
    sounds, the voluminous sound of the engine, noise reduction
    (adding noise here and reducing it there; somewhat crazy!),
    feedback sounds (with Star Trek as general paragon :-), and
    some gimmicks.

    Janis

    I heard an apocryphal story several years or maybe a couple of
    decades ago that one car company hired the Yamaha audio company
    to design the exhaust system of one of their vehicles, and the
    result sounded EXTREMELY good.

    MAY be true.

    IC engines CAN sound REALLY nice - if properly 'tuned'.

    Electrics ... TOO quiet, people use engine sounds as
    part of their instant eval to cross a street. No sound
    equals no safety. The blind should SUE.

    At least in the US, I understand there is a government-imposed
    requirement for electrics to make noise. When I first heard of
    that requirement, I thought about selling a retrofit kit that
    would contain a pack of playing cards and a few clothespins. :-)
    I have described the artificial noise they create as about
    halfway between a flying saucer and a garbage disposal.


    Fun ... but for practical safety they should
    sound similar to a IC-powered car.

    That last-instant decision whether to step out
    in the street is as much based on sound as by
    sight. It's automatic, you don't even think
    about it. The 'broad' spectrum of an IC engine
    offers more cues than any electronic 'beeps'.

    I see Musk is discontinuing the 'CyberTruck' next
    year ... the factory will be dedicated to making
    his self-driving taxicabs.

    Frankly, I'm of the age now where an automatic
    taxi, even kinda slow, may be necessary soon.
    Most of the 'regular cab companies' are GONE
    now, replaced by Uber and such - and I'm not
    really set up well for those schemes and not
    interested in flipping banking numbers between
    fuckin' phones. Something where I got a CC
    charge-per-ride would be great - and no driver
    to tip or offend or worry about going psycho.

    Self-driving MOSTLY works now, though it still
    can miss a few things. Keep the speeds low and
    it has a much better chance of doing right.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From rbowman@bowman@montana.com to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Sun Nov 8 01:23:26 2026
    From Newsgroup: comp.os.linux.misc

    On Sat, 7 Nov 2026 21:53:26 +0000, The Natural Philosopher wrote:

    On 07/11/2026 14:37, Peter Flass wrote:
    On 11/7/26 04:45, The Natural Philosopher wrote:
    On 07/11/2026 04:09, c186282 wrote:
    ASM was also *the* way for the older micro-controllers
    -a-a for a long time. However some of the optimized 'C' compilers can
    -a-a now rival ASM for speed and size and you don't have to hand-code
    -a-a bit-banging serial either

    At some point I idly compiled some gcc C to assembler . I spent a
    pointless half hour trying to work out if I could have done it better.

    I couldn't..

    In terms of modern micros I wouldn't even bother unless I couldn't do
    it in 'C'.


    But then you add all the cruft that GCC drags in and you're back where
    you started (size-wise, at least).
    Hardly.


    When PCs were new I used a compiler called "Aztec C" that had almost no
    overhead in excess libraries, start-up code, etc., unless you asked for
    them.

    Well in general you DO ask for them

    I fixed a 20 year old bug that was caused by the programmer deciding
    adding '-lm' to the compilation would be too heavy. iirc his version of hypot() worked most of the time.



    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From rbowman@bowman@montana.com to comp.os.linux.misc on Sun Nov 8 01:29:52 2026
    From Newsgroup: comp.os.linux.misc

    On 7 Nov 2026 22:48:01 GMT, Robert Riches wrote:

    On 2026-11-07, The Natural Philosopher <tnp@invalid.invalid> wrote:
    On 07/11/2026 04:50, Robert Riches wrote:
    I heard an apocryphal story several years or maybe a couple of decades
    ago that one car company hired the Yamaha audio company to design the
    exhaust system of one of their vehicles, and the result sounded
    EXTREMELY good.

    I like the formula one engine programmed to play 'God Save the Queen'
    https://www.youtube.com/watch?v=XRXwWbo_mX0

    I recognized the tune, but I know it by a different title and different
    text.

    https://www.youtube.com/watch?v=g-38GX2YQig

    That version would be a little more challenging.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lynn Wheeler@lynn@garlic.com to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Sat Nov 7 16:04:20 2026
    From Newsgroup: comp.os.linux.misc

    c186282 <c186282@nnada.net> writes:
    The 'mainframe' era is nearly over - to be replaced
    by cluster/distributed/AI.

    However IBM is already into all that, WILL make money.

    Many baskets ... How To Survive/Thrive.

    Hmmmm ... do remember when IBM was "typewriters" ...

    1988, Nick Donofrio approves HA/6000, originally for NYTimes to move
    their newspaper system ("ATEX") off DEC VAXCluster to RS/6000 (running
    project out at Los Gatos lab). I rename it HA/CMP https://en.wikipedia.org/wiki/IBM_High_Availability_Cluster_Multiprocessing when I start doing technical/scientific cluster scale-up with national
    labs (LANL, LLNL, NCAR, etc) and commercial cluster scale-up with RDBMS
    vendors (Oracle, Sybase, Ingres, Informix) with VAXCluster support in
    same source base with UNIX (planning on using Hursley 9333 for mid-range
    and full FCS for both technical and commercial scaleup). DB2 was still
    all mainframe source, IBM non-mainframe "DB2" had yet to be developed.

    IBM S/88 (relogo'ed Stratus) Product Administrator started taking us
    around to their customers and also had me write a section for the
    corporate continuous availability strategy document (it gets pulled when
    both AS400/Rochester and mainframe/POK complain they couldn't meet requirements). Had coined "disaster survivability" and "geographic survivability" (as counter to disaster/recovery) when out marketing
    HA/CMP.

    One of the first HA/CMP customer installs was new Indian Reservation
    Casino in Connecticut, was suppose to have week of testing before
    opening ... but after 24hrs, they decided to open the doors (based on
    projected revenue; at the time was largest in the US, still one of the
    largest in the country) https://en.wikipedia.org/wiki/Foxwoods_Resort_Casino#Debt_default

    Early Jan92, there was HA/CMP meeting with Oracle CEO and IBM/AWD
    executive Hester tells Ellison that we would have 16-system clusters by
    mid92 and 128-system clusters by ye92. Mid-jan92, I update FSD on HA/CMP
    work with national labs and FSD decides to go with HA/CMP for federal supercomputers. By end of Jan, we are told that cluster scale-up is
    being transferred to Kingston for announce as IBM Supercomputer (SP, technical/scientific *ONLY*) and we aren't allowed to work with anything
    that has more than four systems (we leave IBM a few months later). A
    couple weeks later, 17feb1992, Computerworld news ... IBM establishes laboratory to develop parallel systems (pg8) https://archive.org/details/sim_computerworld_1992-02-17_26_7

    Some speculation that HA/CMP would have eaten the mainframe in the
    commercial market. 1993 industry benchmarks (number of program
    iterations compared to the industry MIPS/BIPS reference platform):

    ES/9000-982 : 8CPU 408MIPS, (51MIPS/CPU)
    RS6000/990 (RIOS chipset) : 1-CPU: 126MIPS

    Executive we had reported to, goes over to head up Somerset/AIM (Apple,
    IBM, Motorola) to do single chip 801/RISC (Power/PC) and uses Motorola
    88k bus/cache enabling SMP implementations.

    In the mid-90s, the i86 chip makers then do hardware layer that
    translates i86 instructions into RISC micro-ops for actual execution,
    largely negating the performance difference with RISC platforms. 1999
    industry benchmarks:
    IBM PowerPC 440: 1,000MIPS
    Intel Pentium3 2,054MIPS

    Dec2000 (IBM mainframe) benchmark:
    IBM z900 16CPU: 2.5BIPS, 156MIPS/CPU

    Also 1988, IBM branch office asked if I could help LLNL (national lab) standardize some serial stuff they were working with ... which quickly
    becomes Fibre Channel Standard ("FCS", including some stuff I had done
    in 1980; initially planning on using 1gbit transfer, full-duplex,
    aggregate transfer 200mbytes/sec & 64-port non-blocking switches, with
    high-end HA/CMP installations). POK mainframe eventually announces its
    serial stuff (when it is already obsolete) as "ESCON" (initially
    10mbytes/sec). Some POK engineers then become involved with FCS and
    define a heavy weight protocol that radically cuts the native
    throughput, eventually announced as FICON.

    Latest public benchmark I've found is 2010 z196 "Peak I/O" getting 2M
    IOPS using 104 FICON (20K IOPS/FICON). About the same time a native FCS
    is announced for E5-2600 server blades getting over million IOPS (two
    such FCS higher throughput than 104 FICON running over FCS). Note IBM
    docs has SAPs (system assist processors that do actual I/O) CPU be kept
    to 70% ... or 1.5M IOPS ... also no CKD DASD have been made for decades
    (just simulated with industry standard fixed-block devices).

    This "Evolution of the System z Channel" (SHARE 2011 presentation at
    wayback machine) https://web.archive.org/web/20170829213251/https://share.confex.com/share/117/webprogram/Handout/Session9931/9934pdhj%20v0.pdf
    talks about after 2010 z196 "Peak I/O" benchmark and "zHPF" that
    increased throughput from 20K IOPS/FICON to 52K IOPS/FICON then to 92K IOPS/FICON (aka from 1/50th to almost 1/10th the native FCS announced in
    2010).
    --
    virtualization experience starting Jan1968, online at home since Mar1970
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Robert Riches@spamtrap42@jacob21819.net to comp.os.linux.misc on Sun Nov 8 04:29:21 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-08, rbowman <bowman@montana.com> wrote:
    On 7 Nov 2026 22:48:01 GMT, Robert Riches wrote:

    On 2026-11-07, The Natural Philosopher <tnp@invalid.invalid> wrote:
    On 07/11/2026 04:50, Robert Riches wrote:
    I heard an apocryphal story several years or maybe a couple of decades >>>> ago that one car company hired the Yamaha audio company to design the
    exhaust system of one of their vehicles, and the result sounded
    EXTREMELY good.

    I like the formula one engine programmed to play 'God Save the Queen'
    https://www.youtube.com/watch?v=XRXwWbo_mX0

    I recognized the tune, but I know it by a different title and different
    text.

    https://www.youtube.com/watch?v=g-38GX2YQig

    That version would be a little more challenging.

    To some ears, the F1 engine revving rapidly sounds fairly similar
    to much of the first 40 seconds of that piece--all I could stand
    to listen to. 1/2 ;-)
    --
    Robert Riches
    spamtrap42@jacob21819.net
    (Yes, that is one of my email addresses.)
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lars Poulsen@lars@beagle-ears.com to comp.os.linux.misc on Sun Nov 8 05:12:38 2026
    From Newsgroup: comp.os.linux.misc

    On 11/7/2026 17:23, rbowman wrote:
    On Sat, 7 Nov 2026 21:53:26 +0000, The Natural Philosopher wrote:

    On 07/11/2026 14:37, Peter Flass wrote:
    On 11/7/26 04:45, The Natural Philosopher wrote:
    On 07/11/2026 04:09, c186282 wrote:
    ASM was also *the* way for the older micro-controllers
    -a-a for a long time. However some of the optimized 'C' compilers can >>>>> -a-a now rival ASM for speed and size and you don't have to hand-code >>>>> -a-a bit-banging serial either

    At some point I idly compiled some gcc C to assembler . I spent a
    pointless half hour trying to work out if I could have done it better. >>>>
    I couldn't..

    In terms of modern micros I wouldn't even bother unless I couldn't do
    it in 'C'.


    But then you add all the cruft that GCC drags in and you're back where
    you started (size-wise, at least).
    Hardly.


    When PCs were new I used a compiler called "Aztec C" that had almost no
    overhead in excess libraries, start-up code, etc., unless you asked for
    them.

    Well in general you DO ask for them

    I fixed a 20 year old bug that was caused by the programmer deciding
    adding '-lm' to the compilation would be too heavy. iirc his version of hypot() worked most of the time.

    After 60 years of programming, I did not know it was in the library. I
    always just did
    h = sqrt(x*x+y*y);
    but if he did not use -lm, did he do his own sqrt() too? And was this on
    a machine without hardware floating point?
    --
    Lars Poulsen - an old geek in Santa Barbara, California
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lars Poulsen@lars@beagle-ears.com to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Sun Nov 8 05:59:55 2026
    From Newsgroup: comp.os.linux.misc

    On 11/7/2026 18:04, Lynn Wheeler wrote:
    1988, Nick Donofrio approves HA/6000, originally for NYTimes to move
    their newspaper system ("ATEX") off DEC VAXCluster to RS/6000 (running project out at Los Gatos lab). I rename it HA/CMP https://en.wikipedia.org/wiki/IBM_High_Availability_Cluster_Multiprocessing when I start doing technical/scientific cluster scale-up with national
    labs (LANL, LLNL, NCAR, etc) and commercial cluster scale-up with RDBMS vendors (Oracle, Sybase, Ingres, Informix) with VAXCluster support in
    same source base with UNIX (planning on using Hursley 9333 for mid-range
    and full FCS for both technical and commercial scaleup). DB2 was still
    all mainframe source, IBM non-mainframe "DB2" had yet to be developed.

    Around 2001, my group had developed a T1-over-UDP-over-Ethernet bridge
    to use in tandem with our wireless ethernet bridge product. After we
    spun out from WiLan, and became Afar, we also marketed it separately.
    Our largest sale was to NYT, to replace the dual T1 lines they had to
    their (16?) regional printing plants, connected to the phototypesetters
    in the print shops. At 42nd street they had a rack with 60(?) of our
    boxes. While they wanted the two lines for each remote site to be on
    separate boxes connected to separate IP networks, two lines to different
    sites could share one of our units, which had two T1 jacks.

    System worked well; I once got to visit the Old Lady shortly before they
    moved down the street.

    IBM S/88 (relogo'ed Stratus) Product Administrator started taking us
    around to their customers and also had me write a section for the
    corporate continuous availability strategy document (it gets pulled when
    both AS400/Rochester and mainframe/POK complain they couldn't meet requirements). Had coined "disaster survivability" and "geographic survivability" (as counter to disaster/recovery) when out marketing
    HA/CMP.

    I was curious about S/88, which I did not know about before. CoPilot
    pulled together a nice synopsis, from which I learned that Stratus and
    Tandem both had figured out how to do rolling upgrades in a non-stop environment; I had thought that was an IBM mainframe invention.

    I also learned that both Stratus and Tandem survive to this day, in a
    fashion:
    Tandem-> Compaq-> HP - is now HPE Non-Stop

    Stratus-> Ascend-> Lucent -> spin-out Stratus Tech
    building HA systems on Xeon processors.
    Now called Penguin Solutions.
    Many techies despise System/Z as an overpriced dinosaur with an
    obsolete architechture. But I wonder if it is any less expensive to dip
    the Intel processors in that magic sauce. I bet IBM has better scale to
    their builds.
    --
    Lars Poulsen - an old geek in Santa Barbara, California
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Leroy H@lh@somewhere.net to comp.os.linux.misc on Sun Nov 8 14:31:00 2026
    From Newsgroup: comp.os.linux.misc

    On Sun, 8 Nov 2026 05:12:38 -0800, Lars Poulsen wrote:


    I did not know it was in the library. I
    always just did
    h = sqrt(x*x+y*y);


    Another case of lack of comp sci training.

    The function exists because it avoids a loss of precision and
    it also checks for under/overflow conditions.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lynn Wheeler@lynn@garlic.com to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Sun Nov 8 08:11:28 2026
    From Newsgroup: comp.os.linux.misc

    Lars Poulsen <lars@beagle-ears.com> writes:
    Around 2001, my group had developed a T1-over-UDP-over-Ethernet bridge
    to use in tandem with our wireless ethernet bridge product. After we
    spun out from WiLan, and became Afar, we also marketed it
    separately. Our largest sale was to NYT, to replace the dual T1 lines
    they had to their (16?) regional printing plants, connected to the phototypesetters in the print shops. At 42nd street they had a rack
    with 60(?) of our boxes. While they wanted the two lines for each
    remote site to be on separate boxes connected to separate IP networks,
    two lines to different sites could share one of our units, which had
    two T1 jacks.

    System worked well; I once got to visit the Old Lady shortly before
    they moved down the street.

    early 80s, I got HSDT project, T1 and faster computer links (both
    terrestrial and satellite) and battles with SNA group (60s, IBM had 2701 supporting T1, 70s with SNA/VTAM and issues, links were capped at 56kbit
    ... and I had to mostly resort to non-IBM hardware).

    Mid-80s, communication group was also fighting off release of mainframe
    TCP/IP support and when that failed, they said that since they had
    corporate responsibility for everything that crossed datacenter walls,
    it had to be release through them. I then add RFC1044 and in some tuning
    tests at Cray Research between a Cray and 4341, got sustained 4341
    channel throughput using only modest amount of 4341 CPU (something like
    500 times improvement in bytes moved per instruction executed).

    HSDT also was working with NSF director and was suppose to get $20M to interconnect the NSF Supercomputer centers. NSF gave grant to Univ. Cal
    to build Berkeley supercomputer center and I was asked to give them
    talks on the interconnect. Turns out regents had san diego getting the
    next new bldg ... and it became the San Diego Supercomputer center
    instead.

    Then congress cuts the budget, some other things happened and eventually
    there was RFP released (in part based on what we already had
    running). NSF 28Mar1986 Preliminary Announcement: https://web.archive.org/web/20260311074443/https://www.garlic.com/~lynn/2002k.html#12
    The OASC has initiated three programs: The Supercomputer Centers Program
    to provide Supercomputer cycles; the New Technologies Program to foster
    new supercomputer software and hardware developments; and the Networking Program to build a National Supercomputer Access Network - NSFnet.

    IBM internal politics was not allowing us to bid. The NSF director tried
    to help by writing the company a letter (3Apr1986, NSF Director to IBM
    Chief Scientist and IBM Senior VP and director of Research, copying IBM
    CEO) with support from other gov. agencies ... but that just made the
    internal politics worse (as did claims that what we already had
    operational was at least 5yrs ahead of the winning bid), as regional
    networks connect in, NSFnet becomes the NSFNET backbone, precursor to
    modern internet

    ... aka NCSA ... grant for new "supercomputer software" https://www.ncsa.illinois.edu/
    --
    virtualization experience starting Jan1968, online at home since Mar1970
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lynn Wheeler@lynn@garlic.com to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Sun Nov 8 08:32:30 2026
    From Newsgroup: comp.os.linux.misc

    Lynn Wheeler <lynn@garlic.com> writes:
    Latest public benchmark I've found is 2010 z196 "Peak I/O" getting 2M
    IOPS using 104 FICON (20K IOPS/FICON). About the same time a native FCS
    is announced for E5-2600 server blades getting over million IOPS (two
    such FCS higher throughput than 104 FICON running over FCS). Note IBM
    docs has SAPs (system assist processors that do actual I/O) CPU be kept
    to 70% ... or 1.5M IOPS ... also no CKD DASD have been made for decades
    (just simulated with industry standard fixed-block devices).

    2010:
    Max configured z196, $30M, 80core, 50BIPS.
    Intel E5-2600, 16cores, 500BIPS

    IBM base list price for E5-2600 server blade was $1815

    2010 Note large cloud operations with multiple megadatacenters around
    the world with half million or more server blades with enormous
    automation and able to run megadatacenter with 70-80 staff and
    assembled their own server blades at 1/3rd the cost of brand
    name vendors.

    There was then industry articles that server chip makers were shipping
    half their products directly to megadatacenters. Shortly later, IBM
    sells off its server blade business.
    --
    virtualization experience starting Jan1968, online at home since Mar1970
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From rbowman@bowman@montana.com to comp.os.linux.misc on Sun Nov 8 18:52:33 2026
    From Newsgroup: comp.os.linux.misc

    On Sun, 8 Nov 2026 05:12:38 -0800, Lars Poulsen wrote:

    After 60 years of programming, I did not know it was in the library. I
    always just did
    h = sqrt(x*x+y*y);
    but if he did not use -lm, did he do his own sqrt() too? And was this on
    a machine without hardware floating point?

    I can't remember exactly what the problem was but he made an assumption
    that didn't require finding a hypotenuse. It worked well enough that
    nobody noticed.

    iirc it had to do with alerts. The clients could add specific points to
    the system, a hazardous storage location, for example and set a radius for
    the trigger. If an incident was reported at point A, any of the alerts
    within N feet should pop up. The client reported an alert was within the configured N feet but didn't show up and provided a screen shot of a map
    in Arc Desktop. It was very close but no cigar.

    As I think about it I believe the simplification was creating a bounding
    box rather than a circle with the radius of N to avoid doing the math.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lawrence =?iso-8859-13?q?D=FFOliveiro?=@ldo@nz.invalid to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Sun Nov 8 21:31:18 2026
    From Newsgroup: comp.os.linux.misc

    On Sun, 8 Nov 2026 05:59:55 -0800, Lars Poulsen wrote:

    ... from which I learned that Stratus and Tandem both had figured
    out how to do rolling upgrades in a non-stop environment; I had
    thought that was an IBM mainframe invention.

    The IBM mega-marketing machine strikes again.

    Or past tense, really; IBM no longer has that power. Microsoft
    inherited that mantle, but it, too, is losing its grip.

    Many techies despise System/Z as an overpriced dinosaur with an
    obsolete architechture. But I wonder if it is any less expensive to
    dip the Intel processors in that magic sauce. I bet IBM has better
    scale to their builds.

    There was a paper on how Google did it, published 10 or 20 years ago.
    At that point, they had a total complement of about 460,000 x86
    servers. They would buy the cheapest component they could for every
    component but one: the power supply.

    With that kind of scale, who cares if 100 or 1000 machines are broken
    at any moment? The users certainly wonrCOt notice.

    Mainframes stopped being relevant back in the 1980s; every further
    development since then has just been driven by customer inertia.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lawrence =?iso-8859-13?q?D=FFOliveiro?=@ldo@nz.invalid to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Sun Nov 8 21:32:13 2026
    From Newsgroup: comp.os.linux.misc

    On Sun, 08 Nov 2026 08:32:30 -1000, Lynn Wheeler wrote:

    There was then industry articles that server chip makers were
    shipping half their products directly to megadatacenters. Shortly
    later, IBM sells off its server blade business.

    Which kind of begs the question, doesnrCOt it: who bought it, and why?
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lars Poulsen@lars@beagle-ears.com to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Sun Nov 8 13:45:03 2026
    From Newsgroup: comp.os.linux.misc

    On 11/8/2026 13:31, Lawrence DrCOOliveiro wrote:
    There was a paper on how Google did it, published 10 or 20 years ago.
    At that point, they had a total complement of about 460,000 x86
    servers. They would buy the cheapest component they could for every
    component but one: the power supply.

    With that kind of scale, who cares if 100 or 1000 machines are broken
    at any moment? The users certainly wonrCOt notice.

    The parallelism of a Google is a very different beast from the robust parallelism of a banking mainframe.

    Google needs a way to route an incoming connection to a random running
    server instance. It does not matter which one, and if the one that was
    picked has died, the first packet retry will get to a better one.

    That is very different from transactions that must execute exactly once.

    Mainframes stopped being relevant back in the 1980s; every further development since then has just been driven by customer inertia.

    If we define mainframes a large, very robust systems, they are still
    very relevant. Although it may be true that the major technology was in
    place before 1990 (non-stop cluster systems, transaction locking with completion/rollback and strategic redundancy). What you call customer
    inertia is also known as vendor lock-in.
    --
    Lars Poulsen - an old geek in Santa Barbara, California
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lynn Wheeler@lynn@garlic.com to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Sun Nov 8 11:55:20 2026
    From Newsgroup: comp.os.linux.misc

    Lawrence DrCOOliveiro <ldo@nz.invalid> writes:
    Which kind of begs the question, doesnrCOt it: who bought it, and why?

    IBM sold server blades to save company that they prevsiouly sold its PC business to. https://www.cnet.com/tech/tech-industry/ibm-sells-its-x86-server-business-to-lenovo-for-2-3-billion/
    (Jan2014) IBM sells its x86 server business to Lenovo for $2.3 billion
    An expected deal to sell off its Intel-based servers to the Chinese
    electronics company has arrived, reflecting the shifting realities of
    the market for powerful networked machines.

    --
    virtualization experience starting Jan1968, online at home since Mar1970
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lynn Wheeler@lynn@garlic.com to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Sun Nov 8 12:21:49 2026
    From Newsgroup: comp.os.linux.misc

    Lars Poulsen <lars@beagle-ears.com> writes:
    If we define mainframes a large, very robust systems, they are still
    very relevant. Although it may be true that the major technology was
    in place before 1990 (non-stop cluster systems, transaction locking
    with completion/rollback and strategic redundancy). What you call
    customer inertia is also known as vendor lock-in.

    trivia, after early 70s "Future System" http://www.jfsowa.com/computer/memo125.htm https://en.wikipedia.org/wiki/IBM_Future_Systems_project https://people.computing.clemson.edu/~mark/fs.html

    imploded, head of POK was convincing corporate to kill the VM370
    product, shutdown the development group and transfer all the people to
    POK for MVS/XA. Endicott eventually acquires VM370 responsibility (for
    the mid-range), but has to recreate a development group from scratch. I
    get asked to help with a 16-CPU 370 and we con the 3033 processor
    engineers into helping in their spare time (a lot more interesting that remapping 168 logic to 20% faster chips). Everybody thought it was great
    until somebody tells the head of POK that it could be decades POK's
    favorite son operating system ("MVS") had effective 16-CPU (MVS
    documents were claiming that MVS 2-CPU support only had 1.2-1.5 times
    the throughput of single CPU, because of high multiprocessor support
    overhead, POK doesn't ship a 16-CPU system until after the turn of the century). Head of POK then invites some of us to never visit POK again
    and directs the 3033 processor engineers, "heads down and no
    distractions".

    About the same time, Endicott talks me into helping with ECPS for the
    138/148 (also used with 4331/4341), identify highest executed 6kbytes of
    VM370 kernel execution for re-implementing directly in machine microcode
    (at 10:1 performance increase, mid-range 370s were avg. 10 native
    instructions for every 370 instruction). Archived post with initial
    analysis, 6kbytes accounted for 79.55% of kernel execution: https://web.archive.org/web/20260416141952/https://www.garlic.com/~lynn/94.html#21

    After being invited to never visit POK again, I transfer out to IBM San
    Jose Research on the west coast and worked with Jim Gray and Vera Watson
    on the original SQL/relational RDBMS, System/R. After Jim Gray left IBM
    SJR for Tandem (fall1980), he did studies of system availability and
    outages. He was finding that hardware was becoming much more reliable
    and outages were shifting to environmental (floods, earthquakes, storms,
    etc) and people mistakes, needing redundant systems at geographically
    separated distances. https://courses.cs.duke.edu/spring11/cps210/papers/Gray_computers_stop.pdf https://web.archive.org/web/20260219063547/https://www.garlic.com/~lynn/grayft84.pdf

    As Jim was leaving, he was foisting off stuff on me including BofA;
    Research had System/R joint study with BofA and they ordered 60 VM4341s
    for System/R. Some amount of resistance from STL, IMS and next great
    DBMS "EAGLE". Was able to do tech transfer to Endicott (under the radar
    while company was pre-occupied with "EAGLE") for SQL/DS. Then when
    "EAGLE" implodes there was request how fast could System/R be ported to
    MVS, eventually released as "DB2" (originally for decision support
    only).

    Also when 1st transfer to west coast, I got to wander around silicon
    valley datacenters (and consolidated HONE was just up the road),
    including disk bldg14/engineering and bldg15/product test across the
    street. They had been doing 7/24, prescheduled stand-alone testing and mentioned that recently they had tried MVS, but it had 15min MTBF (in
    that environment, requiring manual re-ipl). I offer to rewrite I/O
    Supervisor, making it bullet proof and never fail allowing any amount of ondemand, concurrent testing (greatly improving productivity). I then do
    an internal research report "I/O Reliability Enhancement" and happen to
    mention MVS 15min MTBF, bringing down the wrath of the MVS organization
    on my head.

    Bldg15 gets the 1st engineering 3033 outside POK 3033 processor
    engineering (trivia: after 3033s started shipping, 3033 processor
    engineers start on trout/3090). Testing was only percent or two of CPU,
    so we scrounge up a 3830 controller and 3330 string and setup our own
    private online service. Summer 1978, bldg15 also gets engineering
    E5/4341 (year before 1st customer ship). Jan1979, branch office hears
    about it and cons me into doing (CDC6600 Fortran) benchmark for national
    lab that was looking at getting 70 VM4341s for compute farm (sort of
    leading edge of the coming cluster supercomputer tsunami). Spring 1979,
    some USAFDC (in the Pentagon) wanted to come by to talk to me about 20
    VM4341 systems, visit kept being delayed, by the time they came by (six
    months later), it had grown from 20 to 210. Also saw large corporations ordering hundreds of VM4341s at a time for installing out in
    departmental areas (sort of the leading edge of the coming distributed computing tsunami). VM4341s also contributed to HONE system popping up
    all over the world (and inside IBM, departmental conference rooms
    becomes scarce, so many converted to VM4341 rooms).
    --
    virtualization experience starting Jan1968, online at home since Mar1970
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lawrence =?iso-8859-13?q?D=FFOliveiro?=@ldo@nz.invalid to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Sun Nov 8 23:51:39 2026
    From Newsgroup: comp.os.linux.misc

    On Sun, 8 Nov 2026 13:45:03 -0800, Lars Poulsen wrote:

    On 11/8/2026 13:31, Lawrence DrCOOliveiro wrote:

    There was a paper on how Google did it, published 10 or 20 years
    ago. At that point, they had a total complement of about 460,000
    x86 servers. They would buy the cheapest component they could for
    every component but one: the power supply.

    With that kind of scale, who cares if 100 or 1000 machines are
    broken at any moment? The users certainly wonrCOt notice.

    The parallelism of a Google is a very different beast from the
    robust parallelism of a banking mainframe.

    Mainframes have never been that rCLrobustrCY. YourCOve mentioned elsewhere learning about Stratus and Tandem, who took robustness to a new level
    back in the 1980s, beyond anything the mainframe world could match.
    Think of that kind of redundancy multiplied tenfold, a hundredfold, a thousandfold, all built on open-source technologies: thatrCOs the kind
    of world we live in today.

    That is very different from transactions that must execute exactly
    once.

    That technology is pretty commonplace nowadays <https://kafka.apache.org/44/getting-started/introduction/>.

    Mainframes stopped being relevant back in the 1980s; every further
    development since then has just been driven by customer inertia.

    If we define mainframes a large, very robust systems, they are still
    very relevant.

    ThatrCOs kind of a circular definition though, isnrCOt it. IrCOll stick with the traditional definition: one which maximized I/O bandwidth over
    minimizing latency, batch operation efficiency over interactive
    responsiveness. That was very much IBMrCOs core business, and the
    obsolescence of that market directly correlates with IBM dwindling to
    a shadow of its former self.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From c186282@c186282@nnada.net to comp.os.linux.misc on Sun Nov 8 21:29:51 2026
    From Newsgroup: comp.os.linux.misc

    On 11/8/26 09:31, Leroy H wrote:
    On Sun, 8 Nov 2026 05:12:38 -0800, Lars Poulsen wrote:


    I did not know it was in the library. I
    always just did
    h = sqrt(x*x+y*y);


    Another case of lack of comp sci training.

    The function exists because it avoids a loss of precision and
    it also checks for under/overflow conditions.

    Hey, GOTTA do that !

    However it's not a 'lack of Comp Sci' training
    involved. ANY sane programmer would deal with
    those problems.

    Typical 'good programming' - at least a third
    of yer code is for dealing with potential FUCK
    UPS. If the code is for 'user interaction' it
    can become HALF yer code. Dealing with Stupid
    Humans is The Worst !

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From c186282@c186282@nnada.net to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Sun Nov 8 21:49:05 2026
    From Newsgroup: comp.os.linux.misc

    On 11/8/26 16:32, Lawrence DrCOOliveiro wrote:
    On Sun, 08 Nov 2026 08:32:30 -1000, Lynn Wheeler wrote:

    There was then industry articles that server chip makers were
    shipping half their products directly to megadatacenters. Shortly
    later, IBM sells off its server blade business.

    Which kind of begs the question, doesnrCOt it: who bought it, and why?

    Also, how come IBM was smart enough to
    sell it ?

    Of all the Old Biggies, seems only IBM still
    survives/thrives. DEC, Burroughs, Honeywell, DG
    and others - down the toilet. All of those
    made great computers, but SOMETHING went wrong
    and they couldn't cope.

    Somebody at IBM is making smart decisions,
    diversifying properly. Develop, make money,
    then dump (with royalty agreements). No tech
    is 'forever', so exploit "the window" and
    then move on.

    Yea, yea, 'Dilbert' made fun of them - but IBM
    is STILL HERE, and still a 'blue-chip' stock.

    What of their approach might benefit YOUR biz ?

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Peter Flass@Peter@Iron-Spring.com to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Sun Nov 8 20:33:45 2026
    From Newsgroup: comp.os.linux.misc

    On 11/8/26 16:51, Lawrence DrCOOliveiro wrote:
    On Sun, 8 Nov 2026 13:45:03 -0800, Lars Poulsen wrote:

    On 11/8/2026 13:31, Lawrence DrCOOliveiro wrote:

    There was a paper on how Google did it, published 10 or 20 years
    ago. At that point, they had a total complement of about 460,000
    x86 servers. They would buy the cheapest component they could for
    every component but one: the power supply.

    With that kind of scale, who cares if 100 or 1000 machines are
    broken at any moment? The users certainly wonrCOt notice.

    The parallelism of a Google is a very different beast from the
    robust parallelism of a banking mainframe.

    Mainframes have never been that rCLrobustrCY. YourCOve mentioned elsewhere learning about Stratus and Tandem, who took robustness to a new level
    back in the 1980s, beyond anything the mainframe world could match.
    Think of that kind of redundancy multiplied tenfold, a hundredfold, a thousandfold, all built on open-source technologies: thatrCOs the kind
    of world we live in today.


    Mainframes typically achieve an uptime percentage between 99.999% (five
    nines) and 99.9999999% (nine nines)
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From c186282@c186282@nnada.net to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Sun Nov 8 22:59:21 2026
    From Newsgroup: comp.os.linux.misc

    On 11/8/26 22:33, Peter Flass wrote:
    On 11/8/26 16:51, Lawrence DrCOOliveiro wrote:
    On Sun, 8 Nov 2026 13:45:03 -0800, Lars Poulsen wrote:

    On 11/8/2026 13:31, Lawrence DrCOOliveiro wrote:

    There was a paper on how Google did it, published 10 or 20 years
    ago. At that point, they had a total complement of about 460,000
    x86 servers. They would buy the cheapest component they could for
    every component but one: the power supply.

    With that kind of scale, who cares if 100 or 1000 machines are
    broken at any moment? The users certainly wonrCOt notice.

    The parallelism of a Google is a very different beast from the
    robust parallelism of a banking mainframe.

    Mainframes have never been that rCLrobustrCY. YourCOve mentioned elsewhere >> learning about Stratus and Tandem, who took robustness to a new level
    back in the 1980s, beyond anything the mainframe world could match.
    Think of that kind of redundancy multiplied tenfold, a hundredfold, a
    thousandfold, all built on open-source technologies: thatrCOs the kind
    of world we live in today.


    Mainframes typically achieve an uptime percentage between 99.999% (five nines) and 99.9999999% (nine nines)

    They ARE very very very good !

    Noted in this group though - I/O arch diffs between
    mainframes and minis.

    Will someone try to elaborate ?

    Started off on a "mini" - PDP-11 - with a PDP-8
    kinda in the background (but wasn't really allowed
    to use it for anything). This I/O diff, didn't
    NOTICE it, but was likely there. Long-after
    insights would be interesting.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lawrence =?iso-8859-13?q?D=FFOliveiro?=@ldo@nz.invalid to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Mon Nov 9 04:42:48 2026
    From Newsgroup: comp.os.linux.misc

    On Sun, 8 Nov 2026 20:33:45 -0700, Peter Flass wrote:

    Mainframes typically achieve an uptime percentage between 99.999%
    (five nines) and 99.9999999% (nine nines)

    The hyperscalers probably manage that well, if not better.

    There is an article from 1986 on Bitsavers which describes having to
    reboot an IBM mainframe to turn daylight saving on and off.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Nuno Silva@nunojsilva@invalid.invalid to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Mon Nov 9 10:01:13 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-09, Lawrence DrCOOliveiro wrote:

    On Sun, 8 Nov 2026 20:33:45 -0700, Peter Flass wrote:

    Mainframes typically achieve an uptime percentage between 99.999%
    (five nines) and 99.9999999% (nine nines)

    The hyperscalers probably manage that well, if not better.

    There is an article from 1986 on Bitsavers which describes having to
    reboot an IBM mainframe to turn daylight saving on and off.

    *That* sounds like complicating a problem that would be made simpler by
    not using the local time, or at least DST-adjusted local time.

    What did they do when the clock moved backwards? Stop operations for one
    hour? Or eventual clashes would be acceptable? Separate, incrementing
    unique IDs?
    --
    Nuno Silva
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From The Natural Philosopher@tnp@invalid.invalid to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Mon Nov 9 11:08:53 2026
    From Newsgroup: comp.os.linux.misc

    On 09/11/2026 03:33, Peter Flass wrote:
    Mainframes typically achieve an uptime percentage between 99.999% (five nines) and 99.9999999% (nine nines)

    Correcting Lawrence won't change a thing.
    TalkingBollocksrao is a way of life for some people.

    Look at yer president....
    --
    Civilization exists by geological consent, subject to change without notice.
    rCo Will Durant

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From scott@scott@slp53.sl.home (Scott Lurndal) to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Mon Nov 9 15:59:03 2026
    From Newsgroup: comp.os.linux.misc

    Lawrence =?iso-8859-13?q?D=FFOliveiro?= <ldo@nz.invalid> writes:
    On Sun, 8 Nov 2026 13:45:03 -0800, Lars Poulsen wrote:

    On 11/8/2026 13:31, Lawrence DrCOOliveiro wrote:

    There was a paper on how Google did it, published 10 or 20 years
    ago. At that point, they had a total complement of about 460,000
    x86 servers. They would buy the cheapest component they could for
    every component but one: the power supply.

    With that kind of scale, who cares if 100 or 1000 machines are
    broken at any moment? The users certainly wonrCOt notice.

    The parallelism of a Google is a very different beast from the
    robust parallelism of a banking mainframe.

    Mainframes have never been that rCLrobustrCY

    Sure they have. Burroughs had clustered systems by 1968,
    with shared storage and shared peripherals.


    . YourCOve mentioned elsewhere
    learning about Stratus and Tandem, who took robustness to a new level
    back in the 1980s, beyond anything the mainframe world could match.

    Sure, pull the other one. Several of my former Burroughs colleagues
    worked at Tandem, and we at Burroughs/Unisys built systems designed with
    fault tolerance as a key goal (e.g. OPUS, the design of which started
    in 1989).

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lynn Wheeler@lynn@garlic.com to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Mon Nov 9 07:32:09 2026
    From Newsgroup: comp.os.linux.misc

    c186282 <c186282@nnada.net> writes:
    Noted in this group though - I/O arch diffs between
    mainframes and minis.

    previously posted, 1988, ibm branch office asked me if I could help LLNL
    with standardization of some serial stuff they had been working with,
    which quickly becomes fibre-channel standard ("FCS", including some
    stuff I had done in the 80s, initially 1gbit transfer, full-duplex,
    aggregate 200mbytes/sec).
    https://en.wikipedia.org/wiki/Fibre_Channel

    then IBM releases some serial stuff they had been working with for
    nearly a decade (when it is already obsolete) as ESCON (initially 10mbytes/sec). Then some IBM engineers get involved with FCS and define
    a heavy-weight protocol that radically reduced native throughput https://en.wikipedia.org/wiki/FICON

    Latest public benchmark I've found is 2010 z196 "Peak I/O" getting 2M
    IOPS using 104 FICON (20K IOPS/FICON). About the same time a native FCS
    is announced for E5-2600 server blades getting over million IOPS (two
    such FCS higher throughput than 104 FICON running over FCS). Note IBM
    docs has SAPs (system assist processors that now do the actual I/O) CPU
    be kept to 70% ... or 1.5M IOPS ... also no CKD DASD have been made for
    decades (just simulated with industry standard fixed-block devices).

    This "Evolution of the System z Channel" (SHARE 2011 presentation at
    wayback machine) https://web.archive.org/web/20170829213251/https://share.confex.com/share/117/webprogram/Handout/Session9931/9934pdhj%20v0.pdf
    talks about after 2010 z196 "Peak I/O" benchmark and "zHPF" that
    increased throughput from 20K IOPS/FICON to 52K IOPS/FICON then to 92K IOPS/FICON (aka from 1/50th to almost 1/10th the native FCS announced in
    2010).
    --
    virtualization experience starting Jan1968, online at home since Mar1970
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From John Levine@johnl@taugh.com to comp.os.linux.misc,alt.folklore.computers on Mon Nov 9 21:20:02 2026
    From Newsgroup: comp.os.linux.misc

    According to c186282 <c186282@nnada.net>:
    Noted in this group though - I/O arch diffs between
    mainframes and minis.

    Will someone try to elaborate ?

    The classic difference is that mainframes have channels, so the
    processor sets up a bunch of work for the channel to do, the channel
    goes off and does it independently of the CPU. Minis had much less
    complex I/O controllers so they do more detailed stuff. An example I
    gave a few messages back was for a card reader. A mainframe starts a
    channel program, it reads in a card or maybe seveal cards, and reports
    back with an interrupt. On a mini there was an I/O operation to start
    the card reader, an interrupt for each column and an I/O to read it,
    and then something at the end to stop the reader or go to the next
    card.

    But that was 40 or 50 years ago. These days even cheap generic disk controllers do more than channels did in the 1960s, e.g., keep a queue
    of requests and schedule them to minimize delay, and handle spare block replacement for bad blocks.

    Modern mainframes have extremely capable channels but the distinction is
    more about reliability and things like hardware and software upgrades in
    place without rebooting.
    --
    Regards,
    John Levine, johnl@taugh.com, Primary Perpetrator of "The Internet for Dummies",
    Please consider the environment before reading this e-mail. https://jl.ly
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From c186282@c186282@nnada.net to comp.os.linux.misc,alt.folklore.computers on Mon Nov 9 22:07:57 2026
    From Newsgroup: comp.os.linux.misc

    On 11/9/26 16:20, John Levine wrote:
    According to c186282 <c186282@nnada.net>:
    Noted in this group though - I/O arch diffs between
    mainframes and minis.

    Will someone try to elaborate ?

    The classic difference is that mainframes have channels, so the
    processor sets up a bunch of work for the channel to do, the channel
    goes off and does it independently of the CPU. Minis had much less
    complex I/O controllers so they do more detailed stuff. An example I
    gave a few messages back was for a card reader. A mainframe starts a
    channel program, it reads in a card or maybe seveal cards, and reports
    back with an interrupt. On a mini there was an I/O operation to start
    the card reader, an interrupt for each column and an I/O to read it,
    and then something at the end to stop the reader or go to the next
    card.

    But that was 40 or 50 years ago. These days even cheap generic disk controllers do more than channels did in the 1960s, e.g., keep a queue
    of requests and schedule them to minimize delay, and handle spare block replacement for bad blocks.

    Modern mainframes have extremely capable channels but the distinction is
    more about reliability and things like hardware and software upgrades in place without rebooting.

    Ok ... so 'mainframes' as you saw them were more
    'directors' for distributed processing rather
    than THE processor.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Thomas Koenig@tkoenig@netcologne.de to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Tue Nov 10 07:09:17 2026
    From Newsgroup: comp.os.linux.misc

    c186282 <c186282@nnada.net> schrieb:

    Noted in this group though - I/O arch diffs between
    mainframes and minis.

    I have a somewhat personal definition of a mainframe, which I also
    includes the classical ones.

    A mainframe is a computer that

    - is very expensive
    - takes large resources to run (a computer center)
    - has very good performance in at least one field
    - is not a cluster, massively parallel etc

    As computers have evolved, the field of performance has shifted.
    The 704 and the Cray-1 both had high computing performance for
    their day, but low reliability. The 360/30 was slow, but had
    good I/O capabilities.

    Today's mainframes can no longer compete on compute power, but
    they have extremely high reliability (you can literally fire a
    bullet through a CPU, and things will migrate smoothly - if you
    don't hit the Recovery Unit), good I/O and lots of sub-processors
    to lighten the load on the CPU.

    I read that, when minis came along, their first defining
    characteristic was that they didn't need computer center resources
    and fit in a single room.
    --
    This USENET posting was made without artificial intelligence,
    artificial impertinence, artificial arrogance, artificial stupidity,
    artificial flavorings or artificial colorants.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lawrence =?iso-8859-13?q?D=FFOliveiro?=@ldo@nz.invalid to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Tue Nov 10 07:27:50 2026
    From Newsgroup: comp.os.linux.misc

    On Tue, 10 Nov 2026 07:09:17 -0000 (UTC), Thomas Koenig wrote:

    Today's mainframes can no longer compete on compute power, but they
    have extremely high reliability (you can literally fire a bullet
    through a CPU, and things will migrate smoothly - if you don't hit
    the Recovery Unit) ...

    Cloud providers like Amazon, Google, Microsoft etc offer much higher reliability through massive parallelism, based on cheap, yet highly
    redundant hardware running a Free Software stack that is not priced
    per CPU. That is much more cost-effective and any equivalent number of mainframes. And there is no rCLRecovery UnitrCY that you can shoot at to
    take it out.

    The nice thing is, you can run the same sort of rCLenterprise-classrCY
    setup that the big folks operate in your own homelab, because all the foundation software they use is available to you, too <https://docs.xcp-ng.org/vms/vm-migration/>.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Alexander Schreiber@als@usenet.thangorodrim.de to comp.os.linux.misc,alt.folklore.computers on Tue Nov 10 10:37:49 2026
    From Newsgroup: comp.os.linux.misc

    c186282 <c186282@nnada.net> wrote:
    On 11/9/26 16:20, John Levine wrote:
    According to c186282 <c186282@nnada.net>:
    Noted in this group though - I/O arch diffs between
    mainframes and minis.

    Will someone try to elaborate ?

    The classic difference is that mainframes have channels, so the
    processor sets up a bunch of work for the channel to do, the channel
    goes off and does it independently of the CPU. Minis had much less
    complex I/O controllers so they do more detailed stuff. An example I
    gave a few messages back was for a card reader. A mainframe starts a
    channel program, it reads in a card or maybe seveal cards, and reports
    back with an interrupt. On a mini there was an I/O operation to start
    the card reader, an interrupt for each column and an I/O to read it,
    and then something at the end to stop the reader or go to the next
    card.

    But that was 40 or 50 years ago. These days even cheap generic disk
    controllers do more than channels did in the 1960s, e.g., keep a queue
    of requests and schedule them to minimize delay, and handle spare block
    replacement for bad blocks.

    Modern mainframes have extremely capable channels but the distinction is
    more about reliability and things like hardware and software upgrades in
    place without rebooting.

    Ok ... so 'mainframes' as you saw them were more
    'directors' for distributed processing rather
    than THE processor.

    No. The operating system and application software ran on the mainframe
    CPU. But - as described above - a key design aspect of mainframes back
    then was that I/O was not done by the CPU (like it was done on home
    computers were all I/O was entirely handled by the CPU), but fully off-
    loaded to peripherials that were essentially limited computers of their
    own:
    - CPU tells I/O interface (channel): "Do this I/O"
    - CPU goes back to computing, channel does some I/O heavy lifting
    - channel has completed I/O and sends an "Hey, I'm done" interrupt
    to the CPU

    Which leaves most of the CPU free to do compute work instead of busily
    doing lots of I/O. This and other design concepts (like form based
    terminals doing local editing - think web forms filled in the browser, send
    as a whole form and processed as a single request by the server, instead
    of the server handling every keystroke on the terminal) enabled having thousands of interactive users on one mainframe, since the mainframe CPU
    did almost entirely compute (e.g. run the database and software for
    inventory, accounting, ...) and all I/O was offloaded to specialized
    computers (channels) acting as I/O interfaces. You essentially pay with
    extra complexity for a far more capable system.

    Kind regards,
    Alex.
    --
    "Opportunity is missed by most people because it is dressed in overalls and
    looks like work." -- Thomas A. Edison
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Alexander Schreiber@als@usenet.thangorodrim.de to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Tue Nov 10 10:50:32 2026
    From Newsgroup: comp.os.linux.misc

    Lawrence DrCOOliveiro <ldo@nz.invalid> wrote:
    On Tue, 10 Nov 2026 07:09:17 -0000 (UTC), Thomas Koenig wrote:

    Today's mainframes can no longer compete on compute power, but they
    have extremely high reliability (you can literally fire a bullet
    through a CPU, and things will migrate smoothly - if you don't hit
    the Recovery Unit) ...

    Cloud providers like Amazon, Google, Microsoft etc offer much higher reliability through massive parallelism, based on cheap, yet highly
    redundant hardware running a Free Software stack that is not priced
    per CPU. That is much more cost-effective and any equivalent number of mainframes. And there is no rCLRecovery UnitrCY that you can shoot at to
    take it out.

    The nice thing is, you can run the same sort of rCLenterprise-classrCY
    setup that the big folks operate in your own homelab, because all the foundation software they use is available to you, too
    <https://docs.xcp-ng.org/vms/vm-migration/>.

    Not exactly the same. The mainframe design philosophy includes "The
    machine MUST NOT fail", for which one pays in complexity and system
    cost. As long as your workload fits on that single machine (or a
    small cluster (Parallel Sysplex)) and is of the "MUST NOT fail"
    nature, fine.

    Highly distributed systems like those run and offered by the hyperscalers
    are built on an entirely different design philosophy:
    - _everything_ fails all the time (law of big numbers: take the MTBF
    for a compute component, run a lot of them and you've got a steady
    rain of failures, e.g. say you've got an MTBF of 1 million hours and
    you run 2 million of these devices, then on average you'll have a
    failure every half hour)
    - deal with it
    - run a software stack that is highly distributed and highly redundant,
    e.g. do not just store one or two copies of data, store 4 copies
    of the data plus two copies of the recovery checksums so you still
    recover the data after losing some pieces
    - you will end up with impressive complexity because building a reliable
    system on top of unreliable components is not easy
    - it helps a lot if your workloads can tolerate small to tiny amounts
    of failure (e.g. loading a web page times out, uploading a file fails)
    - but reliable operations (such as a planet scale consistent database)
    are doable, they just need more work and smarts
    - one key aspect: well designed highly distributed system scale very
    well, a scale-up factor of 10 without any system changes aside from
    adding more hardware falls under "baseline capability" and much higher
    scale-ups are the norm

    Kind regards,
    Alex.
    --
    "Opportunity is missed by most people because it is dressed in overalls and
    looks like work." -- Thomas A. Edison
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Nuno Silva@nunojsilva@invalid.invalid to comp.os.linux.misc,alt.folklore.computers on Tue Nov 10 10:45:40 2026
    From Newsgroup: comp.os.linux.misc

    [ Removing the shell group that, at least by now, is not relevant
    anymore. Not sure where to set Followup-To, though... maybe
    alt.unix.geeks? ]

    On 2026-11-10, Lawrence DrCOOliveiro wrote:

    On Tue, 10 Nov 2026 07:09:17 -0000 (UTC), Thomas Koenig wrote:

    Today's mainframes can no longer compete on compute power, but they
    have extremely high reliability (you can literally fire a bullet
    through a CPU, and things will migrate smoothly - if you don't hit
    the Recovery Unit) ...

    Cloud providers like Amazon, Google, Microsoft etc offer much higher reliability through massive parallelism, based on cheap, yet highly
    redundant hardware running a Free Software stack that is not priced
    per CPU. That is much more cost-effective and any equivalent number of mainframes. And there is no rCLRecovery UnitrCY that you can shoot at to
    take it out.

    The description by Thomas does make it look mainframes are capable of
    much more fine-grained fail-safe operation.

    There has to be *something* coordinating "recovery" to achieve a similar
    level of fail-safe operation with the stack you describe. Entirely
    possible, but unless you consider fail-safe "this machine has failed,
    here's another one" with no recovery of what was being done, it's
    likely not really comparable?

    The nice thing is, you can run the same sort of rCLenterprise-classrCY
    setup that the big folks operate in your own homelab, because all the foundation software they use is available to you, too <https://docs.xcp-ng.org/vms/vm-migration/>.
    --
    Nuno Silva
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Pancho@Pancho.Jones@protonmail.com to comp.os.linux.misc,alt.folklore.computers on Tue Nov 10 12:22:20 2026
    From Newsgroup: comp.os.linux.misc

    On 11/10/26 09:37, Alexander Schreiber wrote:

    No. The operating system and application software ran on the mainframe
    CPU. But - as described above - a key design aspect of mainframes back
    then was that I/O was not done by the CPU (like it was done on home
    computers were all I/O was entirely handled by the CPU), but fully off- loaded to peripherials that were essentially limited computers of their

    "Fully offloaded" is a bit weasel worded. I thought PCs did offload some duties to io devices, nic chip, wifi chip, gpu card. sound card. Also,
    even a mainframe will need service routines to handle communications
    with IO devices. Where to draw the line seems unclear.

    Caveat: I don't have a clue about these things, it is just that what you
    are saying doesn't sound quite right.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Peter Flass@Peter@Iron-Spring.com to comp.os.linux.misc,alt.folklore.computers on Tue Nov 10 07:35:59 2026
    From Newsgroup: comp.os.linux.misc

    On 11/10/26 02:37, Alexander Schreiber wrote:
    c186282 <c186282@nnada.net> wrote:
    On 11/9/26 16:20, John Levine wrote:
    According to c186282 <c186282@nnada.net>:
    Noted in this group though - I/O arch diffs between
    mainframes and minis.

    Will someone try to elaborate ?

    The classic difference is that mainframes have channels, so the
    processor sets up a bunch of work for the channel to do, the channel
    goes off and does it independently of the CPU. Minis had much less
    complex I/O controllers so they do more detailed stuff. An example I
    gave a few messages back was for a card reader. A mainframe starts a
    channel program, it reads in a card or maybe seveal cards, and reports
    back with an interrupt. On a mini there was an I/O operation to start
    the card reader, an interrupt for each column and an I/O to read it,
    and then something at the end to stop the reader or go to the next
    card.

    But that was 40 or 50 years ago. These days even cheap generic disk
    controllers do more than channels did in the 1960s, e.g., keep a queue
    of requests and schedule them to minimize delay, and handle spare block
    replacement for bad blocks.

    Modern mainframes have extremely capable channels but the distinction is >>> more about reliability and things like hardware and software upgrades in >>> place without rebooting.

    Ok ... so 'mainframes' as you saw them were more
    'directors' for distributed processing rather
    than THE processor.

    No. The operating system and application software ran on the mainframe
    CPU. But - as described above - a key design aspect of mainframes back
    then was that I/O was not done by the CPU (like it was done on home
    computers were all I/O was entirely handled by the CPU), but fully off- loaded to peripherials that were essentially limited computers of their
    own:
    - CPU tells I/O interface (channel): "Do this I/O"
    - CPU goes back to computing, channel does some I/O heavy lifting
    - channel has completed I/O and sends an "Hey, I'm done" interrupt
    to the CPU


    A prime example of neat code is that HASP (now JES) printed an entire
    page with one channel program instead of a line at a time. It also read
    a number of cards with one channel program instead of one card at a
    time, or, as someone has mentioned, one card column at a time.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Peter Flass@Peter@Iron-Spring.com to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Tue Nov 10 07:37:44 2026
    From Newsgroup: comp.os.linux.misc

    On 11/10/26 02:50, Alexander Schreiber wrote:
    Lawrence DrCOOliveiro <ldo@nz.invalid> wrote:
    On Tue, 10 Nov 2026 07:09:17 -0000 (UTC), Thomas Koenig wrote:

    Today's mainframes can no longer compete on compute power, but they
    have extremely high reliability (you can literally fire a bullet
    through a CPU, and things will migrate smoothly - if you don't hit
    the Recovery Unit) ...

    Cloud providers like Amazon, Google, Microsoft etc offer much higher
    reliability through massive parallelism, based on cheap, yet highly
    redundant hardware running a Free Software stack that is not priced
    per CPU. That is much more cost-effective and any equivalent number of
    mainframes. And there is no rCLRecovery UnitrCY that you can shoot at to
    take it out.

    The nice thing is, you can run the same sort of rCLenterprise-classrCY
    setup that the big folks operate in your own homelab, because all the
    foundation software they use is available to you, too
    <https://docs.xcp-ng.org/vms/vm-migration/>.

    Not exactly the same. The mainframe design philosophy includes "The
    machine MUST NOT fail", for which one pays in complexity and system
    cost. As long as your workload fits on that single machine (or a
    small cluster (Parallel Sysplex)) and is of the "MUST NOT fail"
    nature, fine.

    Highly distributed systems like those run and offered by the hyperscalers
    are built on an entirely different design philosophy:
    - _everything_ fails all the time (law of big numbers: take the MTBF
    for a compute component, run a lot of them and you've got a steady
    rain of failures, e.g. say you've got an MTBF of 1 million hours and
    you run 2 million of these devices, then on average you'll have a
    failure every half hour)

    Didn't I read somewhere that techs in large datacenters get around on rollerblades to swap out failing components?

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From scott@scott@slp53.sl.home (Scott Lurndal) to comp.os.linux.misc,alt.folklore.computers on Tue Nov 10 15:14:56 2026
    From Newsgroup: comp.os.linux.misc

    Pancho <Pancho.Jones@protonmail.com> writes:
    On 11/10/26 09:37, Alexander Schreiber wrote:

    No. The operating system and application software ran on the mainframe
    CPU. But - as described above - a key design aspect of mainframes back
    then was that I/O was not done by the CPU (like it was done on home
    computers were all I/O was entirely handled by the CPU), but fully off-
    loaded to peripherials that were essentially limited computers of their

    "Fully offloaded" is a bit weasel worded. I thought PCs did offload some >duties to io devices, nic chip, wifi chip, gpu card. sound card. Also,
    even a mainframe will need service routines to handle communications
    with IO devices. Where to draw the line seems unclear.

    Fully offloaded is accurate. For the Burroughs mainframes, a single instruction -initiated- an I/O request to a channel. The IIO (later
    SPIO) instruction would instruct the I/O processor (in the late 70s
    the IOP used an 8085 uProcessor) to initiate an I/O by passing a
    6-digit (24-bit) opcode/variant word and the start and end addresses
    of the buffer. The IOP would pass the command to a Data-Link Processor
    (DLP) that handled communications to the device controller (magnetic
    tape, disk) or unit record device (printer, card reader, card punch)
    or external communications processor (RS232/20ma current loop) and wait
    for the controller to start passing data back to the IOP. The IOP
    would move the data to the appropriate memory buffer and, when complete,
    signal the processor that the I/O had completed with an interrupt.


    Caveat: I don't have a clue about these things, it is just that what you
    are saying doesn't sound quite right.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From scott@scott@slp53.sl.home (Scott Lurndal) to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Tue Nov 10 15:16:24 2026
    From Newsgroup: comp.os.linux.misc

    Thomas Koenig <tkoenig@netcologne.de> writes:
    c186282 <c186282@nnada.net> schrieb:

    Noted in this group though - I/O arch diffs between
    mainframes and minis.

    I have a somewhat personal definition of a mainframe, which I also
    includes the classical ones.

    A mainframe is a computer that

    - is very expensive
    - takes large resources to run (a computer center)
    - has very good performance in at least one field
    - is not a cluster, massively parallel etc

    Some mainframes were clustered. Burroughs, for example,
    could share storage between up to four independent hosts;
    with shared access to files (with hardware record lockout
    for shared updates).

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From scott@scott@slp53.sl.home (Scott Lurndal) to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Tue Nov 10 15:18:28 2026
    From Newsgroup: comp.os.linux.misc

    Peter Flass <Peter@Iron-Spring.com> writes:
    On 11/10/26 02:50, Alexander Schreiber wrote:
    Lawrence DrCOOliveiro <ldo@nz.invalid> wrote:
    On Tue, 10 Nov 2026 07:09:17 -0000 (UTC), Thomas Koenig wrote:

    Today's mainframes can no longer compete on compute power, but they
    have extremely high reliability (you can literally fire a bullet
    through a CPU, and things will migrate smoothly - if you don't hit
    the Recovery Unit) ...

    Cloud providers like Amazon, Google, Microsoft etc offer much higher
    reliability through massive parallelism, based on cheap, yet highly
    redundant hardware running a Free Software stack that is not priced
    per CPU. That is much more cost-effective and any equivalent number of
    mainframes. And there is no rCLRecovery UnitrCY that you can shoot at to >>> take it out.

    The nice thing is, you can run the same sort of rCLenterprise-classrCY
    setup that the big folks operate in your own homelab, because all the
    foundation software they use is available to you, too
    <https://docs.xcp-ng.org/vms/vm-migration/>.

    Not exactly the same. The mainframe design philosophy includes "The
    machine MUST NOT fail", for which one pays in complexity and system
    cost. As long as your workload fits on that single machine (or a
    small cluster (Parallel Sysplex)) and is of the "MUST NOT fail"
    nature, fine.

    Highly distributed systems like those run and offered by the hyperscalers
    are built on an entirely different design philosophy:
    - _everything_ fails all the time (law of big numbers: take the MTBF
    for a compute component, run a lot of them and you've got a steady
    rain of failures, e.g. say you've got an MTBF of 1 million hours and
    you run 2 million of these devices, then on average you'll have a
    failure every half hour)

    Didn't I read somewhere that techs in large datacenters get around on >rollerblades to swap out failing components?


    The Minneapolis Fed Reserve operators were reputed to have worn
    roller skates when sorting checks (the sorters, with full complement
    of pockets, were quite long).
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lynn Wheeler@lynn@garlic.com to comp.os.linux.misc,alt.folklore.computers on Tue Nov 10 07:29:52 2026
    From Newsgroup: comp.os.linux.misc

    Alexander Schreiber <als@usenet.thangorodrim.de> writes:
    No. The operating system and application software ran on the mainframe
    CPU. But - as described above - a key design aspect of mainframes back
    then was that I/O was not done by the CPU (like it was done on home
    computers were all I/O was entirely handled by the CPU), but fully off- loaded to peripherials that were essentially limited computers of their
    own:
    - CPU tells I/O interface (channel): "Do this I/O"
    - CPU goes back to computing, channel does some I/O heavy lifting
    - channel has completed I/O and sends an "Hey, I'm done" interrupt
    to the CPU

    Original 360, there was limited processor and storage, but a lof of
    excess channel capacity ... as a result things like finding data was
    offloaded to channels ... which would "search" for desired data ... and
    as a result, channels ran half-duplex ... with channel dedicated to
    single operation ... constantly pulling search argument from processor
    storage.

    Move into 80s, storage got lot more abundant ... and channel
    architecture increasingly channel I/O increasingly became
    system throughput bottleneck.

    Early 80s, I wrote tome that 60s era 360/67 could get 70-80 CP67/CMS
    users on system ... and early 80s, with IBM processors getting 40-50
    times faster and storage increasing 40-50 larger, should be able be see 3000-4000 users, but was only seeing 300-400 users (because disks &
    channels throughput only increased 3-5 times (partly because half-duplex operation increasingly played a latency role in fully being able to see benefits of any speed increase).

    A IBM disk division executive took exception and assigned the division performance group to refute the claim. After a couple weeks, they came
    back and said I had slightly understated the issue.

    1988, one of issues of being able to see significant increase in
    throughput with LLNL serial stuff ... becoming fibre-channel standard
    was running full-duplex being able to concurrently transmit and receive
    on 1gbit speed links (w/o having to wait for full round-trip
    hand-shaking).

    Note: 1980, IBM STL (since renamed SVL) was bursting at seams and 300 people/terminals from IMS group were being moved to offsite bldg, with dataprocessing service back to STL datacenter. They had tried "remote
    3270" but found human factors totally unacceptable.

    I get con'ed into doing channel extender support, allowing placing 3270 controllers at offsite bldg with no perceived difference in human
    factors. Unanticipated side-effect was the associated systems'
    throughput increase 10-15 percent. STL had been speading 3270
    controllers across all channels with 3330 drives. The channel-extenders (operating full-duplex) significantly reduced the channel busy
    interference with the 3330 drives (for same amount of 3270 traffic;
    there was some consideration moving all 3270 controllers to
    channel-extenders).
    --
    virtualization experience starting Jan1968, online at home since Mar1970
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lynn Wheeler@lynn@garlic.com to comp.os.linux.misc,alt.folklore.computers on Tue Nov 10 07:55:29 2026
    From Newsgroup: comp.os.linux.misc

    Lynn Wheeler <lynn@garlic.com> writes:
    Note: 1980, IBM STL (since renamed SVL) was bursting at seams and 300 people/terminals from IMS group were being moved to offsite bldg, with dataprocessing service back to STL datacenter. They had tried "remote
    3270" but found human factors totally unacceptable.

    I get con'ed into doing channel extender support, allowing placing 3270 controllers at offsite bldg with no perceived difference in human
    factors. Unanticipated side-effect was the associated systems'
    throughput increase 10-15 percent. STL had been speading 3270
    controllers across all channels with 3330 drives. The channel-extenders (operating full-duplex) significantly reduced the channel busy
    interference with the 3330 drives (for same amount of 3270 traffic;
    there was some consideration moving all 3270 controllers to channel-extenders).

    Note once the 3033 was out the door, the processor engineers start on trout/3090. They configure memory, processor, channels, I/O to meet
    specific throughput goals. However, in San Jose, the administration
    dictated that the 3880 disk controller would use a (really slow,
    resulting in significant channel busy) vertical CISC processor (much
    slower than the previous processor in 3830 controller).

    The 3090 people thought the 3880 would be the same as 3830, but
    supporting 3380 3mbyte/sec transfer, when they found out how bad the
    3880 controllers' channel busy was (there is significantly amount of half-cuplex channel protocol back&forth undernearth the covers with
    latency playing increasing role), they realized they would have to significantly increase the number of I/O channels in order to achieve
    system throughput targets ... which required an additional TCM (3090
    said semi-facetiously they would bill the 3880 organization for the
    increase in 3090 manufacturing cost). Marketing eventually comes up with
    what a fantastic I/O machine the 3090 is with big increase in number of channels (but it was actually role played by slower 3880 controller and increase in channel busy associated with protocol chatter).
    --
    virtualization experience starting Jan1968, online at home since Mar1970
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Charlie Gibbs@cgibbs@kltpzyxm.invalid to comp.os.linux.misc,alt.folklore.computers on Tue Nov 10 18:36:48 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-10, Alexander Schreiber <als@usenet.thangorodrim.de> wrote:

    c186282 <c186282@nnada.net> wrote:

    Ok ... so 'mainframes' as you saw them were more
    'directors' for distributed processing rather
    than THE processor.

    No. The operating system and application software ran on the mainframe
    CPU. But - as described above - a key design aspect of mainframes back
    then was that I/O was not done by the CPU (like it was done on home
    computers were all I/O was entirely handled by the CPU), but fully off- loaded to peripherials that were essentially limited computers of their
    own:
    - CPU tells I/O interface (channel): "Do this I/O"
    - CPU goes back to computing, channel does some I/O heavy lifting
    - channel has completed I/O and sends an "Hey, I'm done" interrupt
    to the CPU

    Some minis offered an option for peripherals to have at least
    some of this kind of smarts. It was called "DMA" (direct memory
    access), since the peripherals directly accessed memory rather
    than having to pass data through the CPU. It was one of those
    gee-whiz things that you might want if you could justify the cost.

    In the mainframe world, you never heard the term "DMA".
    It was taken for granted.
    --
    /~\ Charlie Gibbs | In this world there are
    \ / <cgibbs@kltpzyxm.invalid> | two kinds of people:
    X I'm really at ac.dekanfrus | 1. Those who can extrapolate
    / \ if you read it the right way. | from incomplete data.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Charlie Gibbs@cgibbs@kltpzyxm.invalid to comp.os.linux.misc,alt.folklore.computers on Tue Nov 10 18:36:49 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-10, Peter Flass <Peter@Iron-Spring.com> wrote:

    A prime example of neat code is that HASP (now JES) printed an entire
    page with one channel program instead of a line at a time. It also read
    a number of cards with one channel program instead of one card at a
    time, or, as someone has mentioned, one card column at a time.

    The coolest channel program I ever saw was one I cribbed from a
    boot loader and enhanced a bit. You passed it a file name and
    a device address, and it would go out to the drive, find out
    where the VTOC (volume table of contents) lived, seek to it,
    and search for the format 1 label that described the file you
    were looking for. Mind you, it tied up the channel for the
    duration, but the machine I was working on didn't multitask,
    so it was a win for me.
    --
    /~\ Charlie Gibbs | In this world there are
    \ / <cgibbs@kltpzyxm.invalid> | two kinds of people:
    X I'm really at ac.dekanfrus | 1. Those who can extrapolate
    / \ if you read it the right way. | from incomplete data.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Charlie Gibbs@cgibbs@kltpzyxm.invalid to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Tue Nov 10 18:36:50 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-10, Thomas Koenig <tkoenig@netcologne.de> wrote:

    I read that, when minis came along, their first defining
    characteristic was that they didn't need computer center resources
    and fit in a single room.

    And you could plug them into a wall socket.
    --
    /~\ Charlie Gibbs | In this world there are
    \ / <cgibbs@kltpzyxm.invalid> | two kinds of people:
    X I'm really at ac.dekanfrus | 1. Those who can extrapolate
    / \ if you read it the right way. | from incomplete data.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Peter Flass@Peter@Iron-Spring.com to comp.os.linux.misc,alt.folklore.computers on Tue Nov 10 12:16:57 2026
    From Newsgroup: comp.os.linux.misc

    On 11/10/26 10:29, Lynn Wheeler wrote:
    Alexander Schreiber <als@usenet.thangorodrim.de> writes:
    No. The operating system and application software ran on the mainframe
    CPU. But - as described above - a key design aspect of mainframes back
    then was that I/O was not done by the CPU (like it was done on home
    computers were all I/O was entirely handled by the CPU), but fully off-
    loaded to peripherials that were essentially limited computers of their
    own:
    - CPU tells I/O interface (channel): "Do this I/O"
    - CPU goes back to computing, channel does some I/O heavy lifting
    - channel has completed I/O and sends an "Hey, I'm done" interrupt
    to the CPU

    Original 360, there was limited processor and storage, but a lof of
    excess channel capacity ... as a result things like finding data was offloaded to channels ... which would "search" for desired data ... and
    as a result, channels ran half-duplex ... with channel dedicated to
    single operation ... constantly pulling search argument from processor storage.

    Move into 80s, storage got lot more abundant ... and channel
    architecture increasingly channel I/O increasingly became
    system throughput bottleneck.

    Early 80s, I wrote tome that 60s era 360/67 could get 70-80 CP67/CMS
    users on system ... and early 80s, with IBM processors getting 40-50
    times faster and storage increasing 40-50 larger, should be able be see 3000-4000 users, but was only seeing 300-400 users (because disks &
    channels throughput only increased 3-5 times (partly because half-duplex operation increasingly played a latency role in fully being able to see benefits of any speed increase).

    With different resource constraints the system would have needed
    redesign to really make use of what it had - like ISAM (disk-intensive)
    became VSAM (more memory-intensive).
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Peter Flass@Peter@Iron-Spring.com to comp.os.linux.misc,alt.folklore.computers on Tue Nov 10 12:19:31 2026
    From Newsgroup: comp.os.linux.misc

    On 11/10/26 10:55, Lynn Wheeler wrote:


    Note once the 3033 was out the door, the processor engineers start on trout/3090. They configure memory, processor, channels, I/O to meet
    specific throughput goals. However, in San Jose, the administration
    dictated that the 3880 disk controller would use a (really slow,
    resulting in significant channel busy) vertical CISC processor (much
    slower than the previous processor in 3830 controller).

    The 3090 people thought the 3880 would be the same as 3830, but
    supporting 3380 3mbyte/sec transfer, when they found out how bad the


    Anyone would think that newer=faster. Who would have thought it would be slower?


    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lawrence =?iso-8859-13?q?D=FFOliveiro?=@ldo@nz.invalid to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Tue Nov 10 19:55:46 2026
    From Newsgroup: comp.os.linux.misc

    On Tue, 10 Nov 2026 10:50:32 +0100, Alexander Schreiber wrote:

    Lawrence DrCOOliveiro <ldo@nz.invalid> wrote:

    On Tue, 10 Nov 2026 07:09:17 -0000 (UTC), Thomas Koenig wrote:

    Today's mainframes can no longer compete on compute power, but
    they have extremely high reliability (you can literally fire a
    bullet through a CPU, and things will migrate smoothly - if you
    don't hit the Recovery Unit) ...

    Cloud providers like Amazon, Google, Microsoft etc offer much
    higher reliability through massive parallelism, based on cheap, yet
    highly redundant hardware running a Free Software stack that is not
    priced per CPU. That is much more cost-effective and any equivalent
    number of mainframes. And there is no rCLRecovery UnitrCY that you can
    shoot at to take it out.

    Not exactly the same. The mainframe design philosophy includes "The
    machine MUST NOT fail", for which one pays in complexity and system
    cost. As long as your workload fits on that single machine (or a
    small cluster (Parallel Sysplex)) and is of the "MUST NOT fail"
    nature, fine.

    - The daylight savings change.
    Once you get the GMT time and local offset set correctly the
    other problem is the daylight savings change. Unlike VM, MVS
    allows you to change the local time without an IPL so you should
    be able to make the daylight savings change on the fly. We even
    did it once in production. SMF data uses local time for time
    stamps, so it really got messed up. JES2 automatic commands also
    use local time and they were halted. There are probably other
    problems, but those were enough for us. We IPL for the daylight
    savings change.
    -- Dave Crow, rCLSetting a ComputerrCOs Clock Correctly: Does
    anybody know what time it is?rCY,
    presentation given to SHARE (IBM user group) in 1986
    <https://bitsavers.trailing-edge.com/pdf/ibm/share/SHARE67_Presentation_on_Setting_a_Computers_Clock_Correctly_1986.pdf>
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lawrence =?iso-8859-13?q?D=FFOliveiro?=@ldo@nz.invalid to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Tue Nov 10 19:57:28 2026
    From Newsgroup: comp.os.linux.misc

    On Tue, 10 Nov 2026 18:36:50 GMT, Charlie Gibbs wrote:

    On 2026-11-10, Thomas Koenig <tkoenig@netcologne.de> wrote:

    I read that, when minis came along, their first defining
    characteristic was that they didn't need computer center resources
    and fit in a single room.

    And you could plug them into a wall socket.

    Some kind of power conditioning would have been highly recommended.

    I can remember, during the 1980s, during the least little power
    brownout, our campusrCOs main VAX systems would instantly go down, and
    take about 15 minutes to come back up.

    Meanwhile, the PCs on our desks didnrCOt miss a beat.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Pancho@Pancho.Jones@protonmail.com to comp.os.linux.misc,alt.folklore.computers on Tue Nov 10 20:36:21 2026
    From Newsgroup: comp.os.linux.misc

    On 11/10/26 15:14, Scott Lurndal wrote:

    "Fully offloaded" is a bit weasel worded. I thought PCs did offload some
    duties to io devices, nic chip, wifi chip, gpu card. sound card. Also,
    even a mainframe will need service routines to handle communications
    with IO devices. Where to draw the line seems unclear.

    Fully offloaded is accurate. For the Burroughs mainframes, a single instruction -initiated- an I/O request to a channel. The IIO (later
    SPIO) instruction would instruct the I/O processor (in the late 70s
    the IOP used an 8085 uProcessor) to initiate an I/O by passing a
    6-digit (24-bit) opcode/variant word and the start and end addresses
    of the buffer. The IOP would pass the command to a Data-Link Processor (DLP) that handled communications to the device controller (magnetic
    tape, disk) or unit record device (printer, card reader, card punch)
    or external communications processor (RS232/20ma current loop) and wait
    for the controller to start passing data back to the IOP. The IOP
    would move the data to the appropriate memory buffer and, when complete, signal the processor that the I/O had completed with an interrupt.


    Yes, you can give me a lot of details, but I'm not seeing the
    fundamental difference between that and DMA IO devices that have been
    around since the first IBM PC in 1981.

    As I said I'm naive of this kind of low level stuff, but from my view as
    an application programmer I'm failing to see a fundamental difference
    from a PC, offloading an IO task to a non CPU device with communication
    via registers, interrupts and DMA.



    Caveat: I don't have a clue about these things, it is just that what you
    are saying doesn't sound quite right.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Pancho@Pancho.Jones@protonmail.com to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Tue Nov 10 20:51:47 2026
    From Newsgroup: comp.os.linux.misc

    On 11/9/26 10:01, Nuno Silva wrote:
    On 2026-11-09, Lawrence DrCOOliveiro wrote:

    On Sun, 8 Nov 2026 20:33:45 -0700, Peter Flass wrote:

    Mainframes typically achieve an uptime percentage between 99.999%
    (five nines) and 99.9999999% (nine nines)

    The hyperscalers probably manage that well, if not better.

    There is an article from 1986 on Bitsavers which describes having to
    reboot an IBM mainframe to turn daylight saving on and off.

    *That* sounds like complicating a problem that would be made simpler by
    not using the local time, or at least DST-adjusted local time.


    Yep, obvious really, UTC internally, local time only at the presentation layer.

    When apps were just one country, one timezone, you could kind of forgive
    it, but I was working on new multinational apps right up to the late
    noughties that got it wrong.


    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Alexander Schreiber@als@usenet.thangorodrim.de to comp.os.linux.misc,alt.folklore.computers on Tue Nov 10 21:53:26 2026
    From Newsgroup: comp.os.linux.misc

    Pancho <Pancho.Jones@protonmail.com> wrote:
    On 11/10/26 09:37, Alexander Schreiber wrote:

    No. The operating system and application software ran on the mainframe
    CPU. But - as described above - a key design aspect of mainframes back
    then was that I/O was not done by the CPU (like it was done on home
    computers were all I/O was entirely handled by the CPU), but fully off-
    loaded to peripherials that were essentially limited computers of their

    "Fully offloaded" is a bit weasel worded. I thought PCs did offload some duties to io devices, nic chip, wifi chip, gpu card. sound card. Also,
    even a mainframe will need service routines to handle communications
    with IO devices. Where to draw the line seems unclear.

    That came much later. Even in the 1990s, the typical PC had an IDE
    controller for storage (SCSI, which offloaded some work onto the controller
    was something found in servers, workstations and PCs of people who cared
    about such things), the NIC was a far cry from modern DPUs - there were
    even NICs that essentially consisted of the bus PHY, the Ethernet PHY and
    some minimal glue logic (doing even Ethernet framing in the driver running
    on the CPU). GPUs in the modern sense weren't there yet, but fancy graphics cards with some (mostly 2D, 3D was just getting started) acceleration did exist. And sound cards handled mostly AD/DA and some analog mixing and
    maybe synthesis.

    Caveat: I don't have a clue about these things, it is just that what you
    are saying doesn't sound quite right.

    PC tech has come a loooong way since then.

    Kind regards,
    Alex.
    --
    "Opportunity is missed by most people because it is dressed in overalls and
    looks like work." -- Thomas A. Edison
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Alexander Schreiber@als@usenet.thangorodrim.de to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Tue Nov 10 22:02:01 2026
    From Newsgroup: comp.os.linux.misc

    Peter Flass <Peter@Iron-Spring.com> wrote:
    On 11/10/26 02:50, Alexander Schreiber wrote:
    Lawrence DrCOOliveiro <ldo@nz.invalid> wrote:
    On Tue, 10 Nov 2026 07:09:17 -0000 (UTC), Thomas Koenig wrote:

    Today's mainframes can no longer compete on compute power, but they
    have extremely high reliability (you can literally fire a bullet
    through a CPU, and things will migrate smoothly - if you don't hit
    the Recovery Unit) ...

    Cloud providers like Amazon, Google, Microsoft etc offer much higher
    reliability through massive parallelism, based on cheap, yet highly
    redundant hardware running a Free Software stack that is not priced
    per CPU. That is much more cost-effective and any equivalent number of
    mainframes. And there is no rCLRecovery UnitrCY that you can shoot at to >>> take it out.

    The nice thing is, you can run the same sort of rCLenterprise-classrCY
    setup that the big folks operate in your own homelab, because all the
    foundation software they use is available to you, too
    <https://docs.xcp-ng.org/vms/vm-migration/>.

    Not exactly the same. The mainframe design philosophy includes "The
    machine MUST NOT fail", for which one pays in complexity and system
    cost. As long as your workload fits on that single machine (or a
    small cluster (Parallel Sysplex)) and is of the "MUST NOT fail"
    nature, fine.

    Highly distributed systems like those run and offered by the hyperscalers
    are built on an entirely different design philosophy:
    - _everything_ fails all the time (law of big numbers: take the MTBF
    for a compute component, run a lot of them and you've got a steady
    rain of failures, e.g. say you've got an MTBF of 1 million hours and
    you run 2 million of these devices, then on average you'll have a
    failure every half hour)

    Didn't I read somewhere that techs in large datacenters get around on rollerblades to swap out failing components?

    You don't want that in a data center, because just it increases the risk
    of injury even more. You already need aid stations around on the floor
    for when, not if, someone gets hurt (usually scrapes or shallow cuts by sharp-edged parts in the machines).

    Kind regards,
    Alex.
    --
    "Opportunity is missed by most people because it is dressed in overalls and
    looks like work." -- Thomas A. Edison
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From The Natural Philosopher@tnp@invalid.invalid to comp.os.linux.misc,alt.folklore.computers on Tue Nov 10 21:43:38 2026
    From Newsgroup: comp.os.linux.misc

    On 10/11/2026 09:37, Alexander Schreiber wrote:
    I/O was not done by the CPU (like it was done on home
    computers were all I/O was entirely handled by the CPU), but fully off- loaded to peripherials that were essentially limited computers of their
    own:
    I see you never used the serial cards with up to 32 terminals attached
    to a PC running SCO unix.

    They were essentially IO processors
    --
    Gun Control: The law that ensures that only criminals have guns.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From John Levine@johnl@taugh.com to comp.os.linux.misc,alt.folklore.computers on Tue Nov 10 22:07:36 2026
    From Newsgroup: comp.os.linux.misc

    According to Pancho <Pancho.Jones@protonmail.com>:
    On 11/10/26 09:37, Alexander Schreiber wrote:

    No. The operating system and application software ran on the mainframe
    CPU. But - as described above - a key design aspect of mainframes back
    then was that I/O was not done by the CPU (like it was done on home
    computers were all I/O was entirely handled by the CPU), but fully off-
    loaded to peripherials that were essentially limited computers of their

    "Fully offloaded" is a bit weasel worded.

    It varied a lot. The large 360s had channels that were separate computers sharing memory with the CPU. The 360/30 had one microcode engine that both implemented the CPU and the channel. When it did disk I/O the CPU pretty much stopped because the channel had a lot to do. One of the impressive parts of the 360's architecture is that you could run the same code on large and small systems and it would work so long as it didn't make assumptions about how long things would take.

    I thought PCs did offload some duties to io devices, nic chip, wifi chip, gpu card. sound card.

    Now they do. In the 1970s not so much.
    --
    Regards,
    John Levine, johnl@taugh.com, Primary Perpetrator of "The Internet for Dummies",
    Please consider the environment before reading this e-mail. https://jl.ly
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From John Levine@johnl@taugh.com to comp.os.linux.misc,alt.folklore.computers on Tue Nov 10 22:14:16 2026
    From Newsgroup: comp.os.linux.misc

    According to Charlie Gibbs <cgibbs@kltpzyxm.invalid>:
    Some minis offered an option for peripherals to have at least
    some of this kind of smarts. It was called "DMA" (direct memory
    access), ...

    DEC minis called it "data break". There was one-cycle data break
    where the peripheral controller provided the address to transfer
    and kept track of the count. And there was three-cycle data
    break, where the peripheral provided a fixed memory address,
    the CPU used that address as a data pointer which it incremented,
    the address+1 as a counter which it decremented, then did the
    tranfer and sent a signal saying whether the count was zero.

    Data break controllers were expensive so they only used them
    for devices like disks and drums where programmed I/O wasn't
    fast enough. DECTape, a block addressable tape which was
    the spritual forerunner of floppy disks, originally had
    a PDP-8 three-cycle data break controller, but later on
    there was a cheaper controller that had a little bit
    of buffering that made programmed I/O transfers fast
    enough.

    Keep in mind that data break or DMA was only part of
    what a channel did. The channel also did device control,
    like disk head seeks and track selection, and could
    chain multiple operations togehter, like the example
    someone gave of a channel program that printed an
    entire page of output.
    --
    Regards,
    John Levine, johnl@taugh.com, Primary Perpetrator of "The Internet for Dummies",
    Please consider the environment before reading this e-mail. https://jl.ly
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From John Levine@johnl@taugh.com to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Tue Nov 10 22:19:32 2026
    From Newsgroup: comp.os.linux.misc

    According to Peter Flass <Peter@Iron-Spring.com>:
    Didn't I read somewhere that techs in large datacenters get around on >rollerblades to swap out failing components?

    I don't think so. I've read that Google has people wandering around with grocery carts full of spare parts that they swap into failed servers.
    You don't want them banging into the racks.

    At the Western Union building at 60 Hudson St in NYC, they did manual
    messge switching, in which one teletype punched out a message on paper
    tape, they tore it off, then girls (we'd call them young women now)
    on skates zipped the tapes to other teletypes to send the message
    onward.

    I gather that one of the earliest large scale uses of minicomputers
    was to automate that process.
    --
    Regards,
    John Levine, johnl@taugh.com, Primary Perpetrator of "The Internet for Dummies",
    Please consider the environment before reading this e-mail. https://jl.ly
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lawrence =?iso-8859-13?q?D=FFOliveiro?=@ldo@nz.invalid to comp.os.linux.misc,alt.folklore.computers on Tue Nov 10 22:44:10 2026
    From Newsgroup: comp.os.linux.misc

    On Tue, 10 Nov 2026 10:37:49 +0100, Alexander Schreiber wrote:

    ... a key design aspect of mainframes back then was that I/O was not
    done by the CPU (like it was done on home computers were all I/O was
    entirely handled by the CPU), but fully off- loaded to peripherials
    that were essentially limited computers of their own ...

    Which made sense back when the CPU was a scarce and expensive resource.
    It no longer made sense when that was no longer true.

    Which leaves most of the CPU free to do compute work instead of
    busily doing lots of I/O. This and other design concepts (like form
    based terminals doing local editing - think web forms filled in the
    browser, send as a whole form and processed as a single request by
    the server, instead of the server handling every keystroke on the
    terminal) enabled having thousands of interactive users on one
    mainframe ...

    Nowadays we have millions of users on social media simultaneously, but
    that request/response paradigm is no longer sufficient for the
    real-time interaction requirements.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lawrence =?iso-8859-13?q?D=FFOliveiro?=@ldo@nz.invalid to comp.os.linux.misc,alt.folklore.computers on Tue Nov 10 22:45:34 2026
    From Newsgroup: comp.os.linux.misc

    On Tue, 10 Nov 2026 07:35:59 -0700, Peter Flass wrote:

    A prime example of neat code is that HASP (now JES) printed an
    entire page with one channel program instead of a line at a time.

    ThatrCOs how PostScript printers worked: a whole bunch of commands to
    render the page image, then a rCLshowpagerCY to actually print it out.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Charlie Gibbs@cgibbs@kltpzyxm.invalid to comp.os.linux.misc,alt.folklore.computers on Tue Nov 10 23:09:18 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-10, Peter Flass <Peter@Iron-Spring.com> wrote:

    On 11/10/26 10:55, Lynn Wheeler wrote:

    Note once the 3033 was out the door, the processor engineers start on
    trout/3090. They configure memory, processor, channels, I/O to meet
    specific throughput goals. However, in San Jose, the administration
    dictated that the 3880 disk controller would use a (really slow,
    resulting in significant channel busy) vertical CISC processor (much
    slower than the previous processor in 3830 controller).

    The 3090 people thought the 3880 would be the same as 3830, but
    supporting 3380 3mbyte/sec transfer, when they found out how bad the

    Anyone would think that newer=faster. Who would have thought it would be slower?

    Welcome to the 21st century. How long does it take your TV to boot up?
    (I refer to this as "warming up", as a tribute to the tube sets of old.)
    --
    /~\ Charlie Gibbs | In this world there are
    \ / <cgibbs@kltpzyxm.invalid> | two kinds of people:
    X I'm really at ac.dekanfrus | 1. Those who can extrapolate
    / \ if you read it the right way. | from incomplete data.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Charlie Gibbs@cgibbs@kltpzyxm.invalid to comp.os.linux.misc,alt.folklore.computers on Tue Nov 10 23:09:19 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-10, Pancho <Pancho.Jones@protonmail.com> wrote:

    Yes, you can give me a lot of details, but I'm not seeing the
    fundamental difference between that and DMA IO devices that have been
    around since the first IBM PC in 1981.

    Go back another 10 years...
    --
    /~\ Charlie Gibbs | In this world there are
    \ / <cgibbs@kltpzyxm.invalid> | two kinds of people:
    X I'm really at ac.dekanfrus | 1. Those who can extrapolate
    / \ if you read it the right way. | from incomplete data.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From c186282@c186282@nnada.net to comp.os.linux.misc,alt.folklore.computers on Tue Nov 10 21:23:22 2026
    From Newsgroup: comp.os.linux.misc

    On 11/10/26 07:22, Pancho wrote:
    On 11/10/26 09:37, Alexander Schreiber wrote:

    No. The operating system and application software ran on the mainframe
    CPU. But - as described above - a key design aspect of mainframes back
    then was that I/O was not done by the CPU (like it was done on home
    computers were all I/O was entirely handled by the CPU), but fully off-
    loaded to peripherials that were essentially limited computers of their

    "Fully offloaded" is a bit weasel worded. I thought PCs did offload some duties to io devices, nic chip, wifi chip, gpu card. sound card. Also,
    even a mainframe will need service routines to handle communications
    with IO devices. Where to draw the line seems unclear.

    Caveat: I don't have a clue about these things, it is just that what you
    are saying doesn't sound quite right.

    Well, nothing is SIMPLE here .....

    Kind of why I asked how people defined "mainframe"
    vs "mini".

    No CLEAR line apparently.

    And the line MOVES.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From c186282@c186282@nnada.net to comp.os.linux.misc,alt.folklore.computers on Tue Nov 10 21:42:41 2026
    From Newsgroup: comp.os.linux.misc

    On 11/10/26 04:37, Alexander Schreiber wrote:
    c186282 <c186282@nnada.net> wrote:
    On 11/9/26 16:20, John Levine wrote:
    According to c186282 <c186282@nnada.net>:
    Noted in this group though - I/O arch diffs between
    mainframes and minis.

    Will someone try to elaborate ?

    The classic difference is that mainframes have channels, so the
    processor sets up a bunch of work for the channel to do, the channel
    goes off and does it independently of the CPU. Minis had much less
    complex I/O controllers so they do more detailed stuff. An example I
    gave a few messages back was for a card reader. A mainframe starts a
    channel program, it reads in a card or maybe seveal cards, and reports
    back with an interrupt. On a mini there was an I/O operation to start
    the card reader, an interrupt for each column and an I/O to read it,
    and then something at the end to stop the reader or go to the next
    card.

    But that was 40 or 50 years ago. These days even cheap generic disk
    controllers do more than channels did in the 1960s, e.g., keep a queue
    of requests and schedule them to minimize delay, and handle spare block
    replacement for bad blocks.

    Modern mainframes have extremely capable channels but the distinction is >>> more about reliability and things like hardware and software upgrades in >>> place without rebooting.

    Ok ... so 'mainframes' as you saw them were more
    'directors' for distributed processing rather
    than THE processor.

    No. The operating system and application software ran on the mainframe
    CPU. But - as described above - a key design aspect of mainframes back
    then was that I/O was not done by the CPU (like it was done on home
    computers were all I/O was entirely handled by the CPU), but fully off- loaded to peripherials that were essentially limited computers of their
    own:
    - CPU tells I/O interface (channel): "Do this I/O"
    - CPU goes back to computing, channel does some I/O heavy lifting
    - channel has completed I/O and sends an "Hey, I'm done" interrupt
    to the CPU

    Which leaves most of the CPU free to do compute work instead of busily
    doing lots of I/O. This and other design concepts (like form based
    terminals doing local editing - think web forms filled in the browser, send as a whole form and processed as a single request by the server, instead
    of the server handling every keystroke on the terminal) enabled having thousands of interactive users on one mainframe, since the mainframe CPU
    did almost entirely compute (e.g. run the database and software for inventory, accounting, ...) and all I/O was offloaded to specialized computers (channels) acting as I/O interfaces. You essentially pay with
    extra complexity for a far more capable system.

    Kind regards,
    Alex.

    I kind of see what model you're describing.
    Really not THAT much different from what
    I outlined.

    "Smart" peripherials are good. CPU doesn't have
    to do nearly as much. The downside is that now
    you're hostage to the 'IQ' in the peripherials
    rather than fine managing/optimizing via the
    main CPU.

    Always some trade-offs.

    At least these days the devices are more
    "standardized". Plug in any USB device and
    it almost always 'just works' and uses the
    same set-up/send/receive commands. NOT so
    in the Old Days.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Rich@rich@example.invalid to comp.os.linux.misc on Wed Nov 11 03:22:41 2026
    From Newsgroup: comp.os.linux.misc

    In comp.os.linux.misc Pancho <Pancho.Jones@protonmail.com> wrote:
    On 11/10/26 15:14, Scott Lurndal wrote:

    "Fully offloaded" is a bit weasel worded. I thought PCs did offload some >>> duties to io devices, nic chip, wifi chip, gpu card. sound card. Also,
    even a mainframe will need service routines to handle communications
    with IO devices. Where to draw the line seems unclear.

    Fully offloaded is accurate. For the Burroughs mainframes, a single
    instruction -initiated- an I/O request to a channel. The IIO (later
    SPIO) instruction would instruct the I/O processor (in the late 70s
    the IOP used an 8085 uProcessor) to initiate an I/O by passing a
    6-digit (24-bit) opcode/variant word and the start and end addresses
    of the buffer. The IOP would pass the command to a Data-Link Processor
    (DLP) that handled communications to the device controller (magnetic
    tape, disk) or unit record device (printer, card reader, card punch)
    or external communications processor (RS232/20ma current loop) and wait
    for the controller to start passing data back to the IOP. The IOP
    would move the data to the appropriate memory buffer and, when complete,
    signal the processor that the I/O had completed with an interrupt.


    Yes, you can give me a lot of details, but I'm not seeing the
    fundamental difference between that and DMA IO devices that have been
    around since the first IBM PC in 1981.

    The largest difference is the DMA in the IBM PC was a simple sequencer
    chip. It was programmed with IO port, memory start address, and count
    of bytes/words to move, and it did the transfer. Mainframe IO channels utilized full blown programs for general purpose CPU's to do the same
    thing the DMA chip did (simple move of data between device and memory)
    but the channel programs could be setup to do much more (such as
    translation between internal and external character sets, byte
    swapping, etc.)

    Of course, reality was that probably a majority of channel programs
    were doing little more than the same function as the DMA controller in
    an IBM PC.

    As I said I'm naive of this kind of low level stuff, but from my view as
    an application programmer I'm failing to see a fundamental difference
    from a PC, offloading an IO task to a non CPU device with communication
    via registers, interrupts and DMA.

    The big difference is instead of setting registers, you provided a full "program" (branching, loops, if/then/else, etc.) for how to perform the
    IO. A very similar concept to display lists where one provided a
    "program" for how to draw a display on a CRT instead of letting a fixed sequencer chip stream data to the CRT.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lawrence =?iso-8859-13?q?D=FFOliveiro?=@ldo@nz.invalid to comp.os.linux.misc,alt.folklore.computers on Wed Nov 11 03:59:40 2026
    From Newsgroup: comp.os.linux.misc

    On Tue, 10 Nov 2026 12:19:31 -0700, Peter Flass wrote:

    Anyone would think that newer=faster. Who would have thought it
    would be slower?

    As another example, I recall it said that passenger airliners after
    the Boeing 747 actually have slightly lower cruising speeds. I guess
    itrCOs all part of the goal of saving fuel.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From c186282@c186282@nnada.net to comp.os.linux.misc,alt.folklore.computers on Tue Nov 10 23:25:43 2026
    From Newsgroup: comp.os.linux.misc

    On 11/10/26 09:35, Peter Flass wrote:
    On 11/10/26 02:37, Alexander Schreiber wrote:
    c186282 <c186282@nnada.net> wrote:
    On 11/9/26 16:20, John Levine wrote:
    According to c186282-a <c186282@nnada.net>:
    -a-a-a Noted in this group though - I/O arch diffs between
    -a-a-a mainframes and minis.

    -a-a-a Will someone try to elaborate ?

    The classic difference is that mainframes have channels, so the
    processor sets up a bunch of work for the channel to do, the channel
    goes off and does it independently of the CPU. Minis had much less
    complex I/O controllers so they do more detailed stuff. An example I
    gave a few messages back was for a card reader. A mainframe starts a
    channel program, it reads in a card or maybe seveal cards, and reports >>>> back with an interrupt. On a mini there was an I/O operation to start
    the card reader, an interrupt for each column and an I/O to read it,
    and then something at the end to stop the reader or go to the next
    card.

    But that was 40 or 50 years ago.-a These days even cheap generic disk
    controllers do more than channels did in the 1960s, e.g., keep a queue >>>> of requests and schedule them to minimize delay, and handle spare block >>>> replacement for bad blocks.

    Modern mainframes have extremely capable channels but the
    distinction is
    more about reliability and things like hardware and software
    upgrades in
    place without rebooting.

    -a-a-a Ok ... so 'mainframes' as you saw them were more
    -a-a-a 'directors' for distributed processing rather
    -a-a-a than THE processor.

    No. The operating system and application software ran on the mainframe
    CPU. But - as described above - a key design aspect of mainframes back
    then was that I/O was not done by the CPU (like it was done on home
    computers were all I/O was entirely handled by the CPU), but fully off-
    loaded to peripherials that were essentially limited computers of their
    own:
    -a - CPU tells I/O interface (channel): "Do this I/O"
    -a - CPU goes back to computing, channel does some I/O heavy lifting
    -a - channel has completed I/O and sends an "Hey, I'm done" interrupt
    -a-a-a to the CPU


    A prime example of neat code is that HASP (now JES) printed an entire
    page with one channel program instead of a line at a time. It also read
    a number of cards with one channel program instead of one card at a
    time, or, as someone has mentioned, one card column at a time.

    "Smart" periphs are good ... but also
    a kind of trap.

    Ideally the CPU does almost everything - so you
    can optimize/tweak.

    But esp in the Olde Days the CPU really wasn't
    that strong or had that many resources.



    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From c186282@c186282@nnada.net to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Tue Nov 10 23:29:31 2026
    From Newsgroup: comp.os.linux.misc

    On 11/10/26 09:37, Peter Flass wrote:
    On 11/10/26 02:50, Alexander Schreiber wrote:
    Lawrence DrCOOliveiro <ldo@nz.invalid> wrote:
    On Tue, 10 Nov 2026 07:09:17 -0000 (UTC), Thomas Koenig wrote:

    Today's mainframes can no longer compete on compute power, but they
    have extremely high reliability (you can literally fire a bullet
    through a CPU, and things will migrate smoothly - if you don't hit
    the Recovery Unit) ...

    Cloud providers like Amazon, Google, Microsoft etc offer much higher
    reliability through massive parallelism, based on cheap, yet highly
    redundant hardware running a Free Software stack that is not priced
    per CPU. That is much more cost-effective and any equivalent number of
    mainframes. And there is no rCLRecovery UnitrCY that you can shoot at to >>> take it out.

    The nice thing is, you can run the same sort of rCLenterprise-classrCY
    setup that the big folks operate in your own homelab, because all the
    foundation software they use is available to you, too
    <https://docs.xcp-ng.org/vms/vm-migration/>.

    Not exactly the same. The mainframe design philosophy includes "The
    machine MUST NOT fail", for which one pays in complexity and system
    cost. As long as your workload fits on that single machine (or a
    small cluster (Parallel Sysplex)) and is of the "MUST NOT fail"
    nature, fine.

    Highly distributed systems like those run and offered by the hyperscalers
    are built on an entirely different design philosophy:
    -a - _everything_ fails all the time (law of big numbers: take the MTBF
    -a-a-a for a compute component, run a lot of them and you've got a steady
    -a-a-a rain of failures, e.g. say you've got an MTBF of 1 million hours and >> -a-a-a you run 2 million of these devices, then on average you'll have a
    -a-a-a failure every half hour)

    Didn't I read somewhere that techs in large datacenters get around on rollerblades to swap out failing components?

    IS an issue !!!

    Linux/UNIX are 'better' at this.

    As for 'failures every half hour' ... can't
    HAVE that !


    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From c186282@c186282@nnada.net to comp.os.linux.misc,alt.folklore.computers on Tue Nov 10 23:46:08 2026
    From Newsgroup: comp.os.linux.misc

    On 11/10/26 10:14, Scott Lurndal wrote:
    Pancho <Pancho.Jones@protonmail.com> writes:
    On 11/10/26 09:37, Alexander Schreiber wrote:

    No. The operating system and application software ran on the mainframe
    CPU. But - as described above - a key design aspect of mainframes back
    then was that I/O was not done by the CPU (like it was done on home
    computers were all I/O was entirely handled by the CPU), but fully off-
    loaded to peripherials that were essentially limited computers of their

    "Fully offloaded" is a bit weasel worded. I thought PCs did offload some
    duties to io devices, nic chip, wifi chip, gpu card. sound card. Also,
    even a mainframe will need service routines to handle communications
    with IO devices. Where to draw the line seems unclear.

    Fully offloaded is accurate. For the Burroughs mainframes, a single instruction -initiated- an I/O request to a channel. The IIO (later
    SPIO) instruction would instruct the I/O processor (in the late 70s
    the IOP used an 8085 uProcessor) to initiate an I/O by passing a
    6-digit (24-bit) opcode/variant word and the start and end addresses
    of the buffer. The IOP would pass the command to a Data-Link Processor (DLP) that handled communications to the device controller (magnetic
    tape, disk) or unit record device (printer, card reader, card punch)
    or external communications processor (RS232/20ma current loop) and wait
    for the controller to start passing data back to the IOP. The IOP
    would move the data to the appropriate memory buffer and, when complete, signal the processor that the I/O had completed with an interrupt.

    Esp in the Bad Old Days, "smart perips" were VERY
    important. The CPU just sent a LITTLE stuff to them.
    Very CPU/RAM/Cycle efficient.

    Flip - left you hostage to the periph "IQ", often
    no ways to tweak/enhance it all.


    Caveat: I don't have a clue about these things, it is just that what you
    are saying doesn't sound quite right.

    Me ... SEEN it all - the plusses AND minuses.

    The somewhat older group members will have
    seen MORE of that.

    The newbies think NOTHING from the past can
    possibly inform them, pave the way.

    Horribly wrong ... but you'll never convince them.
    For them it's now all "AI" ... until it displaces
    them entirely.

    Can ya say "Would You Like Fries With That ?" in
    three or four languages ??? :-)

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Nuno Silva@nunojsilva@invalid.invalid to comp.os.linux.misc,alt.folklore.computers on Wed Nov 11 09:08:10 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-11, Lawrence DrCOOliveiro wrote:

    On Tue, 10 Nov 2026 12:19:31 -0700, Peter Flass wrote:

    Anyone would think that newer=faster. Who would have thought it
    would be slower?

    As another example, I recall it said that passenger airliners after
    the Boeing 747 actually have slightly lower cruising speeds. I guess
    itrCOs all part of the goal of saving fuel.

    It is possible if the 747 was inefficient. IIRC there's a sweet spot (or range?) for efficient fuel usage in subsonic jets.

    (For supersonic, IIRC Aerospatiale-BAC Concorde was actually quite
    efficient in cruise (compared to other supersonic aircraft), and I think
    the proposed redesign would even do away with the need for wet thrust (necessary for TO and transonic climb?).)
    --
    Nuno Silva
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Alexander Schreiber@als@usenet.thangorodrim.de to comp.os.linux.misc,alt.folklore.computers on Wed Nov 11 12:03:15 2026
    From Newsgroup: comp.os.linux.misc

    The Natural Philosopher <tnp@invalid.invalid> wrote:
    On 10/11/2026 09:37, Alexander Schreiber wrote:
    I/O was not done by the CPU (like it was done on home
    computers were all I/O was entirely handled by the CPU), but fully off-
    loaded to peripherials that were essentially limited computers of their
    own:
    I see you never used the serial cards with up to 32 terminals attached
    to a PC running SCO unix.

    I never touched SCO Unix, I got my Unix start on SunOS/Solaris and very
    early Linux.

    They were essentially IO processors

    Sure, they have to be to handle than many serial lines.

    But even in the "PC server" space, those were not that common.

    Kind regards,
    Alex.
    --
    "Opportunity is missed by most people because it is dressed in overalls and
    looks like work." -- Thomas A. Edison
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Alexander Schreiber@als@usenet.thangorodrim.de to comp.os.linux.misc,alt.folklore.computers on Wed Nov 11 12:00:47 2026
    From Newsgroup: comp.os.linux.misc

    Pancho <Pancho.Jones@protonmail.com> wrote:
    On 11/10/26 15:14, Scott Lurndal wrote:

    "Fully offloaded" is a bit weasel worded. I thought PCs did offload some >>> duties to io devices, nic chip, wifi chip, gpu card. sound card. Also,
    even a mainframe will need service routines to handle communications
    with IO devices. Where to draw the line seems unclear.

    Fully offloaded is accurate. For the Burroughs mainframes, a single
    instruction -initiated- an I/O request to a channel. The IIO (later
    SPIO) instruction would instruct the I/O processor (in the late 70s
    the IOP used an 8085 uProcessor) to initiate an I/O by passing a
    6-digit (24-bit) opcode/variant word and the start and end addresses
    of the buffer. The IOP would pass the command to a Data-Link Processor
    (DLP) that handled communications to the device controller (magnetic
    tape, disk) or unit record device (printer, card reader, card punch)
    or external communications processor (RS232/20ma current loop) and wait
    for the controller to start passing data back to the IOP. The IOP
    would move the data to the appropriate memory buffer and, when complete,
    signal the processor that the I/O had completed with an interrupt.


    Yes, you can give me a lot of details, but I'm not seeing the
    fundamental difference between that and DMA IO devices that have been
    around since the first IBM PC in 1981.

    First of all, those capabilities where standard - to the point where DMA
    wasn't mentioned as a feature because it was an assumed part of the tech baseline - in mainframes long before the 1980s.

    As I said I'm naive of this kind of low level stuff, but from my view as
    an application programmer I'm failing to see a fundamental difference
    from a PC, offloading an IO task to a non CPU device with communication
    via registers, interrupts and DMA.

    Does not compare. In the PC case, the controller with DMA handles simple requests such as "fetch X blocks of data from address Y on block device
    Z". For mainframe channel systems, while they also do this, they can do
    a lot more since they can run dedicated programs send to them as part of
    the request - including processing the data beyond "get me some bytes".

    Other folks already mentioned examples like telling the channel controller
    to go grab a specific file from the filesystem on an attached block device.

    On a PC, only the actual fetching of disk blocks would be offloaded to
    the controller and the filesystem handling (reading the start of the file- system, traversing the file system structures to find the list of blocks occupied by the file and its size) would have been done by the CPU.

    Kind regards,
    Alex.
    --
    "Opportunity is missed by most people because it is dressed in overalls and
    looks like work." -- Thomas A. Edison
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Alexander Schreiber@als@usenet.thangorodrim.de to comp.os.linux.misc,alt.folklore.computers on Wed Nov 11 12:15:20 2026
    From Newsgroup: comp.os.linux.misc

    Lawrence DrCOOliveiro <ldo@nz.invalid> wrote:
    On Tue, 10 Nov 2026 10:37:49 +0100, Alexander Schreiber wrote:

    ... a key design aspect of mainframes back then was that I/O was not
    done by the CPU (like it was done on home computers were all I/O was
    entirely handled by the CPU), but fully off- loaded to peripherials
    that were essentially limited computers of their own ...

    Which made sense back when the CPU was a scarce and expensive resource.
    It no longer made sense when that was no longer true.

    It still makes sense and that is borne out in hardware design. We haven't flipped back to polled I/O because "CPUs are faster than we know what to
    do with", quite the opposite: modern GPUs offload most the fancy 3D graphics processing (originally, these days it's also a lot of matrix math for AI), modern NICs offer a lot of offload (modern DC NICs are full on computers running Linux and e.g. handle network traffic encryption on their own,
    see e.g. NVidia Bluefield DPUs) and so on.

    Which leaves most of the CPU free to do compute work instead of
    busily doing lots of I/O. This and other design concepts (like form
    based terminals doing local editing - think web forms filled in the
    browser, send as a whole form and processed as a single request by
    the server, instead of the server handling every keystroke on the
    terminal) enabled having thousands of interactive users on one
    mainframe ...

    Nowadays we have millions of users on social media simultaneously, but
    that request/response paradigm is no longer sufficient for the
    real-time interaction requirements.

    The "Unix serial terminal" model never applied to web interactive traffic,
    that was always running on the "smart terminal doing local edits - mainframe style" model. Mostly because that's a lot easier to implement and also it scales a lot better.

    Kind regards,
    Alex.
    --
    "Opportunity is missed by most people because it is dressed in overalls and
    looks like work." -- Thomas A. Edison
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Alexander Schreiber@als@usenet.thangorodrim.de to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Wed Nov 11 12:19:16 2026
    From Newsgroup: comp.os.linux.misc

    c186282 <c186282@nnada.net> wrote:
    On 11/10/26 09:37, Peter Flass wrote:
    On 11/10/26 02:50, Alexander Schreiber wrote:
    Lawrence DrCOOliveiro <ldo@nz.invalid> wrote:
    On Tue, 10 Nov 2026 07:09:17 -0000 (UTC), Thomas Koenig wrote:

    Today's mainframes can no longer compete on compute power, but they
    have extremely high reliability (you can literally fire a bullet
    through a CPU, and things will migrate smoothly - if you don't hit
    the Recovery Unit) ...

    Cloud providers like Amazon, Google, Microsoft etc offer much higher
    reliability through massive parallelism, based on cheap, yet highly
    redundant hardware running a Free Software stack that is not priced
    per CPU. That is much more cost-effective and any equivalent number of >>>> mainframes. And there is no rCLRecovery UnitrCY that you can shoot at to >>>> take it out.

    The nice thing is, you can run the same sort of rCLenterprise-classrCY >>>> setup that the big folks operate in your own homelab, because all the
    foundation software they use is available to you, too
    <https://docs.xcp-ng.org/vms/vm-migration/>.

    Not exactly the same. The mainframe design philosophy includes "The
    machine MUST NOT fail", for which one pays in complexity and system
    cost. As long as your workload fits on that single machine (or a
    small cluster (Parallel Sysplex)) and is of the "MUST NOT fail"
    nature, fine.

    Highly distributed systems like those run and offered by the hyperscalers >>> are built on an entirely different design philosophy:
    -a - _everything_ fails all the time (law of big numbers: take the MTBF
    -a-a-a for a compute component, run a lot of them and you've got a steady >>> -a-a-a rain of failures, e.g. say you've got an MTBF of 1 million hours and >>> -a-a-a you run 2 million of these devices, then on average you'll have a >>> -a-a-a failure every half hour)

    Didn't I read somewhere that techs in large datacenters get around on
    rollerblades to swap out failing components?

    IS an issue !!!

    Linux/UNIX are 'better' at this.

    The failures I mentioned are hardware failing. OS and application level failures sit on top of that.

    As for 'failures every half hour' ... can't
    HAVE that !

    True, in real large data center environments the failure frequency is a
    lot higher than one failure per half hour (across all easily replacable component such as RAM, CPU, disk, PSU, mainboard, switches, ...).

    Kind regards,
    Alex.
    --
    "Opportunity is missed by most people because it is dressed in overalls and
    looks like work." -- Thomas A. Edison
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Alexander Schreiber@als@usenet.thangorodrim.de to comp.os.linux.misc,comp.unix.shell,alt.folklore.computers on Wed Nov 11 12:09:20 2026
    From Newsgroup: comp.os.linux.misc

    John Levine <johnl@taugh.com> wrote:
    According to Peter Flass <Peter@Iron-Spring.com>:
    Didn't I read somewhere that techs in large datacenters get around on >>rollerblades to swap out failing components?

    I don't think so. I've read that Google has people wandering around with grocery carts full of spare parts that they swap into failed servers.

    Not exactly "wandering around with grocery carts". The time of the service techs is not cheap, so you want to optimize its use - their repair jobs
    are scheduled to maximize time spent fixing machines and minimize time "wandering around" - so there is software that e.g. figures out if a
    given rack has enough pending repairs to make it worth sending a tech
    with a full cart loaded with spares there. You do not want your techs
    hopping across the DC floor to e.g. replace just a single stuck fan per
    rack as that would give them lots of exercise but the employer would
    receive little in useful work.

    Kind regards,
    Alex.
    --
    "Opportunity is missed by most people because it is dressed in overalls and
    looks like work." -- Thomas A. Edison
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From The Natural Philosopher@tnp@invalid.invalid to comp.os.linux.misc,alt.folklore.computers on Wed Nov 11 12:31:11 2026
    From Newsgroup: comp.os.linux.misc

    On 11/11/2026 09:08, Nuno Silva wrote:
    On 2026-11-11, Lawrence DrCOOliveiro wrote:

    On Tue, 10 Nov 2026 12:19:31 -0700, Peter Flass wrote:

    Anyone would think that newer=faster. Who would have thought it
    would be slower?

    As another example, I recall it said that passenger airliners
    after the Boeing 747 actually have slightly lower cruising speeds.
    I guess itrCOs all part of the goal of saving fuel.

    It is possible if the 747 was inefficient. IIRC there's a sweet spot
    (or range?) for efficient fuel usage in subsonic jets.


    707 was not a turbofan. As bypass engines became widely used, the
    overall economics shifted to lower speed lower drag.
    Airline economics is complex with passenger miles per day, and
    maintenance costs per flight hour being balanced by fuel costs per
    passenger mile.

    Higher fuel costs as well as rising bypass ratios have increased the
    tendency towards lower speed cruising

    (For supersonic, IIRC Aerospatiale-BAC Concorde was actually quite
    efficient in cruise (compared to other supersonic aircraft), and I
    think the proposed redesign would even do away with the need for wet
    thrust (necessary for TO and transonic climb?).)

    With rising fuel costs I somehow doubt supersonic airliners would ever
    be profitable again
    --
    ThererCOs a mighty big difference between good, sound reasons and reasons
    that sound good.

    Burton Hillis (William Vaughn, American columnist)

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From The Natural Philosopher@tnp@invalid.invalid to comp.os.linux.misc,alt.folklore.computers on Wed Nov 11 12:36:43 2026
    From Newsgroup: comp.os.linux.misc

    On 11/11/2026 11:03, Alexander Schreiber wrote:
    The Natural Philosopher <tnp@invalid.invalid> wrote:
    On 10/11/2026 09:37, Alexander Schreiber wrote:
    I/O was not done by the CPU (like it was done on home
    computers were all I/O was entirely handled by the CPU), but fully off-
    loaded to peripherials that were essentially limited computers of their
    own:
    I see you never used the serial cards with up to 32 terminals attached
    to a PC running SCO unix.

    I never touched SCO Unix, I got my Unix start on SunOS/Solaris and very
    early Linux.

    They were essentially IO processors

    Sure, they have to be to handle than many serial lines.

    But even in the "PC server" space, those were not that common.

    Depends on your age and experience. I saw many. A multiuser system that
    could handle 32 screens running a database oriented program for <
    $10,000 was a very good alternative to a VAX.

    Adding in TCP/IP rendered the serial ports obsolete and made it even
    better, apart from the cost per desktop of a PC running
    telnet...compared to a VT220 or a Wyse 50.



    Kind regards,
    Alex.
    --
    ItrCOs easier to fool people than to convince them that they have been fooled. Mark Twain



    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lars Poulsen@lars@beagle-ears.com to comp.os.linux.misc,alt.folklore.computers on Wed Nov 11 06:32:18 2026
    From Newsgroup: comp.os.linux.misc

    On 11/11/2026 03:00, Alexander Schreiber wrote:
    Does not compare. In the PC case, the controller with DMA handles simple requests such as "fetch X blocks of data from address Y on block device
    Z". For mainframe channel systems, while they also do this, they can do
    a lot more since they can run dedicated programs send to them as part of
    the request - including processing the data beyond "get me some bytes".

    Other folks already mentioned examples like telling the channel controller
    to go grab a specific file from the filesystem on an attached block device.

    On a PC, only the actual fetching of disk blocks would be offloaded to
    the controller and the filesystem handling (reading the start of the file- system, traversing the file system structures to find the list of blocks occupied by the file and its size) would have been done by the CPU.
    I never did channel programming at the SIO instruction level, but a
    colleague who did, told me that once mounted, you could copy one tape to another with a single SIO instruction, including backspace-retry on read
    and write errors.

    My direct experience with channel programming relates to programming a
    Z80 on a UNIBUS card to simulate reads and writes on a selector channel
    in order for a VAX to drive a line printer (really a Xerox 9700 laser
    printer - Paul Kyzivat of DEC write the VMS symbiont for it).
    --
    Lars Poulsen - an old geek in Santa Barbara, California
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Peter Flass@Peter@Iron-Spring.com to comp.os.linux.misc,alt.folklore.computers on Wed Nov 11 07:41:27 2026
    From Newsgroup: comp.os.linux.misc

    On 11/11/26 04:00, Alexander Schreiber wrote:
    Pancho <Pancho.Jones@protonmail.com> wrote:
    On 11/10/26 15:14, Scott Lurndal wrote:

    "Fully offloaded" is a bit weasel worded. I thought PCs did offload some >>>> duties to io devices, nic chip, wifi chip, gpu card. sound card. Also, >>>> even a mainframe will need service routines to handle communications
    with IO devices. Where to draw the line seems unclear.

    Fully offloaded is accurate. For the Burroughs mainframes, a single
    instruction -initiated- an I/O request to a channel. The IIO (later
    SPIO) instruction would instruct the I/O processor (in the late 70s
    the IOP used an 8085 uProcessor) to initiate an I/O by passing a
    6-digit (24-bit) opcode/variant word and the start and end addresses
    of the buffer. The IOP would pass the command to a Data-Link Processor >>> (DLP) that handled communications to the device controller (magnetic
    tape, disk) or unit record device (printer, card reader, card punch)
    or external communications processor (RS232/20ma current loop) and wait
    for the controller to start passing data back to the IOP. The IOP
    would move the data to the appropriate memory buffer and, when complete, >>> signal the processor that the I/O had completed with an interrupt.


    Yes, you can give me a lot of details, but I'm not seeing the
    fundamental difference between that and DMA IO devices that have been
    around since the first IBM PC in 1981.

    First of all, those capabilities where standard - to the point where DMA wasn't mentioned as a feature because it was an assumed part of the tech baseline - in mainframes long before the 1980s.

    As I said I'm naive of this kind of low level stuff, but from my view as
    an application programmer I'm failing to see a fundamental difference
    from a PC, offloading an IO task to a non CPU device with communication
    via registers, interrupts and DMA.

    Does not compare. In the PC case, the controller with DMA handles simple requests such as "fetch X blocks of data from address Y on block device
    Z". For mainframe channel systems, while they also do this, they can do
    a lot more since they can run dedicated programs send to them as part of
    the request - including processing the data beyond "get me some bytes".

    Other folks already mentioned examples like telling the channel controller
    to go grab a specific file from the filesystem on an attached block device.

    On a PC, only the actual fetching of disk blocks would be offloaded to
    the controller and the filesystem handling (reading the start of the file- system, traversing the file system structures to find the list of blocks occupied by the file and its size) would have been done by the CPU.

    Kind regards,
    Alex.

    ISAM channel programs would look up the desired key in the index, read
    the index record and store the disk address further down in the channel program (I'm a bit vague on details, so I don't know if this was done by
    the channel program or if there was a "program controlled interrupt"
    (PCI) bit set to have an interrupt routine do this), and then continue
    on and read the record - all in one I/O operation.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Peter Flass@Peter@Iron-Spring.com to comp.os.linux.misc,alt.folklore.computers on Wed Nov 11 07:45:23 2026
    From Newsgroup: comp.os.linux.misc

    On 11/11/26 04:15, Alexander Schreiber wrote:
    Lawrence DrCOOliveiro <ldo@nz.invalid> wrote:
    On Tue, 10 Nov 2026 10:37:49 +0100, Alexander Schreiber wrote:

    ... a key design aspect of mainframes back then was that I/O was not
    done by the CPU (like it was done on home computers were all I/O was
    entirely handled by the CPU), but fully off- loaded to peripherials
    that were essentially limited computers of their own ...

    Which made sense back when the CPU was a scarce and expensive resource.
    It no longer made sense when that was no longer true.

    It still makes sense and that is borne out in hardware design. We haven't flipped back to polled I/O because "CPUs are faster than we know what to
    do with", quite the opposite: modern GPUs offload most the fancy 3D graphics processing (originally, these days it's also a lot of matrix math for AI), modern NICs offer a lot of offload (modern DC NICs are full on computers running Linux and e.g. handle network traffic encryption on their own,
    see e.g. NVidia Bluefield DPUs) and so on.

    Which leaves most of the CPU free to do compute work instead of
    busily doing lots of I/O. This and other design concepts (like form
    based terminals doing local editing - think web forms filled in the
    browser, send as a whole form and processed as a single request by
    the server, instead of the server handling every keystroke on the
    terminal) enabled having thousands of interactive users on one
    mainframe ...

    Nowadays we have millions of users on social media simultaneously, but
    that request/response paradigm is no longer sufficient for the
    real-time interaction requirements.

    The "Unix serial terminal" model never applied to web interactive traffic, that was always running on the "smart terminal doing local edits - mainframe style" model. Mostly because that's a lot easier to implement and also it scales a lot better.

    Kind regards,
    Alex.


    Computing HARDWARE has gotten so much cheaper it makes sense to put some intelligence into any device so "the computer" can just ask it to do
    something and go do something else.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Peter Flass@Peter@Iron-Spring.com to comp.os.linux.misc,alt.folklore.computers on Wed Nov 11 07:56:40 2026
    From Newsgroup: comp.os.linux.misc

    On 11/11/26 07:32, Lars Poulsen wrote:
    I never did channel programming at the SIO instruction level, but a colleague who did, told me that once mounted, you could copy one tape to another with a single SIO instruction, including backspace-retry on read
    and write errors.

    I guess you could: read a block, transfer-in-channel (TIC) to read the
    next. You'd get an exception at end-of-tape. The error recovery was
    handled automatically by Data Management: back up and retry the read
    three times. If it still failed the programmer would have specified
    whether the bad block was skipped of if they'd get an error.

    I only got to play with EXCP once or twice.


    My direct experience with channel programming relates to programming a
    Z80 on a UNIBUS card to simulate reads and writes on a selector channel
    in order for a VAX to drive a line printer (really a Xerox 9700 laser printer - Paul Kyzivat of DEC write the VMS symbiont for it).

    Sounds like a fun project!

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Robert Riches@spamtrap42@jacob21819.net to comp.os.linux.misc on Wed Nov 11 15:19:39 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-11, Alexander Schreiber <als@usenet.thangorodrim.de> wrote:
    John Levine <johnl@taugh.com> wrote:
    According to Peter Flass <Peter@Iron-Spring.com>:
    Didn't I read somewhere that techs in large datacenters get around on >>>rollerblades to swap out failing components?

    I don't think so. I've read that Google has people wandering around with
    grocery carts full of spare parts that they swap into failed servers.

    Not exactly "wandering around with grocery carts". The time of the service techs is not cheap, so you want to optimize its use - their repair jobs
    are scheduled to maximize time spent fixing machines and minimize time "wandering around" - so there is software that e.g. figures out if a
    given rack has enough pending repairs to make it worth sending a tech
    with a full cart loaded with spares there. You do not want your techs
    hopping across the DC floor to e.g. replace just a single stuck fan per
    rack as that would give them lots of exercise but the employer would
    receive little in useful work.

    Kind regards,
    Alex.

    That description sounds like a possible use for something similar
    to the "elevator algorithm", perhaps involving _real_ elevators.
    ... interesting how life sometimes goes full circle.
    --
    Robert Riches
    spamtrap42@jacob21819.net
    (Yes, that is one of my email addresses.)
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From scott@scott@slp53.sl.home (Scott Lurndal) to comp.os.linux.misc,alt.folklore.computers on Wed Nov 11 15:23:19 2026
    From Newsgroup: comp.os.linux.misc

    c186282 <c186282@nnada.net> writes:
    On 11/10/26 04:37, Alexander Schreiber wrote:
    c186282 <c186282@nnada.net> wrote:
    On 11/9/26 16:20, John Levine wrote:
    According to c186282 <c186282@nnada.net>:
    Noted in this group though - I/O arch diffs between
    mainframes and minis.

    Will someone try to elaborate ?
    <snip>
    Which leaves most of the CPU free to do compute work instead of busily
    doing lots of I/O. This and other design concepts (like form based
    terminals doing local editing - think web forms filled in the browser, send >> as a whole form and processed as a single request by the server, instead
    of the server handling every keystroke on the terminal) enabled having
    thousands of interactive users on one mainframe, since the mainframe CPU
    did almost entirely compute (e.g. run the database and software for
    inventory, accounting, ...) and all I/O was offloaded to specialized
    computers (channels) acting as I/O interfaces. You essentially pay with
    extra complexity for a far more capable system.

    Kind regards,
    Alex.

    I kind of see what model you're describing.
    Really not THAT much different from what
    I outlined.

    "Smart" peripherials are good. CPU doesn't have
    to do nearly as much. The downside is that now
    you're hostage to the 'IQ' in the peripherials
    rather than fine managing/optimizing via the
    main CPU.

    How is that a downside?

    Modern peripherals seem to do quite well at
    both maanging and optimizing without interactions
    with the main CPU. There are lots of knobs
    that can be used to tune the behavior of the
    peripheral controller(s). Particularly in the
    ethernet networking space - managing traffic classes,
    handling encryption, adding/removing headers
    (e.g. MPLS or other tunnelling schemes) are all
    implemented in the controller (with control and
    tuning knobs available to the operating system).

    Some of the controllers I work with are even more
    complex (e.g. radio area networks, front-haul,
    back-haul, etc).
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From scott@scott@slp53.sl.home (Scott Lurndal) to comp.os.linux.misc,alt.folklore.computers on Wed Nov 11 15:31:31 2026
    From Newsgroup: comp.os.linux.misc

    c186282 <c186282@nnada.net> writes:
    On 11/10/26 09:35, Peter Flass wrote:
    On 11/10/26 02:37, Alexander Schreiber wrote:



    A prime example of neat code is that HASP (now JES) printed an entire
    page with one channel program instead of a line at a time. It also read
    a number of cards with one channel program instead of one card at a
    time, or, as someone has mentioned, one card column at a time.

    "Smart" periphs are good ... but also
    a kind of trap.

    Not really. The Mainframe controllers (Burroughs) in the
    1970s and 1980s offloaded work from the central CPU
    that could be used for actual work. Whether it was the
    line printers, card readers, tape drives, disk drives
    or terminal concentrators.


    Ideally the CPU does almost everything - so you
    can optimize/tweak.

    Can you describe _anything_ that could be optimized or
    tweaked, on the host CPU that cannot be more easily done
    by the adapter/controller hardware?



    But esp in the Olde Days the CPU really wasn't
    that strong or had that many resources.

    Especially today, it would be a huge waste to have the CPU
    interacting closely with the controllers on a per
    request basis, much less on each data transfer.

    NVMe controllers support multiple queues - the host
    OS simple queues requests(s) to the controller asynchronously
    and is notified (if desired) upon completion. To avoid
    per-request overhead, interrupt moderation techniques are
    used in the networking space (which use similar queueing
    mechanisms) and in the NVMe space.

    Additionally, most modern high-end devices support SRIOV
    virtualization capabilities, which allows guest operating
    systems direct access to the peripheral hardware, eliminating
    multiple trips throught the hypervisor on each request.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From scott@scott@slp53.sl.home (Scott Lurndal) to comp.os.linux.misc,alt.folklore.computers on Wed Nov 11 15:35:22 2026
    From Newsgroup: comp.os.linux.misc

    Alexander Schreiber <als@usenet.thangorodrim.de> writes:
    Pancho <Pancho.Jones@protonmail.com> wrote:
    On 11/10/26 15:14, Scott Lurndal wrote:



    Does not compare. In the PC case, the controller with DMA handles simple >requests such as "fetch X blocks of data from address Y on block device
    Z". For mainframe channel systems, while they also do this, they can do
    a lot more since they can run dedicated programs send to them as part of
    the request - including processing the data beyond "get me some bytes".

    In the early days of the PC, the "DMA controller" was not part of
    the peripheral itself, but rather external to the peripheral. It was
    programmed to simply send a sequence of load or store requests to
    the actual controller (e.g. pre-IDE disks, 8250 UART, etc).

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From scott@scott@slp53.sl.home (Scott Lurndal) to comp.os.linux.misc,alt.folklore.computers on Wed Nov 11 15:41:39 2026
    From Newsgroup: comp.os.linux.misc

    The Natural Philosopher <tnp@invalid.invalid> writes:
    On 11/11/2026 11:03, Alexander Schreiber wrote:


    But even in the "PC server" space, those were not that common.

    Depends on your age and experience. I saw many. A multiuser system that >could handle 32 screens running a database oriented program for <
    $10,000 was a very good alternative to a VAX.

    They were, as Alexander noted, not that common.


    Adding in TCP/IP rendered the serial ports obsolete and made it even
    better, apart from the cost per desktop of a PC running
    telnet...compared to a VT220 or a Wyse 50.

    That's not necessarily true, a high-end networking
    SoC (40-96 cores, lots of hardware accelerators/peripherals)
    has 16 UARTs (alongside multiple 200Gb ethernet ports).

    They're hardly obsolete.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From scott@scott@slp53.sl.home (Scott Lurndal) to comp.os.linux.misc,alt.folklore.computers on Wed Nov 11 15:47:13 2026
    From Newsgroup: comp.os.linux.misc

    Alexander Schreiber <als@usenet.thangorodrim.de> writes:
    Lawrence DrCOOliveiro <ldo@nz.invalid> wrote:
    On Tue, 10 Nov 2026 10:37:49 +0100, Alexander Schreiber wrote:

    ... a key design aspect of mainframes back then was that I/O was not
    done by the CPU (like it was done on home computers were all I/O was
    entirely handled by the CPU), but fully off- loaded to peripherials
    that were essentially limited computers of their own ...

    Which made sense back when the CPU was a scarce and expensive resource.
    It no longer made sense when that was no longer true.

    It still makes sense and that is borne out in hardware design. We haven't >flipped back to polled I/O because "CPUs are faster than we know what to
    do with", quite the opposite: modern GPUs offload most the fancy 3D graphics >processing (originally, these days it's also a lot of matrix math for AI), >modern NICs offer a lot of offload (modern DC NICs are full on computers >running Linux and e.g. handle network traffic encryption on their own,
    see e.g. NVidia Bluefield DPUs) and so on.

    https://www.marvell.com/products/data-processing-units.html

    These are all peripherals, either directly soldered on
    the mainboard, or via PCI-Express cards slots.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From John Ames@commodorejohn@gmail.com to comp.os.linux.misc,alt.folklore.computers on Wed Nov 11 08:42:44 2026
    From Newsgroup: comp.os.linux.misc

    On Wed, 11 Nov 2026 12:36:43 +0000
    The Natural Philosopher <tnp@invalid.invalid> wrote:

    I see you never used the serial cards with up to 32 terminals
    attached to a PC running SCO unix.

    I never touched SCO Unix, I got my Unix start on SunOS/Solaris and
    very early Linux.

    They were essentially IO processors

    Sure, they have to be to handle than many serial lines.

    But even in the "PC server" space, those were not that common.

    Depends on your age and experience. I saw many. A multiuser system
    that could handle 32 screens running a database oriented program for
    < $10,000 was a very good alternative to a VAX.

    Some of those setups persisted a surprisingly long time in the field.
    Used to work in the back room at a call center that was still running a terminal-based setup 'til early 2012; when all you need is a form for
    the reps to punch orders into, it Just Works.

    (Never got into the backend enough to learn what the host was running;
    some flavor of *nix that also managed the PBX, but the business end had
    been built on a Win2k Server box in Visual FoxPro, of all things.)

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From The Natural Philosopher@tnp@invalid.invalid to comp.os.linux.misc,alt.folklore.computers on Wed Nov 11 16:49:02 2026
    From Newsgroup: comp.os.linux.misc

    On 11/11/2026 16:42, John Ames wrote:
    On Wed, 11 Nov 2026 12:36:43 +0000
    The Natural Philosopher <tnp@invalid.invalid> wrote:

    I see you never used the serial cards with up to 32 terminals
    attached to a PC running SCO unix.

    I never touched SCO Unix, I got my Unix start on SunOS/Solaris and
    very early Linux.

    They were essentially IO processors

    Sure, they have to be to handle than many serial lines.

    But even in the "PC server" space, those were not that common.

    Depends on your age and experience. I saw many. A multiuser system
    that could handle 32 screens running a database oriented program for
    < $10,000 was a very good alternative to a VAX.

    Some of those setups persisted a surprisingly long time in the field.
    Used to work in the back room at a call center that was still running a terminal-based setup 'til early 2012; when all you need is a form for
    the reps to punch orders into, it Just Works.

    (Never got into the backend enough to learn what the host was running;
    some flavor of *nix that also managed the PBX, but the business end had
    been built on a Win2k Server box in Visual FoxPro, of all things.)

    Went into a bank in the mid noughties and the PC was running some sort
    of remote terminal APP into a mainframe somewhere

    A touch typist who can tab through fields is the fastest way to enter
    data, and any eye candy is mere distraction
    --
    To ban Christmas, simply give turkeys the vote.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Thomas Koenig@tkoenig@netcologne.de to comp.os.linux.misc,alt.folklore.computers on Wed Nov 11 17:18:15 2026
    From Newsgroup: comp.os.linux.misc

    The Natural Philosopher <tnp@invalid.invalid> schrieb:

    Went into a bank in the mid noughties and the PC was running some sort
    of remote terminal APP into a mainframe somewhere

    The colored variant of the 3270 terminals or terminal emulators
    (distinctive coloring easy to recognize) are in use at a furniture
    store a few km from where I live. No idea what system they have
    in the background, though.

    But terminals are interesting. I read a bit on the documentation of the
    VT52, where they talked a lot about microcode. The VT100 had a
    microprocessor already, an 8080.

    Which is why it would have been trivial for DEC to build a reasonable
    PC when they were building terminals.
    --
    This USENET posting was made without artificial intelligence,
    artificial impertinence, artificial arrogance, artificial stupidity,
    artificial flavorings or artificial colorants.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From John Ames@commodorejohn@gmail.com to comp.os.linux.misc,alt.folklore.computers on Wed Nov 11 10:06:03 2026
    From Newsgroup: comp.os.linux.misc

    On Wed, 11 Nov 2026 17:18:15 -0000 (UTC)
    Thomas Koenig <tkoenig@netcologne.de> wrote:

    But terminals are interesting. I read a bit on the documentation of
    the VT52, where they talked a lot about microcode. The VT100 had a microprocessor already, an 8080.

    Which is why it would have been trivial for DEC to build a reasonable
    PC when they were building terminals.

    Indeed - in fact, a number of contemporary microcomputers (e.g. the
    Vector 3) were basically just a terminal with more RAM and some form of
    mass storage, at the hardware level. DEC's lateness to the micro market
    was, unfortunately, a management issue rather than a technical one.

    (Not that they didn't have some technical hurdles in their own efforts,
    thanks to the almost-but-critically-not-quite-compatible LSI versions
    of the PDP-8, but Olsen's resistance to the idea and unwillingness to
    price down to the market doomed them right out the gate.)

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From The Natural Philosopher@tnp@invalid.invalid to comp.os.linux.misc,alt.folklore.computers on Wed Nov 11 18:13:42 2026
    From Newsgroup: comp.os.linux.misc

    On 11/11/2026 17:18, Thomas Koenig wrote:
    The Natural Philosopher <tnp@invalid.invalid> schrieb:

    Went into a bank in the mid noughties and the PC was running some sort
    of remote terminal APP into a mainframe somewhere

    The colored variant of the 3270 terminals or terminal emulators
    (distinctive coloring easy to recognize) are in use at a furniture
    store a few km from where I live. No idea what system they have
    in the background, though.

    But terminals are interesting. I read a bit on the documentation of the VT52, where they talked a lot about microcode. The VT100 had a
    microprocessor already, an 8080.

    Which is why it would have been trivial for DEC to build a reasonable
    PC when they were building terminals.

    That wasn't their market segment

    It's quite rare for a company that has seen success in one sector to
    reinvent itself in another.

    At the time, no one knew what use a PC would be.,
    --
    "An intellectual is a person knowledgeable in one field who speaks out
    only in others...rCY

    Tom Wolfe

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Carlos E.R.@robin_listas@es.invalid to comp.os.linux.misc on Wed Nov 11 20:16:22 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-07 23:41, Robert Riches wrote:
    On 2026-11-07, c186282 <c186282@nnada.net> wrote:
    On 11/6/26 23:50, Robert Riches wrote:
    On 2026-11-07, Janis Papanagnou <janis_papanagnou+ng@hotmail.com> wrote:


    I heard an apocryphal story several years or maybe a couple of
    decades ago that one car company hired the Yamaha audio company
    to design the exhaust system of one of their vehicles, and the
    result sounded EXTREMELY good.

    MAY be true.

    IC engines CAN sound REALLY nice - if properly 'tuned'.

    Electrics ... TOO quiet, people use engine sounds as
    part of their instant eval to cross a street. No sound
    equals no safety. The blind should SUE.

    At least in the US, I understand there is a government-imposed
    requirement for electrics to make noise. When I first heard of
    that requirement, I thought about selling a retrofit kit that
    would contain a pack of playing cards and a few clothespins. :-)
    I have described the artificial noise they create as about
    halfway between a flying saucer and a garbage disposal.

    I heard an EV in a parking lot the other day sounding as a flying saucer.


    Yes, there is some kind of requirement to make a noise, which they do
    with speakers somehow. I rented a fiat 600e this summer, and it made no
    noise at all. I did not see a config for it.
    --
    Cheers, Carlos.
    ESEfc-Efc+, EUEfc-Efc|;
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Charlie Gibbs@cgibbs@kltpzyxm.invalid to comp.os.linux.misc,alt.folklore.computers on Wed Nov 11 20:33:52 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-11, The Natural Philosopher <tnp@invalid.invalid> wrote:

    A touch typist who can tab through fields is the fastest way to enter
    data, and any eye candy is mere distraction

    This harks back to my frequently-stated belief that computer systems
    should be ugly and boring. Ugly as in lacking the eye candy that just
    gets in the way, and boring as in it Just Works without any unpleasant surprises.

    It's too bad that fast, efficient, easy-to-use designs have gone out
    of style, having been replaced by increasing layers of complexity that
    make systems cumbersome and prone to failure. (But they're awfully
    pretty...) I'm seeing more and more breakdowns that remind me of
    E.M. Forster's 1909 story "The Machine Stops". (His characters
    even used an equivalent to today's iPad.)
    --
    /~\ Charlie Gibbs | In this world there are
    \ / <cgibbs@kltpzyxm.invalid> | two kinds of people:
    X I'm really at ac.dekanfrus | 1. Those who can extrapolate
    / \ if you read it the right way. | from incomplete data.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Charlie Gibbs@cgibbs@kltpzyxm.invalid to comp.os.linux.misc,alt.folklore.computers on Wed Nov 11 20:33:53 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-11, The Natural Philosopher <tnp@invalid.invalid> wrote:

    With rising fuel costs I somehow doubt supersonic airliners would ever
    be profitable again

    Besides, no matter how fast the plane flies you're still
    going to be waiting for your baggage at the other end.

    Actual air time is only part of the total travel time -
    once you add in the trip to the airport, check-in time,
    baggage retrieval, and transport from the destination
    terminal to wherever you're really going, air time becomes
    a relatively small portion of the total time.
    --
    /~\ Charlie Gibbs | In this world there are
    \ / <cgibbs@kltpzyxm.invalid> | two kinds of people:
    X I'm really at ac.dekanfrus | 1. Those who can extrapolate
    / \ if you read it the right way. | from incomplete data.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Pancho@Pancho.Jones@protonmail.com to comp.os.linux.misc,alt.folklore.computers on Wed Nov 11 21:19:02 2026
    From Newsgroup: comp.os.linux.misc

    On 11/11/26 12:31, The Natural Philosopher wrote:
    On 11/11/2026 09:08, Nuno Silva wrote:
    On 2026-11-11, Lawrence DrCOOliveiro wrote:

    On Tue, 10 Nov 2026 12:19:31 -0700, Peter Flass wrote:

    Anyone would think that newer=faster. Who would have thought it
    would be slower?

    As another example, I recall it said that passenger airliners
    after the Boeing 747 actually have slightly lower cruising speeds.
    I guess itrCOs all part of the goal of saving fuel.

    It is possible if the 747 was inefficient. IIRC there's a sweet spot
    (or range?) for efficient fuel usage in subsonic jets.


    707 was not a turbofan. As bypass engines became widely used, the
    overall economics shifted to lower-a speed lower drag.
    Airline economics is complex with passenger miles per day, and
    maintenance costs per flight hour being balanced by fuel costs per
    passenger mile.

    Higher fuel costs as well as rising bypass ratios have increased the
    tendency towards lower speed cruising

    (For supersonic, IIRC Aerospatiale-BAC Concorde was actually quite
    efficient in cruise (compared to other supersonic aircraft), and I
    think the proposed redesign would even do away with the need for wet
    thrust (necessary for TO and transonic climb?).)

    With rising fuel costs I somehow doubt supersonic airliners would ever
    be profitable again


    I don't think supersonic airlines ever were profitable. My last business flight was on Concorde. Not something my company paid for, just a
    standard business flight randomly upgraded to Concorde, because there
    was so little demand for the direction, New York to London.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lawrence =?iso-8859-13?q?D=FFOliveiro?=@ldo@nz.invalid to comp.os.linux.misc,alt.folklore.computers on Wed Nov 11 21:48:26 2026
    From Newsgroup: comp.os.linux.misc

    On Wed, 11 Nov 2026 07:41:27 -0700, Peter Flass wrote:

    ISAM channel programs would look up the desired key in the index,
    read the index record and store the disk address further down in the
    channel program (I'm a bit vague on details, so I don't know if this
    was done by the channel program or if there was a "program
    controlled interrupt" (PCI) bit set to have an interrupt routine do
    this), and then continue on and read the record - all in one I/O
    operation.

    DECrCOs platforms never had this kind of complex hardware. On the other
    hand, doing everything in software meant they could add features to
    their ISAM stack like key compression and even record compression.

    To handle this sort of thing in an I/O channel controller, you end up
    making it a general-purpose CPU anyway.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lawrence =?iso-8859-13?q?D=FFOliveiro?=@ldo@nz.invalid to comp.os.linux.misc,alt.folklore.computers on Wed Nov 11 21:50:31 2026
    From Newsgroup: comp.os.linux.misc

    On Wed, 11 Nov 2026 21:19:02 +0000, Pancho wrote:

    I don't think supersonic airlines ever were profitable.

    Concorde was never profitable, even at the eye-watering prices they
    charged for a ticket. But it gave BA and Air France terrific bragging
    rights ...
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lawrence =?iso-8859-13?q?D=FFOliveiro?=@ldo@nz.invalid to comp.os.linux.misc,alt.folklore.computers on Wed Nov 11 21:57:28 2026
    From Newsgroup: comp.os.linux.misc

    On Wed, 11 Nov 2026 17:18:15 -0000 (UTC), Thomas Koenig wrote:

    But terminals are interesting. I read a bit on the documentation of
    the VT52, where they talked a lot about microcode. The VT100 had a microprocessor already, an 8080.

    The first VT100 line had some interesting firmware bugs. You could
    engage smooth-scroll mode (why donrCOt more apps have that now?), do
    some fancy cursor addressing, and suddenly find yourself looking at 3
    copies of the line you were typing into. Each time you entered text
    and pressed return, more copies would appear.

    And the VT100 series spawned unbelievable numbers of clones. The most
    elaborate one I saw was the C.Itoh CIT-101. That even had an escape
    sequence to trigger the key-click sound.

    The later VT200 series were nice and sleek and all that, but I felt
    they were a step backwards in one respect: press the rCLSET-UPrCY key, and
    it took perceptibly longer to switch to the setup menu than the old
    VT100 did -- that was instant. I assume the older model had some double-buffering display hardware support which was dropped from the
    newer range ...

    Which is why it would have been trivial for DEC to build a
    reasonable PC when they were building terminals.

    Their VT180 was able to run CP/M, complete with floppy drives.
    Presumably that took a separate 8080 (or Z80?) from the one that
    managed the terminal part.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Alexander Schreiber@als@usenet.thangorodrim.de to comp.os.linux.misc,alt.folklore.computers on Wed Nov 11 22:38:52 2026
    From Newsgroup: comp.os.linux.misc

    The Natural Philosopher <tnp@invalid.invalid> wrote:
    On 11/11/2026 16:42, John Ames wrote:
    On Wed, 11 Nov 2026 12:36:43 +0000
    The Natural Philosopher <tnp@invalid.invalid> wrote:

    I see you never used the serial cards with up to 32 terminals
    attached to a PC running SCO unix.

    I never touched SCO Unix, I got my Unix start on SunOS/Solaris and
    very early Linux.

    They were essentially IO processors

    Sure, they have to be to handle than many serial lines.

    But even in the "PC server" space, those were not that common.

    Depends on your age and experience. I saw many. A multiuser system
    that could handle 32 screens running a database oriented program for
    < $10,000 was a very good alternative to a VAX.

    Some of those setups persisted a surprisingly long time in the field.
    Used to work in the back room at a call center that was still running a
    terminal-based setup 'til early 2012; when all you need is a form for
    the reps to punch orders into, it Just Works.

    (Never got into the backend enough to learn what the host was running;
    some flavor of *nix that also managed the PBX, but the business end had
    been built on a Win2k Server box in Visual FoxPro, of all things.)

    Went into a bank in the mid noughties and the PC was running some sort
    of remote terminal APP into a mainframe somewhere

    Flying in the early 2010s, peeked at the display of the attendants
    desktop at checkin: Some sort of Windows, but 90% of the screen was
    occupied by a (IIRC) 3270 terminal emulator connected to the airline's
    system and clearly showing a typical mainframe app.

    A touch typist who can tab through fields is the fastest way to enter
    data, and any eye candy is mere distraction

    Indeed. Just the learning curve is a bit steeper than most people these
    days are used to.

    Kind regards,
    Alex.
    --
    "Opportunity is missed by most people because it is dressed in overalls and
    looks like work." -- Thomas A. Edison
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Alexander Schreiber@als@usenet.thangorodrim.de to comp.os.linux.misc,alt.folklore.computers on Wed Nov 11 22:36:09 2026
    From Newsgroup: comp.os.linux.misc

    Scott Lurndal <scott@slp53.sl.home> wrote:
    Alexander Schreiber <als@usenet.thangorodrim.de> writes:
    Lawrence DrCOOliveiro <ldo@nz.invalid> wrote:
    On Tue, 10 Nov 2026 10:37:49 +0100, Alexander Schreiber wrote:

    ... a key design aspect of mainframes back then was that I/O was not
    done by the CPU (like it was done on home computers were all I/O was
    entirely handled by the CPU), but fully off- loaded to peripherials
    that were essentially limited computers of their own ...

    Which made sense back when the CPU was a scarce and expensive resource.
    It no longer made sense when that was no longer true.

    It still makes sense and that is borne out in hardware design. We haven't >>flipped back to polled I/O because "CPUs are faster than we know what to
    do with", quite the opposite: modern GPUs offload most the fancy 3D graphics >>processing (originally, these days it's also a lot of matrix math for AI), >>modern NICs offer a lot of offload (modern DC NICs are full on computers >>running Linux and e.g. handle network traffic encryption on their own,
    see e.g. NVidia Bluefield DPUs) and so on.

    https://www.marvell.com/products/data-processing-units.html

    These are all peripherals, either directly soldered on
    the mainboard, or via PCI-Express cards slots.

    I know, I've got some NVidia Bluefield 2 DPUs right here. 2x 25 GBit/s
    SFP+ ports and an nice computer (8x Cortex A72 at 2+GHz, 16/32G of
    ECC RAM and 64/128G of eMMC, running Linux) inside the card (with its
    own Gigabit Ethernet port).

    Friends of mine are playing with those a lot, first one of them managed
    to run DOOM on it and the latest trick was connecting a BlueField 2 via
    PCIe interface to a regular (if somewhat older) PC GPU and having the
    Bluefield drive _that_ to run graphics. No "regular" PC required at all.

    Kind regards,
    Alex.
    --
    "Opportunity is missed by most people because it is dressed in overalls and
    looks like work." -- Thomas A. Edison
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From scott@scott@slp53.sl.home (Scott Lurndal) to comp.os.linux.misc,alt.folklore.computers on Wed Nov 11 22:21:49 2026
    From Newsgroup: comp.os.linux.misc

    Alexander Schreiber <als@usenet.thangorodrim.de> writes:
    The Natural Philosopher <tnp@invalid.invalid> wrote:
    On 11/11/2026 16:42, John Ames wrote:
    On Wed, 11 Nov 2026 12:36:43 +0000
    The Natural Philosopher <tnp@invalid.invalid> wrote:

    I see you never used the serial cards with up to 32 terminals
    attached to a PC running SCO unix.

    I never touched SCO Unix, I got my Unix start on SunOS/Solaris and
    very early Linux.

    They were essentially IO processors

    Sure, they have to be to handle than many serial lines.

    But even in the "PC server" space, those were not that common.

    Depends on your age and experience. I saw many. A multiuser system
    that could handle 32 screens running a database oriented program for
    < $10,000 was a very good alternative to a VAX.

    Some of those setups persisted a surprisingly long time in the field.
    Used to work in the back room at a call center that was still running a
    terminal-based setup 'til early 2012; when all you need is a form for
    the reps to punch orders into, it Just Works.

    (Never got into the backend enough to learn what the host was running;
    some flavor of *nix that also managed the PBX, but the business end had
    been built on a Win2k Server box in Visual FoxPro, of all things.)

    Went into a bank in the mid noughties and the PC was running some sort
    of remote terminal APP into a mainframe somewhere

    Flying in the early 2010s, peeked at the display of the attendants
    desktop at checkin: Some sort of Windows, but 90% of the screen was
    occupied by a (IIRC) 3270 terminal emulator connected to the airline's
    system and clearly showing a typical mainframe app.

    It's entirely likely that was communicating with a Unisys
    2200/Clearpath system (UNIVAC Standard Airline System).

    https://en.wikipedia.org/wiki/USAS_(application)
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From John Levine@johnl@taugh.com to comp.os.linux.misc,alt.folklore.computers on Wed Nov 11 22:52:22 2026
    From Newsgroup: comp.os.linux.misc

    According to Nuno Silva <nunojsilva@invalid.invalid>:
    As another example, I recall it said that passenger airliners after
    the Boeing 747 actually have slightly lower cruising speeds. I guess
    itrCOs all part of the goal of saving fuel.

    It is possible if the 747 was inefficient. IIRC there's a sweet spot (or >range?) for efficient fuel usage in subsonic jets.

    The 747 was never particularly efficient, but in its era it was so much bigger than other planes that the cost per passenger was reasonable because there
    was relatively fewer crew than on a 707 or DC-8. All of them had three
    people on the flight deck, captain, first officer, and flight engineer.

    The two engine 777 had more advanced and reliable engines. Once the 777
    was ceritfied for long trips over water (known as ETOPS) the 747 was
    dead. The 777 carried nearly as many people but with two engines and
    a two person flight crew.

    If someone's going to ask about the even bigger A380, it had the same
    problems as the 747 and is also being retired in favor of two engine
    A350. The 747 was (still is) a reasonable cargo jet, with the
    motivation for the upper deck being to put the flight deck up high
    enough that the nose could open to roll large cargo in and out. The
    A380 didn't do that.
    --
    Regards,
    John Levine, johnl@taugh.com, Primary Perpetrator of "The Internet for Dummies",
    Please consider the environment before reading this e-mail. https://jl.ly
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Alexander Schreiber@als@usenet.thangorodrim.de to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 00:20:51 2026
    From Newsgroup: comp.os.linux.misc

    Scott Lurndal <scott@slp53.sl.home> wrote:
    Alexander Schreiber <als@usenet.thangorodrim.de> writes:
    The Natural Philosopher <tnp@invalid.invalid> wrote:
    On 11/11/2026 16:42, John Ames wrote:
    On Wed, 11 Nov 2026 12:36:43 +0000
    The Natural Philosopher <tnp@invalid.invalid> wrote:

    I see you never used the serial cards with up to 32 terminals
    attached to a PC running SCO unix.

    I never touched SCO Unix, I got my Unix start on SunOS/Solaris and >>>>>> very early Linux.

    They were essentially IO processors

    Sure, they have to be to handle than many serial lines.

    But even in the "PC server" space, those were not that common.

    Depends on your age and experience. I saw many. A multiuser system
    that could handle 32 screens running a database oriented program for >>>>> < $10,000 was a very good alternative to a VAX.

    Some of those setups persisted a surprisingly long time in the field.
    Used to work in the back room at a call center that was still running a >>>> terminal-based setup 'til early 2012; when all you need is a form for
    the reps to punch orders into, it Just Works.

    (Never got into the backend enough to learn what the host was running; >>>> some flavor of *nix that also managed the PBX, but the business end had >>>> been built on a Win2k Server box in Visual FoxPro, of all things.)

    Went into a bank in the mid noughties and the PC was running some sort
    of remote terminal APP into a mainframe somewhere

    Flying in the early 2010s, peeked at the display of the attendants
    desktop at checkin: Some sort of Windows, but 90% of the screen was >>occupied by a (IIRC) 3270 terminal emulator connected to the airline's >>system and clearly showing a typical mainframe app.

    It's entirely likely that was communicating with a Unisys
    2200/Clearpath system (UNIVAC Standard Airline System).

    https://en.wikipedia.org/wiki/USAS_(application)

    Bingo! IIRC the airline was Lufthansa and they were indeed a longtime
    user of that system from what I could find.

    Kind regards,
    Alex.
    --
    "Opportunity is missed by most people because it is dressed in overalls and
    looks like work." -- Thomas A. Edison
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lars Poulsen@lars@beagle-ears.com to comp.os.linux.misc,alt.folklore.computers on Wed Nov 11 17:25:41 2026
    From Newsgroup: comp.os.linux.misc

    On 11/11/2026 12:33, Charlie Gibbs wrote:
    On 2026-11-11, The Natural Philosopher <tnp@invalid.invalid> wrote:

    With rising fuel costs I somehow doubt supersonic airliners would ever
    be profitable again

    Besides, no matter how fast the plane flies you're still
    going to be waiting for your baggage at the other end.

    Actual air time is only part of the total travel time -
    once you add in the trip to the airport, check-in time,
    baggage retrieval, and transport from the destination
    terminal to wherever you're really going, air time becomes
    a relatively small portion of the total time.

    On a transoceanic route, time in the air is quite significant.
    On a short "puddlejumper" route, a high-speed train (TGV in French) may
    well be faster.
    --
    Lars Poulsen - an old geek in Santa Barbara, California
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From rbowman@bowman@montana.com to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 02:06:05 2026
    From Newsgroup: comp.os.linux.misc

    On Wed, 11 Nov 2026 20:33:53 GMT, Charlie Gibbs wrote:

    Actual air time is only part of the total travel time -
    once you add in the trip to the airport, check-in time,
    baggage retrieval, and transport from the destination terminal to
    wherever you're really going, air time becomes a relatively small
    portion of the total time.

    I'm not planning on any more air travel in this lifetime but the less time locked in a flying tin can the better. The possible exception was business class when the 747s had a comfortable lounge on the upper deck.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lars Poulsen@lars@beagle-ears.com to comp.os.linux.misc,alt.folklore.computers on Wed Nov 11 18:50:17 2026
    From Newsgroup: comp.os.linux.misc

    On 11/11/2026 08:49, The Natural Philosopher wrote:
    Went into a bank in the mid noughties and the PC was running some sort
    of remote terminal APP into a mainframe somewhere

    A touch typist who can tab through fields is the fastest way to enter
    data, and any eye candy is mere distraction
    By that time, I suspect the interface protocol would have been TN3270
    (3270 over telnet).
    --
    Lars Poulsen - an old geek in Santa Barbara, California
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lars Poulsen@lars@beagle-ears.com to comp.os.linux.misc,alt.folklore.computers on Wed Nov 11 19:03:01 2026
    From Newsgroup: comp.os.linux.misc

    On 11/11/2026 09:18, Thomas Koenig wrote:
    The Natural Philosopher <tnp@invalid.invalid> schrieb:

    Went into a bank in the mid noughties and the PC was running some sort
    of remote terminal APP into a mainframe somewhere

    The colored variant of the 3270 terminals or terminal emulators
    (distinctive coloring easy to recognize) are in use at a furniture
    store a few km from where I live. No idea what system they have
    in the background, though.

    But terminals are interesting. I read a bit on the documentation of the VT52, where they talked a lot about microcode. The VT100 had a
    microprocessor already, an 8080.

    Which is why it would have been trivial for DEC to build a reasonable
    PC when they were building terminals.
    Indeed they did - The Rainbow PC-100 from 1982 was a dual-CPU version
    with a VT220 like screen. It had a Z80 to run CP/M and an 8088 to run
    MS-DOS.
    An engineer would say that with 896 KB of RAM it was superior to the
    IBM-PC; but it did not have an ISA bus, so it would never be clone-level compatible.

    https://en.wikipedia.org/wiki/Rainbow_100

    But in my opinion, the DecMate based on a single-chip PDP-8 was a much superior PC for office use.
    --
    Lars Poulsen - an old geek in Santa Barbara, California
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Charlie Gibbs@cgibbs@kltpzyxm.invalid to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 03:44:52 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-11, Alexander Schreiber <als@usenet.thangorodrim.de> wrote:

    The Natural Philosopher <tnp@invalid.invalid> wrote:

    A touch typist who can tab through fields is the fastest way to enter
    data, and any eye candy is mere distraction

    Indeed. Just the learning curve is a bit steeper than most people these
    days are used to.

    Perhaps, but oh, the advantages once you've learned it.
    And you only had to learn it once, rather than having to
    start all over again when each year's new model comes out.
    --
    /~\ Charlie Gibbs | In this world there are
    \ / <cgibbs@kltpzyxm.invalid> | two kinds of people:
    X I'm really at ac.dekanfrus | 1. Those who can extrapolate
    / \ if you read it the right way. | from incomplete data.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Charlie Gibbs@cgibbs@kltpzyxm.invalid to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 03:44:53 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-11, Lawrence DrCOOliveiro <ldo@nz.invalid> wrote:

    The later VT200 series were nice and sleek and all that, but I felt
    they were a step backwards in one respect: press the rCLSET-UPrCY key, and
    it took perceptibly longer to switch to the setup menu than the old
    VT100 did -- that was instant. I assume the older model had some double-buffering display hardware support which was dropped from the
    newer range ...

    It was a step backwards in another respect too: it indulged
    in the same sort of keyboard layout shenanigans as the IBM PC.
    I refused to touch one for that reason.

    It took the industry at least 10 years to recover from the
    silliness that the IBM PC keyboard layout inflicted on it.
    --
    /~\ Charlie Gibbs | In this world there are
    \ / <cgibbs@kltpzyxm.invalid> | two kinds of people:
    X I'm really at ac.dekanfrus | 1. Those who can extrapolate
    / \ if you read it the right way. | from incomplete data.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Charlie Gibbs@cgibbs@kltpzyxm.invalid to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 03:44:52 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-11, John Levine <johnl@taugh.com> wrote:

    The two engine 777 had more advanced and reliable engines. Once the 777
    was ceritfied for long trips over water (known as ETOPS) the 747 was
    dead. The 777 carried nearly as many people but with two engines and
    a two person flight crew.

    ETOPS officially stood for "Extended Twin-engine OPerationS",
    but it was colloquially known as "Engines Turn Or Passengers Swim".
    --
    /~\ Charlie Gibbs | In this world there are
    \ / <cgibbs@kltpzyxm.invalid> | two kinds of people:
    X I'm really at ac.dekanfrus | 1. Those who can extrapolate
    / \ if you read it the right way. | from incomplete data.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From c186282@c186282@nnada.net to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 00:39:38 2026
    From Newsgroup: comp.os.linux.misc

    On 11/11/26 22:44, Charlie Gibbs wrote:
    On 2026-11-11, Lawrence DrCOOliveiro <ldo@nz.invalid> wrote:

    The later VT200 series were nice and sleek and all that, but I felt
    they were a step backwards in one respect: press the rCLSET-UPrCY key, and >> it took perceptibly longer to switch to the setup menu than the old
    VT100 did -- that was instant. I assume the older model had some
    double-buffering display hardware support which was dropped from the
    newer range ...

    It was a step backwards in another respect too: it indulged
    in the same sort of keyboard layout shenanigans as the IBM PC.
    I refused to touch one for that reason.

    It took the industry at least 10 years to recover from the
    silliness that the IBM PC keyboard layout inflicted on it.

    Umm ... I always kinda liked the PC/AT keyboards.

    Plus they were heavy enough to use as weapons :-)

    Kept using my AT keyboard, with multiple adapters,
    right up until I retired a few years ago. Loved
    the positive clickety-click :-)

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From c186282@c186282@nnada.net to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 00:50:11 2026
    From Newsgroup: comp.os.linux.misc

    On 11/11/26 22:44, Charlie Gibbs wrote:
    On 2026-11-11, John Levine <johnl@taugh.com> wrote:

    The two engine 777 had more advanced and reliable engines. Once the 777
    was ceritfied for long trips over water (known as ETOPS) the 747 was
    dead. The 777 carried nearly as many people but with two engines and
    a two person flight crew.

    ETOPS officially stood for "Extended Twin-engine OPerationS",
    but it was colloquially known as "Engines Turn Or Passengers Swim".

    Heh heh !

    Frankly I'd be MUCH happier with at least THREE
    engines if going over wide expanses.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lawrence =?iso-8859-13?q?D=FFOliveiro?=@ldo@nz.invalid to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 06:18:29 2026
    From Newsgroup: comp.os.linux.misc

    On Thu, 12 Nov 2026 03:44:53 GMT, Charlie Gibbs wrote:

    On 2026-11-11, Lawrence DrCOOliveiro <ldo@nz.invalid> wrote:

    The later VT200 series were nice and sleek and all that, but I felt
    they were a step backwards in one respect: press the rCLSET-UPrCY key,
    and it took perceptibly longer to switch to the setup menu than the
    old VT100 did -- that was instant. I assume the older model had
    some double-buffering display hardware support which was dropped
    from the newer range ...

    It was a step backwards in another respect too: it indulged in the
    same sort of keyboard layout shenanigans as the IBM PC.

    No it didnrCOt. Compare <http://xahlee.info/kbd/vt220_terminal.html>.

    I refused to touch one for that reason.

    The short travel was unusual for the time, and was recognized as
    allowing for faster typing, but could also aggravate RSI.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lawrence =?iso-8859-13?q?D=FFOliveiro?=@ldo@nz.invalid to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 06:22:07 2026
    From Newsgroup: comp.os.linux.misc

    On Wed, 11 Nov 2026 19:03:01 -0800, Lars Poulsen wrote:

    An engineer would say that with 896 KB of RAM [the DEC Rainbow] was
    superior to the IBM-PC; but it did not have an ISA bus, so it would
    never be clone-level compatible.

    I remember when a local Microsoft guy came to tell us all about this
    newfangled thing called OS/2, he insisted that the 640kiB limit was
    something fundamental to MS-DOS, and not an IBM PC hardware limitation
    at all. I pointed out to him that there was a machine just on the
    floor above the hall where we were, that proved him wrong if he wanted
    to take a look.

    He was not pleased ...

    But in my opinion, the DecMate based on a single-chip PDP-8 was a much superior PC for office use.

    The only non-byte-addressable machine still shipping in the 1980s??
    (Outside of the mainframe market, perhaps.)
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lawrence =?iso-8859-13?q?D=FFOliveiro?=@ldo@nz.invalid to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 06:32:48 2026
    From Newsgroup: comp.os.linux.misc

    On Wed, 11 Nov 2026 12:15:20 +0100, Alexander Schreiber wrote:

    Lawrence DrCOOliveiro <ldo@nz.invalid> wrote:
    .
    On Tue, 10 Nov 2026 10:37:49 +0100, Alexander Schreiber wrote:

    ... a key design aspect of mainframes back then was that I/O was
    not done by the CPU (like it was done on home computers were all
    I/O was entirely handled by the CPU), but fully off- loaded to
    peripherials that were essentially limited computers of their own
    ...

    Which made sense back when the CPU was a scarce and expensive
    resource. It no longer made sense when that was no longer true.

    It still makes sense and that is borne out in hardware design. We
    haven't flipped back to polled I/O because "CPUs are faster than we
    know what to do with", quite the opposite: modern GPUs offload most
    the fancy 3D graphics processing (originally, these days it's also a
    lot of matrix math for AI), modern NICs offer a lot of offload
    (modern DC NICs are full on computers running Linux and e.g. handle
    network traffic encryption on their own, see e.g. NVidia Bluefield
    DPUs) and so on.

    So really, you are saying that general-purpose CPUs are now offloading
    work ... onto other general-purpose CPUs.

    Nowadays we have millions of users on social media simultaneously,
    but that request/response paradigm is no longer sufficient for the
    real-time interaction requirements.

    The "Unix serial terminal" model never applied to web interactive
    traffic, that was always running on the "smart terminal doing local
    edits - mainframe style" model.

    Again, yourCOre thinking in terms of an analogue of the old mainframe block-mode terminal. Real-time interaction requirements on the Web
    involve more than just the strict request/response model: you also
    have AJAX and WebSockets to provide a far greater level of
    interactivity. Block-mode terminals have no analogue to such
    functionality.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lawrence =?iso-8859-13?q?D=FFOliveiro?=@ldo@nz.invalid to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 06:35:24 2026
    From Newsgroup: comp.os.linux.misc

    On Wed, 11 Nov 2026 07:45:23 -0700, Peter Flass wrote:

    Computing HARDWARE has gotten so much cheaper it makes sense to put
    some intelligence into any device so "the computer" can just ask it
    to do something and go do something else.

    Or, to put it another way, multiple CPU cores have become so
    commonplace that it is no longer considered a big deal to rCLwasterCY some
    of them busy-waiting without bogging down others that are doing actual
    useful work.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lawrence =?iso-8859-13?q?D=FFOliveiro?=@ldo@nz.invalid to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 06:41:19 2026
    From Newsgroup: comp.os.linux.misc

    On Wed, 11 Nov 2026 22:36:09 +0100, Alexander Schreiber wrote:

    Friends of mine are playing with those a lot, first one of them
    managed to run DOOM on it and the latest trick was connecting a
    BlueField 2 via PCIe interface to a regular (if somewhat older) PC
    GPU and having the Bluefield drive _that_ to run graphics. No
    "regular" PC required at all.

    As an example of something that really does hark back to old
    mainframe-style ways of doing things, but in a good way, the RP2040
    chip introduced in the Raspberry Pi Pico included a pair of onboard programmable I/O channel controllers that were specifically optimized
    for high-speed bit-banging, in a way that a general-purpose CPU could
    not match.

    How high-speed? Here <https://www.youtube.com/watch?v=WaPJmCgseQw> is
    an early example: emulating the analog video output signal encoding
    from a 1980s-vintage BBC micro, in real time.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Pancho@Pancho.Jones@protonmail.com to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 07:26:07 2026
    From Newsgroup: comp.os.linux.misc

    On 11/12/26 06:35, Lawrence DrCOOliveiro wrote:
    On Wed, 11 Nov 2026 07:45:23 -0700, Peter Flass wrote:

    Computing HARDWARE has gotten so much cheaper it makes sense to put
    some intelligence into any device so "the computer" can just ask it
    to do something and go do something else.

    Or, to put it another way, multiple CPU cores have become so
    commonplace that it is no longer considered a big deal to rCLwasterCY some
    of them busy-waiting without bogging down others that are doing actual
    useful work.

    I read their comments to be the exact opposite, and, FWIW, it does seem
    to me that event/interrupt driven software and devices have become more prevalent.

    I'm still having difficulty understanding the difference between a
    mainframe and a PC. The argument that mainframes have independent IO
    devices appears to become "more" independent IO devices, when it is
    pointed out PCs have independent IO devices. It sounds a bit like an
    argument that mainframes are mainframes because they use electricity,
    and when it is pointed out PCs also use electricity the argument becomes
    there is no comparison because mainframes use kilowatts and PCs only use
    a few watts. Everything is true, but the distinctions are blurred.

    But after Windows NT, I never understood the difference between a mini computer and a PC. Perhaps next we can do the difference between a SoC computer (which I'm actually using now) and a PC ;-).
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From rbowman@bowman@montana.com to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 07:28:48 2026
    From Newsgroup: comp.os.linux.misc

    On Wed, 11 Nov 2026 19:03:01 -0800, Lars Poulsen wrote:

    Indeed they did - The Rainbow PC-100 from 1982 was a dual-CPU version
    with a VT220 like screen. It had a Z80 to run CP/M and an 8088 to run
    MS-DOS. An engineer would say that with 896 KB of RAM it was superior to
    the IBM-PC; but it did not have an ISA bus, so it would never be
    clone-level compatible.

    Another problem is DEC attempted to capture the diskette sales with a proprietary format available only from them at a much higher cost than
    generic Elephant or Memorex diskettes. DEC was still living in the mini
    world.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lawrence =?iso-8859-13?q?D=FFOliveiro?=@ldo@nz.invalid to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 07:55:47 2026
    From Newsgroup: comp.os.linux.misc

    On Thu, 12 Nov 2026 07:26:07 +0000, Pancho wrote:

    On 11/12/26 06:35, Lawrence DrCOOliveiro wrote:

    On Wed, 11 Nov 2026 07:45:23 -0700, Peter Flass wrote:

    Computing HARDWARE has gotten so much cheaper it makes sense to
    put some intelligence into any device so "the computer" can just
    ask it to do something and go do something else.

    Or, to put it another way, multiple CPU cores have become so
    commonplace that it is no longer considered a big deal to rCLwasterCY
    some of them busy-waiting without bogging down others that are
    doing actual useful work.

    I read their comments to be the exact opposite, and, FWIW, it does
    seem to me that event/interrupt driven software and devices have
    become more prevalent.

    *Everything* has rCLbecome more prevalentrCY, if you want to call it that, e.g.:

    * USB -- polling-based
    * Screen drawing -- locked to a fixed refresh rate
    * Networks and storage -- DMA-heavy, plus interrupts
    * Multiprocessor -- shared-memory-based communications, plus interrupts
    and polling (spinlocks) as the situation dictates

    So everything has its place somewhere.

    I'm still having difficulty understanding the difference between a
    mainframe and a PC.

    The difference is more to do with economics than technology. Once the
    economics no longer made sense, the distinction ceased to be
    important. That happened back in the 1980s; it is a testament to the
    sheer marketing prowess of the biggest mainframe company, IBM, that it
    took over another decade before the market actually noticed that it
    had become irrelevant.

    But after Windows NT, I never understood the difference between a
    mini computer and a PC.

    In the early days (1970s-1980s), the difference was, PCs were too
    limited to run rCLrealrCY operating systems. As the hardware grew more powerful, PCs became more able to run minicomputer-class operating
    systems like Unix.

    Microsoft Windows was born in the era before this, and kept a lot of
    baggage from that early time. This became more and more of a handicap
    as time went on, and thatrCOs what allowed Linux to flourish and
    basically conquer the entire rest of the computing market, apart from
    the desktop.

    Similarly to IBM, MicrosoftrCOs sheer marketing prowess helped it to
    hold on to that market segment, even as it found itself completely
    surrounded as soon as it tried to step outside that.

    Perhaps next we can do the difference between a SoC computer (which
    I'm actually using now) and a PC ;-).

    rCLSoCrCY stands for rCLSystem-on-ChiprCY. ItrCOs just the level of integration of the components, nothing more.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From The Natural Philosopher@tnp@invalid.invalid to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 08:11:46 2026
    From Newsgroup: comp.os.linux.misc

    On 11/11/2026 20:33, Charlie Gibbs wrote:
    Actual air time is only part of the total travel time -
    once you add in the trip to the airport, check-in time,
    baggage retrieval, and transport from the destination
    terminal to wherever you're really going, air time becomes
    a relatively small portion of the total time.

    Not if you are flying to Australia from the UK

    But yes, the break-even with driving is around 400 miles - less if
    there is congestion or a ferry or ocean in between.

    London to New York would be about 3 days by fast nuclear powered ship.
    --
    "When a true genius appears in the world, you may know him by this sign,
    that the dunces are all in confederacy against him."

    Jonathan Swift.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From The Natural Philosopher@tnp@invalid.invalid to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 08:24:44 2026
    From Newsgroup: comp.os.linux.misc

    On 12/11/2026 07:26, Pancho wrote:
    But after Windows NT, I never understood the difference between a mini computer and a PC. Perhaps next we can do the difference between a SoC computer (which I'm actually using now) and a PC

    Fundamentally there isnt much.

    What counts is the target market for it.

    Its rare to have a SoC running 1024 users...
    ...Or a single user mainframe

    In fact if you go back to the TMS9900 series of chips these made lousy
    single user computers. The chipset was designed for efficient context switching in a multitasking environment

    PCs were single tasking single user machines and were never very good at multi-tasking - even running NT

    At that point, SCO Unix was already a better option and when Linux
    arrived it was a no brainer.

    Linux has all the features of a decent minicomputer system when running
    on *86 arch.

    Mainframes are optimised for extremely large numbers of users. Probably
    using intelligent terminals that send a batch of data after many
    keystrokes. Attached to extremely large databases. On very resilient
    hardware.

    Its not Duke Nukem, but it is for all that, a specialised application.,
    --
    rCLThe urge to save humanity is almost always only a false face for the
    urge to rule it.rCY
    rCo H. L. Mencken

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Alexander Schreiber@als@usenet.thangorodrim.de to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 10:40:02 2026
    From Newsgroup: comp.os.linux.misc

    Charlie Gibbs <cgibbs@kltpzyxm.invalid> wrote:
    On 2026-11-11, Alexander Schreiber <als@usenet.thangorodrim.de> wrote:

    The Natural Philosopher <tnp@invalid.invalid> wrote:

    A touch typist who can tab through fields is the fastest way to enter
    data, and any eye candy is mere distraction

    Indeed. Just the learning curve is a bit steeper than most people these
    days are used to.

    Perhaps, but oh, the advantages once you've learned it.

    Wholeheartedly agree. I'm very much a child of the commandline, my primary email client is mutt, my newsreader is slrn and my idea of a graphical user interface is "20+ xterms and a few browser windows".

    And you only had to learn it once, rather than having to
    start all over again when each year's new model comes out.

    Indeed.

    Kind regards,
    Alex.
    --
    "Opportunity is missed by most people because it is dressed in overalls and
    looks like work." -- Thomas A. Edison
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Alexander Schreiber@als@usenet.thangorodrim.de to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 10:49:34 2026
    From Newsgroup: comp.os.linux.misc

    Lawrence DrCOOliveiro <ldo@nz.invalid> wrote:
    On Wed, 11 Nov 2026 12:15:20 +0100, Alexander Schreiber wrote:

    Lawrence DrCOOliveiro <ldo@nz.invalid> wrote:
    .
    On Tue, 10 Nov 2026 10:37:49 +0100, Alexander Schreiber wrote:

    ... a key design aspect of mainframes back then was that I/O was
    not done by the CPU (like it was done on home computers were all
    I/O was entirely handled by the CPU), but fully off- loaded to
    peripherials that were essentially limited computers of their own
    ...

    Which made sense back when the CPU was a scarce and expensive
    resource. It no longer made sense when that was no longer true.

    It still makes sense and that is borne out in hardware design. We
    haven't flipped back to polled I/O because "CPUs are faster than we
    know what to do with", quite the opposite: modern GPUs offload most
    the fancy 3D graphics processing (originally, these days it's also a
    lot of matrix math for AI), modern NICs offer a lot of offload
    (modern DC NICs are full on computers running Linux and e.g. handle
    network traffic encryption on their own, see e.g. NVidia Bluefield
    DPUs) and so on.

    So really, you are saying that general-purpose CPUs are now offloading
    work ... onto other general-purpose CPUs.

    Yes and no. GPUs are very much not general purpose CPUs, as they are
    highly optimizied for specific operations useful initially for fast
    3D graphics and later on also for AI compute. The general purpose CPUs
    in e.g. enterprise NICs are not for shovelling packets, but for extra
    services like traffic encryption and other offloads, while the shovelling
    of raw network packets is done by specialized hardware in the NICs.

    Nowadays we have millions of users on social media simultaneously,
    but that request/response paradigm is no longer sufficient for the
    real-time interaction requirements.

    The "Unix serial terminal" model never applied to web interactive
    traffic, that was always running on the "smart terminal doing local
    edits - mainframe style" model.

    Again, yourCOre thinking in terms of an analogue of the old mainframe block-mode terminal. Real-time interaction requirements on the Web
    involve more than just the strict request/response model: you also
    have AJAX and WebSockets to provide a far greater level of
    interactivity. Block-mode terminals have no analogue to such
    functionality.

    You are thinking too much from a "what the user sees" perspective. Under
    the hood all the "Real-time interaction requirements on the Web" boil down
    to RPC traffic - be it a browser loading a page or the JavaScript running
    on that page fetching more data. In fact one can think of the mainframe
    block mode terminals as RPC clients, with the terminal user filling in
    the RPC request and reading the RPC response directly, making the mainframe (among other things) an RPC server long before that term became standard technical lingo.

    Kind regards,
    Alex.
    --
    "Opportunity is missed by most people because it is dressed in overalls and
    looks like work." -- Thomas A. Edison
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Alexander Schreiber@als@usenet.thangorodrim.de to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 10:51:28 2026
    From Newsgroup: comp.os.linux.misc

    Lawrence DrCOOliveiro <ldo@nz.invalid> wrote:
    On Wed, 11 Nov 2026 07:45:23 -0700, Peter Flass wrote:

    Computing HARDWARE has gotten so much cheaper it makes sense to put
    some intelligence into any device so "the computer" can just ask it
    to do something and go do something else.

    Or, to put it another way, multiple CPU cores have become so
    commonplace that it is no longer considered a big deal to rCLwasterCY some
    of them busy-waiting without bogging down others that are doing actual
    useful work.

    Fun fact: spin locks (aka busy waiting) can be a much more efficient
    method than the alternative because it avoids context switches ...

    Kind regards,
    Alex.
    --
    "Opportunity is missed by most people because it is dressed in overalls and
    looks like work." -- Thomas A. Edison
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Alexander Schreiber@als@usenet.thangorodrim.de to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 10:36:51 2026
    From Newsgroup: comp.os.linux.misc

    Lars Poulsen <lars@beagle-ears.com> wrote:
    On 11/11/2026 12:33, Charlie Gibbs wrote:
    On 2026-11-11, The Natural Philosopher <tnp@invalid.invalid> wrote:

    With rising fuel costs I somehow doubt supersonic airliners would ever
    be profitable again

    Besides, no matter how fast the plane flies you're still
    going to be waiting for your baggage at the other end.

    Actual air time is only part of the total travel time -
    once you add in the trip to the airport, check-in time,
    baggage retrieval, and transport from the destination
    terminal to wherever you're really going, air time becomes
    a relatively small portion of the total time.

    On a transoceanic route, time in the air is quite significant.
    On a short "puddlejumper" route, a high-speed train (TGV in French) may
    well be faster.

    In fact the rollout of the TGV network largely killed off domestic flights
    in France, e.g. the Paris-Lyon TGV line opening in 1981 caused domestic
    air travel on that corridor to drop by 60% in one year. And once you factor
    in all the overhead (not flying) time of air travel, taking a train can actually be faster on shorter (domestic) routes. It tends to be more
    convenient too.

    Kind regards,
    Alex.
    --
    "Opportunity is missed by most people because it is dressed in overalls and
    looks like work." -- Thomas A. Edison
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Nuno Silva@nunojsilva@invalid.invalid to comp.os.linux.misc,alt.folklore.computers,alt.unix.geeks on Thu Nov 12 10:32:52 2026
    From Newsgroup: comp.os.linux.misc

    [ Followup-To: alt.unix.geeks, given this is now definitely getting out
    of scope? ]

    On 2026-11-12, Lars Poulsen wrote:

    On 11/11/2026 12:33, Charlie Gibbs wrote:
    On 2026-11-11, The Natural Philosopher <tnp@invalid.invalid> wrote:

    With rising fuel costs I somehow doubt supersonic airliners would ever
    be profitable again

    Besides, no matter how fast the plane flies you're still
    going to be waiting for your baggage at the other end.

    Actual air time is only part of the total travel time -
    once you add in the trip to the airport, check-in time,
    baggage retrieval, and transport from the destination
    terminal to wherever you're really going, air time becomes
    a relatively small portion of the total time.

    On a transoceanic route, time in the air is quite significant.
    On a short "puddlejumper" route, a high-speed train (TGV in French)
    may well be faster.

    I think there are places where high-speed rail is served with the same
    kind of security checks as air travel, which probably doesn't help with
    travel times.
    --
    Nuno Silva
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Peter Flass@Peter@Iron-Spring.com to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 07:32:20 2026
    From Newsgroup: comp.os.linux.misc

    On 11/11/26 20:44, Charlie Gibbs wrote:
    On 2026-11-11, Lawrence DrCOOliveiro <ldo@nz.invalid> wrote:

    The later VT200 series were nice and sleek and all that, but I felt
    they were a step backwards in one respect: press the rCLSET-UPrCY key, and >> it took perceptibly longer to switch to the setup menu than the old
    VT100 did -- that was instant. I assume the older model had some
    double-buffering display hardware support which was dropped from the
    newer range ...

    It was a step backwards in another respect too: it indulged
    in the same sort of keyboard layout shenanigans as the IBM PC.
    I refused to touch one for that reason.

    It took the industry at least 10 years to recover from the
    silliness that the IBM PC keyboard layout inflicted on it.

    1. I loved the VT220.
    2. The keyboard I'm typing this on a Unicomp a PC-101 clone.
    I refuse to touch a keyboard with an abomination called a "Windows
    Key" on it.



    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Peter Flass@Peter@Iron-Spring.com to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 07:37:01 2026
    From Newsgroup: comp.os.linux.misc

    On 11/11/26 23:32, Lawrence DrCOOliveiro wrote:
    On Wed, 11 Nov 2026 12:15:20 +0100, Alexander Schreiber wrote:

    Lawrence DrCOOliveiro <ldo@nz.invalid> wrote:
    .
    On Tue, 10 Nov 2026 10:37:49 +0100, Alexander Schreiber wrote:

    ... a key design aspect of mainframes back then was that I/O was
    not done by the CPU (like it was done on home computers were all
    I/O was entirely handled by the CPU), but fully off- loaded to
    peripherials that were essentially limited computers of their own
    ...

    Which made sense back when the CPU was a scarce and expensive
    resource. It no longer made sense when that was no longer true.

    It still makes sense and that is borne out in hardware design. We
    haven't flipped back to polled I/O because "CPUs are faster than we
    know what to do with", quite the opposite: modern GPUs offload most
    the fancy 3D graphics processing (originally, these days it's also a
    lot of matrix math for AI), modern NICs offer a lot of offload
    (modern DC NICs are full on computers running Linux and e.g. handle
    network traffic encryption on their own, see e.g. NVidia Bluefield
    DPUs) and so on.

    So really, you are saying that general-purpose CPUs are now offloading
    work ... onto other general-purpose CPUs.

    If you're going to develop an intelligent peripheral or controller, it
    makes both economic and technical sense to grab an off-the-shelf general purpose CPU. Postscript printers were an early example of this in the PC world. I think there were some systems where the printer had more
    compute power than the using system.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Peter Flass@Peter@Iron-Spring.com to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 07:40:00 2026
    From Newsgroup: comp.os.linux.misc

    On 11/12/26 00:26, Pancho wrote:
    On 11/12/26 06:35, Lawrence DrCOOliveiro wrote:
    On Wed, 11 Nov 2026 07:45:23 -0700, Peter Flass wrote:

    Computing HARDWARE has gotten so much cheaper it makes sense to put
    some intelligence into any device so "the computer" can just ask it
    to do something and go do something else.

    Or, to put it another way, multiple CPU cores have become so
    commonplace that it is no longer considered a big deal to rCLwasterCY some >> of them busy-waiting without bogging down others that are doing actual
    useful work.

    I read their comments to be the exact opposite, and, FWIW, it does seem
    to me that event/interrupt driven software and devices have become more prevalent.

    I'm still having difficulty understanding the difference between a
    mainframe and a PC. The argument that mainframes have independent IO
    devices appears to become "more" independent IO devices, when it is
    pointed out PCs have independent IO devices. It sounds a bit like an argument that mainframes are mainframes because they use electricity,
    and when it is pointed out PCs also use electricity the argument becomes there is no comparison because mainframes use kilowatts and PCs only use
    a few watts. Everything is true, but the distinctions are blurred.

    The difference is whether you choose to call it a mainframe or not.


    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Peter Flass@Peter@Iron-Spring.com to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 07:41:53 2026
    From Newsgroup: comp.os.linux.misc

    On 11/12/26 00:28, rbowman wrote:
    On Wed, 11 Nov 2026 19:03:01 -0800, Lars Poulsen wrote:

    Indeed they did - The Rainbow PC-100 from 1982 was a dual-CPU version
    with a VT220 like screen. It had a Z80 to run CP/M and an 8088 to run
    MS-DOS. An engineer would say that with 896 KB of RAM it was superior to
    the IBM-PC; but it did not have an ISA bus, so it would never be
    clone-level compatible.

    Another problem is DEC attempted to capture the diskette sales with a proprietary format available only from them at a much higher cost than generic Elephant or Memorex diskettes. DEC was still living in the mini world.

    DEC had so many offerings people can be forgiven for not keeping them straight. They had Rainbow and Robin (PCs, I think) and VaxStation and DECStation. Anything else? All incompatible.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From scott@scott@slp53.sl.home (Scott Lurndal) to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 15:09:04 2026
    From Newsgroup: comp.os.linux.misc

    Peter Flass <Peter@Iron-Spring.com> writes:
    On 11/12/26 00:28, rbowman wrote:
    On Wed, 11 Nov 2026 19:03:01 -0800, Lars Poulsen wrote:

    Indeed they did - The Rainbow PC-100 from 1982 was a dual-CPU version
    with a VT220 like screen. It had a Z80 to run CP/M and an 8088 to run
    MS-DOS. An engineer would say that with 896 KB of RAM it was superior to >>> the IBM-PC; but it did not have an ISA bus, so it would never be
    clone-level compatible.

    Another problem is DEC attempted to capture the diskette sales with a
    proprietary format available only from them at a much higher cost than
    generic Elephant or Memorex diskettes. DEC was still living in the mini
    world.

    DEC had so many offerings people can be forgiven for not keeping them >straight. They had Rainbow and Robin (PCs, I think) and VaxStation and >DECStation. Anything else?

    https://en.wikipedia.org/wiki/DEC_Professional
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From scott@scott@slp53.sl.home (Scott Lurndal) to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 15:10:58 2026
    From Newsgroup: comp.os.linux.misc

    Charlie Gibbs <cgibbs@kltpzyxm.invalid> writes:
    On 2026-11-11, Lawrence DrCOOliveiro <ldo@nz.invalid> wrote:

    The later VT200 series were nice and sleek and all that, but I felt
    they were a step backwards in one respect: press the rCLSET-UPrCY key, and >> it took perceptibly longer to switch to the setup menu than the old
    VT100 did -- that was instant. I assume the older model had some
    double-buffering display hardware support which was dropped from the
    newer range ...

    It was a step backwards in another respect too: it indulged
    in the same sort of keyboard layout shenanigans as the IBM PC.
    I refused to touch one for that reason.

    It took the industry at least 10 years to recover from the
    silliness that the IBM PC keyboard layout inflicted on it.

    It hasn't really recovered - the caps lock and control
    keys are still in the wrong place on PC keyboards. Fortunately,
    that's easy to fix on Unix/Linux and MacOs systems.

    I don't use windows (never have), so I can't speak for it.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From scott@scott@slp53.sl.home (Scott Lurndal) to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 15:21:18 2026
    From Newsgroup: comp.os.linux.misc

    The Natural Philosopher <tnp@invalid.invalid> writes:
    On 11/11/2026 20:33, Charlie Gibbs wrote:
    Actual air time is only part of the total travel time -
    once you add in the trip to the airport, check-in time,
    baggage retrieval, and transport from the destination
    terminal to wherever you're really going, air time becomes
    a relatively small portion of the total time.

    Not if you are flying to Australia from the UK

    But yes, the break-even with driving is around 400 miles - less if
    there is congestion or a ferry or ocean in between.

    I dunno. I used to drive from San Jose to Pasadena (350 miles). The
    trip took 4 hours 50 minutes[*]. Flight time from San Jose to Burbank
    is less than an hour (and before 2001, one could arrive at the
    gate 10 minutes before departure) and the rental cars were right
    there at the terminal, a five minute walk from baggage claim.

    [*] Assuming no significant traffic blockages and no speeding tickets
    (before 1987, the speed limit on I5 was 55mph - raised from 65 to 70 in 1996).

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From The Natural Philosopher@tnp@invalid.invalid to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 15:27:34 2026
    From Newsgroup: comp.os.linux.misc

    On 12/11/2026 14:37, Peter Flass wrote:
    f you're going to develop an intelligent peripheral or controller, it
    makes both economic and technical sense to grab an off-the-shelf general purpose CPU. Postscript printers were an early example of this in the PC world. I think there were some systems where the printer had more
    compute power than the using system.

    Indeed. The Apple Laserwriter had a Motorola 68000 CPU running at 12 MHz
    The Mac at that time had an 8MHz 68000...
    --
    If I had all the money I've spent on drink...
    ..I'd spend it on drink.

    Sir Henry (at Rawlinson's End)

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From The Natural Philosopher@tnp@invalid.invalid to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 15:29:37 2026
    From Newsgroup: comp.os.linux.misc

    On 12/11/2026 15:21, Scott Lurndal wrote:
    The Natural Philosopher <tnp@invalid.invalid> writes:
    On 11/11/2026 20:33, Charlie Gibbs wrote:
    Actual air time is only part of the total travel time -
    once you add in the trip to the airport, check-in time,
    baggage retrieval, and transport from the destination
    terminal to wherever you're really going, air time becomes
    a relatively small portion of the total time.

    Not if you are flying to Australia from the UK

    But yes, the break-even with driving is around 400 miles - less if
    there is congestion or a ferry or ocean in between.

    I dunno. I used to drive from San Jose to Pasadena (350 miles). The
    trip took 4 hours 50 minutes[*]. Flight time from San Jose to Burbank
    is less than an hour (and before 2001, one could arrive at the
    gate 10 minutes before departure) and the rental cars were right
    there at the terminal, a five minute walk from baggage claim.

    That was before the days of security checks and in a country where 400
    miles doesnt take you across a border

    [*] Assuming no significant traffic blockages and no speeding tickets
    (before 1987, the speed limit on I5 was 55mph - raised from 65 to 70 in 1996).

    --
    "It was a lot more fun being 20 in the 70's that it is being 70 in the 20's" Joew Walsh

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Alexander Schreiber@als@usenet.thangorodrim.de to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 17:16:11 2026
    From Newsgroup: comp.os.linux.misc

    Peter Flass <Peter@Iron-Spring.com> wrote:
    On 11/11/26 20:44, Charlie Gibbs wrote:
    On 2026-11-11, Lawrence DrCOOliveiro <ldo@nz.invalid> wrote:

    The later VT200 series were nice and sleek and all that, but I felt
    they were a step backwards in one respect: press the rCLSET-UPrCY key, and >>> it took perceptibly longer to switch to the setup menu than the old
    VT100 did -- that was instant. I assume the older model had some
    double-buffering display hardware support which was dropped from the
    newer range ...

    It was a step backwards in another respect too: it indulged
    in the same sort of keyboard layout shenanigans as the IBM PC.
    I refused to touch one for that reason.

    It took the industry at least 10 years to recover from the
    silliness that the IBM PC keyboard layout inflicted on it.

    1. I loved the VT220.
    2. The keyboard I'm typing this on a Unicomp a PC-101 clone.
    I refuse to touch a keyboard with an abomination called a "Windows
    Key" on it.

    The Windows key can be made useful by remapping it to Compose. Which is
    what I do on my laptops - because they're not running Windows anyway.

    Kind regards,
    Alex.
    --
    "Opportunity is missed by most people because it is dressed in overalls and
    looks like work." -- Thomas A. Edison
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From John Ames@commodorejohn@gmail.com to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 09:50:18 2026
    From Newsgroup: comp.os.linux.misc

    On Thu, 12 Nov 2026 08:24:44 +0000
    The Natural Philosopher <tnp@invalid.invalid> wrote:

    In fact if you go back to the TMS9900 series of chips these made
    lousy single user computers. The chipset was designed for efficient
    context switching in a multitasking environment

    Oh, the travails of the TI-99 series had nothing at all to do with the
    CPU per se; the design decisions on that platform were truly in the
    Absolutely Gorram Bonkers category:

    Start with the fact that the TMS9900 CPU, while it presents a byte-
    addressable memory model to the programmer, doesn't actually do byte-
    level addressing on the bus; all 8-bit writes are actually 16-bit read- modify-write operations, and 8-bit reads are just 16-bit reads where
    the irrelevant byte is discarded.

    Now consider that the TI-99 only has 256 bytes (128 words) of 16-bit
    RAM, which is used as a register file/scratchpad area for the firmware.
    All the rest is 8-bit. Moreover, it's not even mapped into main memory; instead, it doubles as video memory, and must be accessed byte-serial
    fashion via the VDP (set the address & do sequential access until you
    need to hit some other part of RAM, at which point you set the address
    *again* and repeat.) This also puts video-bound wait states on the bulk
    of memory accesses, on top of the 8-bit clunkiness.

    (There was an optional RAM cartridge add-on which mapped directly into
    CPU address space, but even this was 8-bit.)

    Moreover (as this makes *normal* programming effectively impossible on
    an unexpanded unit) the firmware actually implements a virtual-machine interpreter for programs stored in the video memory or in "graphic ROM" cartridges which are accessed in the same byte-serial fashion. And it's
    *this* instruction set that the built-in TI Basic is implemented in.

    So, to recap, when you power on the unit and enter your 10 PRINT "WTF
    WERE THEY SMOKING!?" : GOTO 10 test:

    * It's running in an intrepreted language.
    * Which is implemented in *another* interpreted language.
    * Which is being read a byte at a time.
    * Through a straw that the video chip is periodically pinching shut.
    * By a CPU that can't actually do byte accesses.

    Gee, I wonder why it was slow...?

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Charlie Gibbs@cgibbs@kltpzyxm.invalid to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 18:46:28 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-12, rbowman <bowman@montana.com> wrote:

    On Wed, 11 Nov 2026 19:03:01 -0800, Lars Poulsen wrote:

    Indeed they did - The Rainbow PC-100 from 1982 was a dual-CPU version
    with a VT220 like screen. It had a Z80 to run CP/M and an 8088 to run
    MS-DOS. An engineer would say that with 896 KB of RAM it was superior to
    the IBM-PC; but it did not have an ISA bus, so it would never be
    clone-level compatible.

    Another problem is DEC attempted to capture the diskette sales with a proprietary format available only from them at a much higher cost than generic Elephant or Memorex diskettes. DEC was still living in the mini world.

    We had a couple of dedicated word processors from an outfit called AES. (Remember the era of the dedicated word processor?) They did the same
    trick with diskettes. However, we had an option that would let the
    machines run CP/M - and FORMAT.COM would happily format any diskette
    in a way that the word processor liked. (And these were 5 1/4-inch hard-sectored diskettes - Dysan made them at the time, and sold them
    for half of what AES wanted for their preformatted diskettes.)
    --
    /~\ Charlie Gibbs | In this world there are
    \ / <cgibbs@kltpzyxm.invalid> | two kinds of people:
    X I'm really at ac.dekanfrus | 1. Those who can extrapolate
    / \ if you read it the right way. | from incomplete data.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Charlie Gibbs@cgibbs@kltpzyxm.invalid to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 18:46:28 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-12, Scott Lurndal <scott@slp53.sl.home> wrote:

    Charlie Gibbs <cgibbs@kltpzyxm.invalid> writes:

    It took the industry at least 10 years to recover from the
    silliness that the IBM PC keyboard layout inflicted on it.

    It hasn't really recovered - the caps lock and control
    keys are still in the wrong place on PC keyboards. Fortunately,
    that's easy to fix on Unix/Linux and MacOs systems.

    At least the shift, enter, and backspace keys are back
    where they belong.

    I remember sitting down at a SPARC workstation and trying
    to type in something. I reached for the left-hand shift
    key and wound up hitting the extra key between it and the
    Z key. I reached for the backspace key to correct it and
    hit something else again. I managed to catch myself before
    hurling the keyboard at the wall, maintaining the presence
    of mind to get up and walk back to my desk, where I telnetted
    into the SPARC from my own terminal. Then all I had to do
    was keep track of whether I was working with SysV or BSD:

    "Hey, the ps command isn't working right!"
    "That's because you're logged into the SPARC."
    "Oh..."

    I don't use windows (never have), so I can't speak for it.

    Lucky soul. :-)
    --
    /~\ Charlie Gibbs | In this world there are
    \ / <cgibbs@kltpzyxm.invalid> | two kinds of people:
    X I'm really at ac.dekanfrus | 1. Those who can extrapolate
    / \ if you read it the right way. | from incomplete data.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Charlie Gibbs@cgibbs@kltpzyxm.invalid to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 18:46:29 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-12, c186282 <c186282@nnada.net> wrote:

    On 11/11/26 22:44, Charlie Gibbs wrote:

    On 2026-11-11, John Levine <johnl@taugh.com> wrote:

    The two engine 777 had more advanced and reliable engines. Once the 777 >>> was ceritfied for long trips over water (known as ETOPS) the 747 was
    dead. The 777 carried nearly as many people but with two engines and
    a two person flight crew.

    ETOPS officially stood for "Extended Twin-engine OPerationS",
    but it was colloquially known as "Engines Turn Or Passengers Swim".

    Heh heh !

    Frankly I'd be MUCH happier with at least THREE
    engines if going over wide expanses.

    That's why ETOPS was devised. It contains measures to improve
    reliability. For instance, a different crew will work on each
    engine, to prevent one group's possible systematic errors from
    affecting the other.

    Still, shit happens. The DC-10 in the Sioux City crash had
    three separate hydraulic systems - and the disintegrating
    engine took out all three. BTW that plane did have three
    engines.

    Many redundant aircraft systems are routed (or re-routed if
    a potential problem is later discovered) to try to prevent
    one incident from damaging them all.
    --
    /~\ Charlie Gibbs | In this world there are
    \ / <cgibbs@kltpzyxm.invalid> | two kinds of people:
    X I'm really at ac.dekanfrus | 1. Those who can extrapolate
    / \ if you read it the right way. | from incomplete data.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Charlie Gibbs@cgibbs@kltpzyxm.invalid to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 18:46:30 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-12, Lawrence DrCOOliveiro <ldo@nz.invalid> wrote:

    On Thu, 12 Nov 2026 03:44:53 GMT, Charlie Gibbs wrote:

    On 2026-11-11, Lawrence DrCOOliveiro <ldo@nz.invalid> wrote:

    The later VT200 series were nice and sleek and all that, but I felt
    they were a step backwards in one respect: press the rCLSET-UPrCY key,
    and it took perceptibly longer to switch to the setup menu than the
    old VT100 did -- that was instant. I assume the older model had
    some double-buffering display hardware support which was dropped
    from the newer range ...

    It was a step backwards in another respect too: it indulged in the
    same sort of keyboard layout shenanigans as the IBM PC.

    No it didnrCOt. Compare <http://xahlee.info/kbd/vt220_terminal.html>.

    I see the same IBM tricks there: the extra key between Z and the
    left-hand shift key, and the return key moved farther away.

    I refused to touch one for that reason.

    The short travel was unusual for the time, and was recognized as
    allowing for faster typing, but could also aggravate RSI.

    I can use keys with short travel, but I don't like them.
    --
    /~\ Charlie Gibbs | In this world there are
    \ / <cgibbs@kltpzyxm.invalid> | two kinds of people:
    X I'm really at ac.dekanfrus | 1. Those who can extrapolate
    / \ if you read it the right way. | from incomplete data.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Charlie Gibbs@cgibbs@kltpzyxm.invalid to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 18:46:30 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-12, The Natural Philosopher <tnp@invalid.invalid> wrote:

    Mainframes are optimised for extremely large numbers of users. Probably using intelligent terminals that send a batch of data after many
    keystrokes. Attached to extremely large databases. On very resilient hardware.

    OK, let's take another stab at it, starting with your word "batch".
    Mainframes are traditionally batch-oriented. You submit jobs and
    data files, and it chews through them and leaves you the results.
    Interaction with the user is limited and often nonexistent.

    Yes, minis and PCs can work this way too. But mainframes are
    heavily optimized for this kind of work, to the point where
    they can outperform minis on batch jobs even while minis
    outperform mainframes on interactive applications.

    It's a philosophical thing.

    Let's see how long it takes someone to tear that one down...
    --
    /~\ Charlie Gibbs | In this world there are
    \ / <cgibbs@kltpzyxm.invalid> | two kinds of people:
    X I'm really at ac.dekanfrus | 1. Those who can extrapolate
    / \ if you read it the right way. | from incomplete data.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From John Levine@johnl@taugh.com to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 19:38:31 2026
    From Newsgroup: comp.os.linux.misc

    According to The Natural Philosopher <tnp@invalid.invalid>:
    But yes, the break-even with driving is around 400 miles - less if
    there is congestion or a ferry or ocean in between.

    London to New York would be about 3 days by fast nuclear powered ship.

    The SS United States went frm New York to England in 3 days, 10 hours, 40 minutes.
    That was in 1952.

    The ship was recently in the news having been towed to Florida and sunk yesterday
    to make an artificial reef for divers.

    R's,
    John
    --
    Regards,
    John Levine, johnl@taugh.com, Primary Perpetrator of "The Internet for Dummies",
    Please consider the environment before reading this e-mail. https://jl.ly
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From John Levine@johnl@taugh.com to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 19:42:56 2026
    From Newsgroup: comp.os.linux.misc

    According to Charlie Gibbs <cgibbs@kltpzyxm.invalid>:
    On 2026-11-12, The Natural Philosopher <tnp@invalid.invalid> wrote:

    Mainframes are optimised for extremely large numbers of users. Probably
    using intelligent terminals that send a batch of data after many
    keystrokes. Attached to extremely large databases. On very resilient
    hardware.

    OK, let's take another stab at it, starting with your word "batch". >Mainframes are traditionally batch-oriented. You submit jobs and
    data files, and it chews through them and leaves you the results.
    Interaction with the user is limited and often nonexistent.

    That was definitely true in the 1950s and 60s, somewhat true in the 1970s.
    Now it's not true at all. A modern mainframe supports thousands of simultaneous virtual machines.
    --
    Regards,
    John Levine, johnl@taugh.com, Primary Perpetrator of "The Internet for Dummies",
    Please consider the environment before reading this e-mail. https://jl.ly
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From rbowman@bowman@montana.com to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 19:44:02 2026
    From Newsgroup: comp.os.linux.misc

    On Thu, 12 Nov 2026 08:24:44 +0000, The Natural Philosopher wrote:

    In fact if you go back to the TMS9900 series of chips these made lousy
    single user computers. The chipset was designed for efficient context switching in a multitasking environment

    I worked on one project that used that processor. Its claim to fame was TI
    had radhard varieties.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From rbowman@bowman@montana.com to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 19:56:37 2026
    From Newsgroup: comp.os.linux.misc

    On Thu, 12 Nov 2026 07:32:20 -0700, Peter Flass wrote:

    2. The keyboard I'm typing this on a Unicomp a PC-101 clone.
    I refuse to touch a keyboard with an abomination called a "Windows
    Key" on it.

    The Sway config file refers to Mod4 as the Logo key. By default it's used
    for most of the key sequences like $mod+Return to open the $term.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From The Natural Philosopher@tnp@invalid.invalid to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 20:11:28 2026
    From Newsgroup: comp.os.linux.misc

    On 12/11/2026 19:38, John Levine wrote:
    According to The Natural Philosopher <tnp@invalid.invalid>:
    But yes, the break-even with driving is around 400 miles - less if
    there is congestion or a ferry or ocean in between.

    London to New York would be about 3 days by fast nuclear powered ship.

    The SS United States went frm New York to England in 3 days, 10 hours, 40 minutes.
    That was in 1952.

    There is a speed limit in displacement hulls irrespective of how much
    power you push through them.
    I only said nuclear because no one would build a new coaler or oil
    burner for that job now, not to run at that speed anyway. Far too much
    power and [expensive] fuel needed.

    Uranium is dirt cheap.


    The ship was recently in the news having been towed to Florida and sunk yesterday
    to make an artificial reef for divers.

    Pity they didn't sacrifice the orange Jesus along with it.

    R's,
    John
    --
    Climate is what you expect but weather is what you get.
    Mark Twain

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lawrence =?iso-8859-13?q?D=FFOliveiro?=@ldo@nz.invalid to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 20:33:17 2026
    From Newsgroup: comp.os.linux.misc

    On Thu, 12 Nov 2026 07:37:01 -0700, Peter Flass wrote:

    On 11/11/26 23:32, Lawrence DrCOOliveiro wrote:

    So really, you are saying that general-purpose CPUs are now
    offloading work ... onto other general-purpose CPUs.

    If you're going to develop an intelligent peripheral or controller, it
    makes both economic and technical sense to grab an off-the-shelf general purpose CPU.

    But the whole point of special-purpose hardware was that it could perform special-purpose tasks faster/more efficiently than a general-purpose CPU.

    But then those tasks gradually get more complex over time, and the special-purpose hardware gains more of the characteristics of a
    general-purpose CPU ...

    ... until you have to start the cycle again.

    <https://www.deviantart.com/default-cube/art/Computer-Graphics-And-The-Wheel-Of-Reincarnation-654204041>

    Of course, this topic drift is itself reinforcing the point that
    the technological justification for the rCLmainframerCY computing paradigm
    just doesnrCOt stack up any more.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lawrence =?iso-8859-13?q?D=FFOliveiro?=@ldo@nz.invalid to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 20:43:19 2026
    From Newsgroup: comp.os.linux.misc

    On Thu, 12 Nov 2026 07:32:20 -0700, Peter Flass wrote:

    I refuse to touch a keyboard with an abomination called a "Windows
    Key" on it.

    My Das Keyboard 4 has an interesting wing-like symbol on that key.
    Other hardware (like laptops which didnrCOt come with Windows) uses
    other symbols -- my System76 laptop labels it the rCLSuperrCY key, in the
    true hacker tradition. I think I have seen other Linux laptops that
    put a penguin symbol there.

    Microsoft even tried adding another one lately, the rCLCopilotrCY key. I donrCOt think thatrCOs been a big success ...

    EfEo:\>
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lawrence =?iso-8859-13?q?D=FFOliveiro?=@ldo@nz.invalid to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 20:38:19 2026
    From Newsgroup: comp.os.linux.misc

    On Thu, 12 Nov 2026 18:46:30 GMT, Charlie Gibbs wrote:

    On 2026-11-12, Lawrence DrCOOliveiro <ldo@nz.invalid> wrote:

    On Thu, 12 Nov 2026 03:44:53 GMT, Charlie Gibbs wrote:

    It was a step backwards in another respect too: it indulged in the
    same sort of keyboard layout shenanigans as the IBM PC.

    No it didnrCOt. Compare <http://xahlee.info/kbd/vt220_terminal.html>.

    I see the same IBM tricks there: the extra key between Z and the
    left-hand shift key, and the return key moved farther away.

    You mean, the angle brackets moved to their own key, so that rCL,rCY and rCL.rCY are still accessible with or without shift?

    In some ways, I think that makes sense. I still regularly hit rCL,rCY or rCL.rCY without letting go of the shift key, and get rCL<rCY and rCL>rCY instead
    without meaning to. Having dedicated keys would have been another way
    that typists could work even faster.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lawrence =?iso-8859-13?q?D=FFOliveiro?=@ldo@nz.invalid to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 20:46:59 2026
    From Newsgroup: comp.os.linux.misc

    On Thu, 12 Nov 2026 17:16:11 +0100, Alexander Schreiber wrote:

    The Windows key can be made useful by remapping it to Compose.

    I use Caps Lock for that.

    You want a non-modifier key for Compose, because it initiates a
    multi-keystroke sequence. And if yourCOre an Emacs user, you want to
    keep all the modifier keys you can get. EfOa
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Alexander Schreiber@als@usenet.thangorodrim.de to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 21:51:19 2026
    From Newsgroup: comp.os.linux.misc

    Charlie Gibbs <cgibbs@kltpzyxm.invalid> wrote:
    On 2026-11-12, c186282 <c186282@nnada.net> wrote:

    On 11/11/26 22:44, Charlie Gibbs wrote:

    On 2026-11-11, John Levine <johnl@taugh.com> wrote:

    The two engine 777 had more advanced and reliable engines. Once the 777 >>>> was ceritfied for long trips over water (known as ETOPS) the 747 was
    dead. The 777 carried nearly as many people but with two engines and
    a two person flight crew.

    ETOPS officially stood for "Extended Twin-engine OPerationS",
    but it was colloquially known as "Engines Turn Or Passengers Swim".

    Heh heh !

    Frankly I'd be MUCH happier with at least THREE
    engines if going over wide expanses.

    That's why ETOPS was devised. It contains measures to improve
    reliability. For instance, a different crew will work on each
    engine, to prevent one group's possible systematic errors from
    affecting the other.

    And we have seen accidents that root-caused to "mechanic messed up
    during maintenance", e.g. the SmartLynx Flight 9001 Incident (hard
    landing after loss of control, root-caused to wrong lubricant applied).

    Still, shit happens. The DC-10 in the Sioux City crash had
    three separate hydraulic systems - and the disintegrating
    engine took out all three. BTW that plane did have three
    engines.

    And that day it was learned that "three independent things" is not
    enough, their components also need to be set up not to share failure
    risks - such as being torn apart by a disintegrating engine near
    all three hydraulic lines. Sadly, this lesson did cost lives.

    Many redundant aircraft systems are routed (or re-routed if
    a potential problem is later discovered) to try to prevent
    one incident from damaging them all.

    Indeed. At least aviation is pretty good at learning from previous failures.

    Kind regards,
    Alex.
    --
    "Opportunity is missed by most people because it is dressed in overalls and
    looks like work." -- Thomas A. Edison
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Peter Flass@Peter@Iron-Spring.com to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 14:11:15 2026
    From Newsgroup: comp.os.linux.misc

    On 11/12/26 13:43, Lawrence DrCOOliveiro wrote:
    On Thu, 12 Nov 2026 07:32:20 -0700, Peter Flass wrote:

    I refuse to touch a keyboard with an abomination called a "Windows
    Key" on it.

    My Das Keyboard 4 has an interesting wing-like symbol on that key.
    Other hardware (like laptops which didnrCOt come with Windows) uses
    other symbols -- my System76 laptop labels it the rCLSuperrCY key, in the true hacker tradition. I think I have seen other Linux laptops that
    put a penguin symbol there.

    Microsoft even tried adding another one lately, the rCLCopilotrCY key. I donrCOt think thatrCOs been a big success ...


    That's because God is my copilot ;-)

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Charlie Gibbs@cgibbs@kltpzyxm.invalid to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 22:41:01 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-12, Peter Flass <Peter@Iron-Spring.com> wrote:

    On 11/12/26 13:43, Lawrence DrCOOliveiro wrote:

    On Thu, 12 Nov 2026 07:32:20 -0700, Peter Flass wrote:

    I refuse to touch a keyboard with an abomination called a "Windows
    Key" on it.

    My Das Keyboard 4 has an interesting wing-like symbol on that key.
    Other hardware (like laptops which didnrCOt come with Windows) uses
    other symbols -- my System76 laptop labels it the rCLSuperrCY key, in the
    true hacker tradition. I think I have seen other Linux laptops that
    put a penguin symbol there.

    Microsoft even tried adding another one lately, the rCLCopilotrCY key. I
    donrCOt think thatrCOs been a big success ...

    That's because God is my copilot ;-)

    That's better than Copilot being your god, I guess.
    --
    /~\ Charlie Gibbs | In this world there are
    \ / <cgibbs@kltpzyxm.invalid> | two kinds of people:
    X I'm really at ac.dekanfrus | 1. Those who can extrapolate
    / \ if you read it the right way. | from incomplete data.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Charlie Gibbs@cgibbs@kltpzyxm.invalid to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 22:41:02 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-12, Alexander Schreiber <als@usenet.thangorodrim.de> wrote:

    Charlie Gibbs <cgibbs@kltpzyxm.invalid> wrote:

    Many redundant aircraft systems are routed (or re-routed if
    a potential problem is later discovered) to try to prevent
    one incident from damaging them all.

    Indeed. At least aviation is pretty good at learning from previous failures.

    That's one of the things I like about aviation: it's one of
    the few remaining places where safety is taken seriously.
    --
    /~\ Charlie Gibbs | In this world there are
    \ / <cgibbs@kltpzyxm.invalid> | two kinds of people:
    X I'm really at ac.dekanfrus | 1. Those who can extrapolate
    / \ if you read it the right way. | from incomplete data.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Charlie Gibbs@cgibbs@kltpzyxm.invalid to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 22:41:02 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-12, Lawrence DrCOOliveiro <ldo@nz.invalid> wrote:

    On Thu, 12 Nov 2026 18:46:30 GMT, Charlie Gibbs wrote:

    On 2026-11-12, Lawrence DrCOOliveiro <ldo@nz.invalid> wrote:

    On Thu, 12 Nov 2026 03:44:53 GMT, Charlie Gibbs wrote:

    It was a step backwards in another respect too: it indulged in the
    same sort of keyboard layout shenanigans as the IBM PC.

    No it didnrCOt. Compare <http://xahlee.info/kbd/vt220_terminal.html>.

    I see the same IBM tricks there: the extra key between Z and the
    left-hand shift key, and the return key moved farther away.

    You mean, the angle brackets moved to their own key, so that rCL,rCY and rCL.rCY are still accessible with or without shift?

    In some ways, I think that makes sense. I still regularly hit rCL,rCY or rCL.rCY without letting go of the shift key, and get rCL<rCY and rCL>rCY instead
    without meaning to. Having dedicated keys would have been another way
    that typists could work even faster.

    That's not been a problem for me. Usually in such cases I'm using
    caps lock, which makes it a non-issue. But if you'd like a separate
    key I won't really object - as long as it's somewhere out of the way,
    not where you're going to hit it every time you reach for the left-hand
    shift key.
    --
    /~\ Charlie Gibbs | In this world there are
    \ / <cgibbs@kltpzyxm.invalid> | two kinds of people:
    X I'm really at ac.dekanfrus | 1. Those who can extrapolate
    / \ if you read it the right way. | from incomplete data.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From John Levine@johnl@taugh.com to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 23:10:19 2026
    From Newsgroup: comp.os.linux.misc

    According to Charlie Gibbs <cgibbs@kltpzyxm.invalid>:
    On 2026-11-12, Alexander Schreiber <als@usenet.thangorodrim.de> wrote:

    Charlie Gibbs <cgibbs@kltpzyxm.invalid> wrote:

    Many redundant aircraft systems are routed (or re-routed if
    a potential problem is later discovered) to try to prevent
    one incident from damaging them all.

    Indeed. At least aviation is pretty good at learning from previous failures.

    That's one of the things I like about aviation: it's one of
    the few remaining places where safety is taken seriously.

    Well, until the current people in charge in D.C.:

    https://www.nytimes.com/2026/11/12/magazine/air-travel-safety-faa-aviation.html --
    Regards,
    John Levine, johnl@taugh.com, Primary Perpetrator of "The Internet for Dummies",
    Please consider the environment before reading this e-mail. https://jl.ly
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Nuno Silva@nunojsilva@invalid.invalid to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 23:31:39 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-12, Alexander Schreiber wrote:

    Charlie Gibbs <cgibbs@kltpzyxm.invalid> wrote:
    On 2026-11-12, c186282 <c186282@nnada.net> wrote:

    On 11/11/26 22:44, Charlie Gibbs wrote:

    On 2026-11-11, John Levine <johnl@taugh.com> wrote:

    The two engine 777 had more advanced and reliable engines. Once the 777 >>>>> was ceritfied for long trips over water (known as ETOPS) the 747 was >>>>> dead. The 777 carried nearly as many people but with two engines and >>>>> a two person flight crew.

    ETOPS officially stood for "Extended Twin-engine OPerationS",
    but it was colloquially known as "Engines Turn Or Passengers Swim".

    Heh heh !

    Frankly I'd be MUCH happier with at least THREE
    engines if going over wide expanses.

    That's why ETOPS was devised. It contains measures to improve
    reliability. For instance, a different crew will work on each
    engine, to prevent one group's possible systematic errors from
    affecting the other.

    And we have seen accidents that root-caused to "mechanic messed up
    during maintenance", e.g. the SmartLynx Flight 9001 Incident (hard
    landing after loss of control, root-caused to wrong lubricant applied).

    Still, shit happens. The DC-10 in the Sioux City crash had
    three separate hydraulic systems - and the disintegrating
    engine took out all three. BTW that plane did have three
    engines.

    And that day it was learned that "three independent things" is not
    enough, their components also need to be set up not to share failure
    risks - such as being torn apart by a disintegrating engine near
    all three hydraulic lines. Sadly, this lesson did cost lives.

    And procedures and training need to stress not *introducing* failure
    sharing as part of the resolution attempt (Air Transat 236 would have
    flown on a single engine hadn't the fuel imbalance procedure been
    followed).


    Many redundant aircraft systems are routed (or re-routed if
    a potential problem is later discovered) to try to prevent
    one incident from damaging them all.

    Indeed. At least aviation is pretty good at learning from previous failures.

    Kind regards,
    Alex.
    --
    Nuno Silva
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Peter Flass@Peter@Iron-Spring.com to comp.os.linux.misc,alt.folklore.computers on Thu Nov 12 16:57:24 2026
    From Newsgroup: comp.os.linux.misc

    On 11/12/26 15:41, Charlie Gibbs wrote:
    On 2026-11-12, Peter Flass <Peter@Iron-Spring.com> wrote:

    On 11/12/26 13:43, Lawrence DrCOOliveiro wrote:

    On Thu, 12 Nov 2026 07:32:20 -0700, Peter Flass wrote:

    I refuse to touch a keyboard with an abomination called a "Windows
    Key" on it.

    My Das Keyboard 4 has an interesting wing-like symbol on that key.
    Other hardware (like laptops which didnrCOt come with Windows) uses
    other symbols -- my System76 laptop labels it the rCLSuperrCY key, in the >>> true hacker tradition. I think I have seen other Linux laptops that
    put a penguin symbol there.

    Microsoft even tried adding another one lately, the rCLCopilotrCY key. I >>> donrCOt think thatrCOs been a big success ...

    That's because God is my copilot ;-)

    That's better than Copilot being your god, I guess.

    https://en.wikipedia.org/wiki/God_Is_My_Co-Pilot_(film)

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lawrence =?iso-8859-13?q?D=FFOliveiro?=@ldo@nz.invalid to comp.os.linux.misc,alt.folklore.computers on Fri Nov 13 02:05:34 2026
    From Newsgroup: comp.os.linux.misc

    On Thu, 12 Nov 2026 10:49:34 +0100, Alexander Schreiber wrote:

    Lawrence DrCOOliveiro <ldo@nz.invalid> wrote:

    On Wed, 11 Nov 2026 12:15:20 +0100, Alexander Schreiber wrote:

    Lawrence DrCOOliveiro <ldo@nz.invalid> wrote:
    .
    Nowadays we have millions of users on social media
    simultaneously, but that request/response paradigm is no longer
    sufficient for the real-time interaction requirements.

    The "Unix serial terminal" model never applied to web interactive
    traffic, that was always running on the "smart terminal doing
    local edits - mainframe style" model.

    Again, yourCOre thinking in terms of an analogue of the old mainframe
    block-mode terminal. Real-time interaction requirements on the Web
    involve more than just the strict request/response model: you also
    have AJAX and WebSockets to provide a far greater level of
    interactivity. Block-mode terminals have no analogue to such
    functionality.

    You are thinking too much from a "what the user sees" perspective.

    Perception doesnrCOt count. ThatrCOs reality. You canrCOt *pretend* to offer interactivity: you really have to have support for interactivity built
    into your architecture from the lowest levels up. You canrCOt build per-character response on top of block mode.

    Under the hood all the "Real-time interaction requirements on the
    Web" boil down to RPC traffic - be it a browser loading a page or
    the JavaScript running on that page fetching more data.

    WebSocket connections are asynchronous full-duplex connections --
    thatrCOs not rCLRPCrCY by any stretch.

    Think of how rCLdoomscrollingrCY works: the page is continually fetching
    more data to update itself.

    Think of how Discord works: type some text and press Enter, and your
    message is immediately visible to everyone else on the same channel.
    Just start typing, and others are notified of that, too, so they can
    wait to see what you are going to say before adding their own
    comments.

    In fact one can think of the mainframe block mode terminals as RPC
    clients ...

    Maybe you can, but thatrCOs not how the interactive Web works.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Thomas Koenig@tkoenig@netcologne.de to comp.os.linux.misc,alt.folklore.computers on Fri Nov 13 07:31:58 2026
    From Newsgroup: comp.os.linux.misc

    Charlie Gibbs <cgibbs@kltpzyxm.invalid> schrieb:
    On 2026-11-12, Alexander Schreiber <als@usenet.thangorodrim.de> wrote:

    Charlie Gibbs <cgibbs@kltpzyxm.invalid> wrote:

    Many redundant aircraft systems are routed (or re-routed if
    a potential problem is later discovered) to try to prevent
    one incident from damaging them all.

    Indeed. At least aviation is pretty good at learning from previous failures.

    That's one of the things I like about aviation: it's one of
    the few remaining places where safety is taken seriously.

    I work in the chmemical industry, and we certainly take safety
    seriously there.

    But if you're interested in airline safety, I assume you read the
    Aviation Herald? One disturbing trend is the increasing number
    of lithium ion battery fires from things that passengers brought
    on board.
    --
    This USENET posting was made without artificial intelligence,
    artificial impertinence, artificial arrogance, artificial stupidity,
    artificial flavorings or artificial colorants.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Thomas Koenig@tkoenig@netcologne.de to comp.os.linux.misc,alt.folklore.computers on Fri Nov 13 07:33:48 2026
    From Newsgroup: comp.os.linux.misc

    Lawrence DrCOOliveiro <ldo@nz.invalid> schrieb:

    So really, you are saying that general-purpose CPUs are now offloading
    work ... onto other general-purpose CPUs.

    Not "now".

    Brooks wrote that one of the major design failures of the /360
    was that they failed to realize their channel processors were,
    in fact, computers.

    Later on, I believe IBM also used microprocessors, and the 801 was
    also used as a channel processor.
    --
    This USENET posting was made without artificial intelligence,
    artificial impertinence, artificial arrogance, artificial stupidity,
    artificial flavorings or artificial colorants.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Alexander Schreiber@als@usenet.thangorodrim.de to comp.os.linux.misc,alt.folklore.computers on Fri Nov 13 15:05:00 2026
    From Newsgroup: comp.os.linux.misc

    Lawrence DrCOOliveiro <ldo@nz.invalid> wrote:
    On Thu, 12 Nov 2026 10:49:34 +0100, Alexander Schreiber wrote:

    Lawrence DrCOOliveiro <ldo@nz.invalid> wrote:

    On Wed, 11 Nov 2026 12:15:20 +0100, Alexander Schreiber wrote:

    Lawrence DrCOOliveiro <ldo@nz.invalid> wrote:
    .
    Nowadays we have millions of users on social media
    simultaneously, but that request/response paradigm is no longer
    sufficient for the real-time interaction requirements.

    The "Unix serial terminal" model never applied to web interactive
    traffic, that was always running on the "smart terminal doing
    local edits - mainframe style" model.

    Again, yourCOre thinking in terms of an analogue of the old mainframe
    block-mode terminal. Real-time interaction requirements on the Web
    involve more than just the strict request/response model: you also
    have AJAX and WebSockets to provide a far greater level of
    interactivity. Block-mode terminals have no analogue to such
    functionality.

    You are thinking too much from a "what the user sees" perspective.

    Perception doesnrCOt count. ThatrCOs reality. You canrCOt *pretend* to offer interactivity: you really have to have support for interactivity built
    into your architecture from the lowest levels up. You canrCOt build per-character response on top of block mode.

    You absolutely can and the interactive web proves it. Yes, in the most
    simple case you build the entire page server side and then sent it
    to the client - think 1990s CGI scripts with no JavaScript. These
    days, things are of course typically more complex to enable the kind
    of interactivity that used to be the prerogative of native apps.

    Under the hood all the "Real-time interaction requirements on the
    Web" boil down to RPC traffic - be it a browser loading a page or
    the JavaScript running on that page fetching more data.

    WebSocket connections are asynchronous full-duplex connections --
    thatrCOs not rCLRPCrCY by any stretch.

    Of course it is. RPCs come in various kinds, from the classic one-shot (establish connection, send request, wait for and read response, close connection) to streaming RPCs. Even HTTP, which is not exactly the most efficient RPC method, came up with streaming RPCs: HTTP 1.1 pipelining,
    where multiple requests are handled within the same TCP connection.

    And "asynchronous" just means you don't do blocking reads and instead
    use the framework equivalent of select(2) for a more fluid interaction.

    Think of how rCLdoomscrollingrCY works: the page is continually fetching
    more data to update itself.

    And how does the page update itself? The JavaScript code running notices
    "end of currently rendered page about to come into viewport" and fires
    off RPCs to load more content to be rendered and displayed.

    Think of how Discord works: type some text and press Enter, and your
    message is immediately visible to everyone else on the same channel.

    Because the Discord app sends an RPC to the backend server with that
    text and all the various other clients get either get push notifications
    with that message (over long-lived connections and streaming RPCs) or
    a callback notification "hey, new content" and then do an RPC to get it
    (I haven't looked into the implementation for Discord).

    Just start typing, and others are notified of that, too, so they can
    wait to see what you are going to say before adding their own
    comments.

    Does discord really show the text to all channel members while you type it? That sounds annoying. But anyway, that's just more RPCs, likely streaming
    RPCs to keep the overhead down. How else do you think that works under
    the hood? Distributed crystal balls? Smoke signals?

    In fact one can think of the mainframe block mode terminals as RPC
    clients ...

    Maybe you can, but thatrCOs not how the interactive Web works.

    Once you take off all the colorful user-side front ends, that's exactly
    how it works. It's just not exposed to the user - like it was with
    block oriented terminals - and all hidden under several layers of
    content transport and presentation logic. But peel back the various
    layers of UI code and update logic and at the bottom you'll find RPC
    traffic. I've run enough of the backends for that ;-)

    Just because most web app developers these days never dive down all the way
    to the bottom of their stack where the RPCs happen doesn't mean they aren't happening.

    Kind regards,
    Alex.
    --
    "Opportunity is missed by most people because it is dressed in overalls and
    looks like work." -- Thomas A. Edison
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From The Natural Philosopher@tnp@invalid.invalid to comp.os.linux.misc,alt.folklore.computers on Fri Nov 13 14:17:37 2026
    From Newsgroup: comp.os.linux.misc

    On 13/11/2026 14:05, Alexander Schreiber wrote:
    Just because most web app developers these days never dive down all the way to the bottom of their stack where the RPCs happen doesn't mean they aren't happening.

    Kind regards,
    Alex.
    Exactly. Browser with JavaScript = thick client = equivalent 3270 functionality

    RPCs done via AJAX calls

    Not a submit button in sight
    --
    The higher up the mountainside
    The greener grows the grass.
    The higher up the monkey climbs
    The more he shows his arse.

    Traditional

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Peter Flass@Peter@Iron-Spring.com to comp.os.linux.misc,alt.folklore.computers on Fri Nov 13 07:40:05 2026
    From Newsgroup: comp.os.linux.misc

    On 11/13/26 00:31, Thomas Koenig wrote:
    Charlie Gibbs <cgibbs@kltpzyxm.invalid> schrieb:
    On 2026-11-12, Alexander Schreiber <als@usenet.thangorodrim.de> wrote:

    Charlie Gibbs <cgibbs@kltpzyxm.invalid> wrote:

    Many redundant aircraft systems are routed (or re-routed if
    a potential problem is later discovered) to try to prevent
    one incident from damaging them all.

    Indeed. At least aviation is pretty good at learning from previous failures.

    That's one of the things I like about aviation: it's one of
    the few remaining places where safety is taken seriously.

    I work in the chmemical industry, and we certainly take safety
    seriously there.

    But if you're interested in airline safety, I assume you read the
    Aviation Herald? One disturbing trend is the increasing number
    of lithium ion battery fires from things that passengers brought
    on board.


    I can't understand (well, I can...) why LiO2 batteries were ever
    approved for general use. Just the other day I saw on the news that a
    house burned (actually exploded) because if a battery fire.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Charlie Gibbs@cgibbs@kltpzyxm.invalid to comp.os.linux.misc,alt.folklore.computers on Fri Nov 13 15:54:48 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-12, Peter Flass <Peter@Iron-Spring.com> wrote:

    On 11/12/26 15:41, Charlie Gibbs wrote:

    On 2026-11-12, Peter Flass <Peter@Iron-Spring.com> wrote:

    On 11/12/26 13:43, Lawrence DrCOOliveiro wrote:

    On Thu, 12 Nov 2026 07:32:20 -0700, Peter Flass wrote:

    I refuse to touch a keyboard with an abomination called a "Windows
    Key" on it.

    My Das Keyboard 4 has an interesting wing-like symbol on that key.
    Other hardware (like laptops which didnrCOt come with Windows) uses
    other symbols -- my System76 laptop labels it the rCLSuperrCY key, in the >>>> true hacker tradition. I think I have seen other Linux laptops that
    put a penguin symbol there.

    Microsoft even tried adding another one lately, the rCLCopilotrCY key. I >>>> donrCOt think thatrCOs been a big success ...

    That's because God is my copilot ;-)

    That's better than Copilot being your god, I guess.

    https://en.wikipedia.org/wiki/God_Is_My_Co-Pilot_(film)

    Yes, I know about the film. My point stands -
    especially now that many people's god has let them down. Again.
    --
    /~\ Charlie Gibbs | In this world there are
    \ / <cgibbs@kltpzyxm.invalid> | two kinds of people:
    X I'm really at ac.dekanfrus | 1. Those who can extrapolate
    / \ if you read it the right way. | from incomplete data.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Charlie Gibbs@cgibbs@kltpzyxm.invalid to comp.os.linux.misc,alt.folklore.computers on Fri Nov 13 16:15:48 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-13, Peter Flass <Peter@Iron-Spring.com> wrote:

    On 11/13/26 00:31, Thomas Koenig wrote:

    Charlie Gibbs <cgibbs@kltpzyxm.invalid> schrieb:

    On 2026-11-12, Alexander Schreiber <als@usenet.thangorodrim.de> wrote:

    Charlie Gibbs <cgibbs@kltpzyxm.invalid> wrote:

    Many redundant aircraft systems are routed (or re-routed if
    a potential problem is later discovered) to try to prevent
    one incident from damaging them all.

    Indeed. At least aviation is pretty good at learning from previous failures.

    That's one of the things I like about aviation: it's one of
    the few remaining places where safety is taken seriously.

    I work in the chmemical industry, and we certainly take safety
    seriously there.

    But if you're interested in airline safety, I assume you read the
    Aviation Herald? One disturbing trend is the increasing number
    of lithium ion battery fires from things that passengers brought
    on board.

    I can't understand (well, I can...) why LiO2 batteries were ever
    approved for general use. Just the other day I saw on the news that a
    house burned (actually exploded) because if a battery fire.

    Years and years ago, Transport Canada mandated that all aircraft
    emergency locator transmitters (ELTs) must be equipped with lithium
    batteries because they were the only ones guaranteed to work in
    temperatures as low as -40 (Fahrenheit or Celsius, take your pick).
    (Never mind the fact that few aircraft - especially the puddle-jumpers
    in the general aviation fleet) were never taken that far north.)
    So we all dutifully complied - and then the reports started coming
    in of ELT batteries exploding, catching fire, or emitting harmful
    gases. So another directive came out mandating that all lithium
    batteries be replaced (with magnesium batteries IIRC). When I
    I pulled my ELT, I found corrosion around the connectors, from
    gases that had managed to leak from inside the case.

    I can't remember just what type of battery is in my current ELT.
    I just rely on the biennial inspections to keep things shipshape.
    --
    /~\ Charlie Gibbs | In this world there are
    \ / <cgibbs@kltpzyxm.invalid> | two kinds of people:
    X I'm really at ac.dekanfrus | 1. Those who can extrapolate
    / \ if you read it the right way. | from incomplete data.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From The Natural Philosopher@tnp@invalid.invalid to comp.os.linux.misc,alt.folklore.computers on Fri Nov 13 16:23:40 2026
    From Newsgroup: comp.os.linux.misc

    On 13/11/2026 14:40, Peter Flass wrote:
    On 11/13/26 00:31, Thomas Koenig wrote:
    Charlie Gibbs <cgibbs@kltpzyxm.invalid> schrieb:
    On 2026-11-12, Alexander Schreiber <als@usenet.thangorodrim.de> wrote:

    Charlie Gibbs <cgibbs@kltpzyxm.invalid> wrote:

    Many redundant aircraft systems are routed (or re-routed if
    a potential problem is later discovered) to try to prevent
    one incident from damaging them all.

    Indeed. At least aviation is pretty good at learning from previous
    failures.

    That's one of the things I like about aviation: it's one of
    the few remaining places where safety is taken seriously.

    I work in the chmemical industry, and we certainly take safety
    seriously there.

    But if you're interested in airline safety, I assume you read the
    Aviation Herald?-a One disturbing trend is the increasing number
    of lithium ion battery fires from things that passengers brought
    on board.


    I can't understand (well, I can...) why LiO2 batteries were ever
    approved for general use.

    There is no such battery. Lithium Ion or LIPO for lithium ion polymer or LiFePo for Lithium iron polymer


    Just the other day I saw on the news that a
    house burned (actually exploded) because if a battery fire.

    Batteries made badly and sold out of spec.
    --
    Religion is regarded by the common people as true, by the wise as
    foolish, and by the rulers as useful.

    (Seneca the Younger, 65 AD)


    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From scott@scott@slp53.sl.home (Scott Lurndal) to comp.os.linux.misc,alt.folklore.computers on Fri Nov 13 18:23:37 2026
    From Newsgroup: comp.os.linux.misc

    Charlie Gibbs <cgibbs@kltpzyxm.invalid> writes:
    On 2026-11-12, c186282 <c186282@nnada.net> wrote:

    On 11/11/26 22:44, Charlie Gibbs wrote:

    On 2026-11-11, John Levine <johnl@taugh.com> wrote:

    The two engine 777 had more advanced and reliable engines. Once the 777 >>>> was ceritfied for long trips over water (known as ETOPS) the 747 was
    dead. The 777 carried nearly as many people but with two engines and
    a two person flight crew.

    ETOPS officially stood for "Extended Twin-engine OPerationS",
    but it was colloquially known as "Engines Turn Or Passengers Swim".

    Heh heh !

    Frankly I'd be MUCH happier with at least THREE
    engines if going over wide expanses.

    That's why ETOPS was devised. It contains measures to improve
    reliability. For instance, a different crew will work on each
    engine, to prevent one group's possible systematic errors from
    affecting the other.

    Still, shit happens. The DC-10 in the Sioux City crash had
    three separate hydraulic systems - and the disintegrating
    engine took out all three. BTW that plane did have three
    engines.

    That's because they ran all three hydraulic lines together
    at one spot near the tail, and the uncontained failure of the tail
    mounted engine cut all three.

    Was still a fantastic landing given a complete lack of
    hydraulics.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Robert Riches@spamtrap42@jacob21819.net to comp.os.linux.misc on Fri Nov 13 19:17:15 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-13, Scott Lurndal <scott@slp53.sl.home> wrote:
    Charlie Gibbs <cgibbs@kltpzyxm.invalid> writes:
    On 2026-11-12, c186282 <c186282@nnada.net> wrote:

    On 11/11/26 22:44, Charlie Gibbs wrote:

    On 2026-11-11, John Levine <johnl@taugh.com> wrote:

    The two engine 777 had more advanced and reliable engines. Once the 777 >>>>> was ceritfied for long trips over water (known as ETOPS) the 747 was >>>>> dead. The 777 carried nearly as many people but with two engines and >>>>> a two person flight crew.

    ETOPS officially stood for "Extended Twin-engine OPerationS",
    but it was colloquially known as "Engines Turn Or Passengers Swim".

    Heh heh !

    Frankly I'd be MUCH happier with at least THREE
    engines if going over wide expanses.

    That's why ETOPS was devised. It contains measures to improve
    reliability. For instance, a different crew will work on each
    engine, to prevent one group's possible systematic errors from
    affecting the other.

    Still, shit happens. The DC-10 in the Sioux City crash had
    three separate hydraulic systems - and the disintegrating
    engine took out all three. BTW that plane did have three
    engines.

    That's because they ran all three hydraulic lines together
    at one spot near the tail, and the uncontained failure of the tail
    mounted engine cut all three.

    Was still a fantastic landing given a complete lack of
    hydraulics.

    One of probably several videos on the topic:

    https://www.youtube.com/watch?v=SqDlEgZYgww

    Definitely an amazing landing!
    --
    Robert Riches
    spamtrap42@jacob21819.net
    (Yes, that is one of my email addresses.)
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From =?UTF-8?Q?St=C3=A9phane?= CARPENTIER@sc@fiat-linux.fr to comp.os.linux.misc,alt.folklore.computers on Fri Nov 13 19:51:53 2026
    From Newsgroup: comp.os.linux.misc

    Le 12-11-2026, Alexander Schreiber <als@usenet.thangorodrim.de> a |-crit-a:

    Wholeheartedly agree. I'm very much a child of the commandline, my primary email client is mutt, my newsreader is slrn and my idea of a graphical user interface is "20+ xterms

    For the terminals, I rather prefer kitty. I really like the way to open
    the output of my last command directly in vim.

    and a few browser windows".

    Which browser? Qutebrowser, nixt, Tridactyl in FireFox or something
    else ? A long time ago, I used to use pentadactyl and the day a Firefox
    upgrade broke it, I looked for some browser managing directly the
    keyboard. Now, I'm using Qutebrowser.
    --
    Si vous avez du temps |a perdre :
    https://scarpet42.gitlab.io
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From =?UTF-8?Q?St=C3=A9phane?= CARPENTIER@sc@fiat-linux.fr to comp.os.linux.misc,alt.folklore.computers on Fri Nov 13 19:58:39 2026
    From Newsgroup: comp.os.linux.misc

    Le 12-11-2026, Peter Flass <Peter@Iron-Spring.com> a |-crit-a:
    I refuse to touch a keyboard with an abomination called a "Windows
    Key" on it.

    I've never found a useful way to use a "Window Key" with a Microsoft
    computer. With a Linux computer, this key is very useful. It's used to
    manage my windows knowing that every configuration I make with this key
    won't mess with the shortcut of my programs. It's so useful, I remapped
    the most useless and awful caps lock key on it.
    --
    Si vous avez du temps |a perdre :
    https://scarpet42.gitlab.io
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From =?UTF-8?Q?St=C3=A9phane?= CARPENTIER@sc@fiat-linux.fr to comp.os.linux.misc,alt.folklore.computers on Fri Nov 13 20:00:31 2026
    From Newsgroup: comp.os.linux.misc

    Le 12-11-2026, Alexander Schreiber <als@usenet.thangorodrim.de> a |-crit-a:

    The Windows key can be made useful by remapping it to Compose.

    Yes, but I don't consider it a remap.
    --
    Si vous avez du temps |a perdre :
    https://scarpet42.gitlab.io
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From John Ames@commodorejohn@gmail.com to comp.os.linux.misc,alt.folklore.computers on Fri Nov 13 12:30:58 2026
    From Newsgroup: comp.os.linux.misc

    On 13 Nov 2026 19:58:39 GMT
    Stophane CARPENTIER <sc@fiat-linux.fr> wrote:
    I've never found a useful way to use a "Window Key" with a Microsoft computer. With a Linux computer, this key is very useful. It's used to
    manage my windows knowing that every configuration I make with this
    key won't mess with the shortcut of my programs.
    I use it for window-manager shortcuts for the same reason.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lawrence =?iso-8859-13?q?D=FFOliveiro?=@ldo@nz.invalid to comp.os.linux.misc,alt.folklore.computers on Fri Nov 13 20:47:15 2026
    From Newsgroup: comp.os.linux.misc

    On Fri, 13 Nov 2026 07:33:48 -0000 (UTC), Thomas Koenig wrote:

    Lawrence DrCOOliveiro <ldo@nz.invalid> schrieb:

    So really, you are saying that general-purpose CPUs are now offloading
    work ... onto other general-purpose CPUs.

    Not "now".

    Brooks wrote that one of the major design failures of the /360
    was that they failed to realize their channel processors were,
    in fact, computers.

    But the whole point of special-purpose hardware was that it could perform special-purpose tasks faster/more efficiently than a general-purpose CPU.

    But then those tasks gradually get more complex over time, and the special-purpose hardware gains more of the characteristics of a
    general-purpose CPU ...

    ... until you have to start the cycle again.

    <https://www.deviantart.com/default-cube/art/Computer-Graphics-And-The-Wheel-Of-Reincarnation-654204041>

    Of course, this topic drift is itself reinforcing the point that
    the technological justification for the rCLmainframerCY computing paradigm
    just doesnrCOt stack up any more.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Alexander Schreiber@als@usenet.thangorodrim.de to comp.os.linux.misc,alt.folklore.computers on Fri Nov 13 22:02:26 2026
    From Newsgroup: comp.os.linux.misc

    The Natural Philosopher <tnp@invalid.invalid> wrote:
    On 13/11/2026 14:40, Peter Flass wrote:
    On 11/13/26 00:31, Thomas Koenig wrote:
    Charlie Gibbs <cgibbs@kltpzyxm.invalid> schrieb:
    On 2026-11-12, Alexander Schreiber <als@usenet.thangorodrim.de> wrote: >>>>
    Charlie Gibbs <cgibbs@kltpzyxm.invalid> wrote:

    Many redundant aircraft systems are routed (or re-routed if
    a potential problem is later discovered) to try to prevent
    one incident from damaging them all.

    Indeed. At least aviation is pretty good at learning from previous
    failures.

    That's one of the things I like about aviation: it's one of
    the few remaining places where safety is taken seriously.

    I work in the chmemical industry, and we certainly take safety
    seriously there.

    But if you're interested in airline safety, I assume you read the
    Aviation Herald?-a One disturbing trend is the increasing number
    of lithium ion battery fires from things that passengers brought
    on board.


    I can't understand (well, I can...) why LiO2 batteries were ever
    approved for general use.

    There is no such battery. Lithium Ion or LIPO for lithium ion polymer or LiFePo for Lithium iron polymer

    Well, there are Lithium-Air batteries. But while various implementations
    of them have been demonstrated, AFAIK they are nowhere near
    commercialization.

    Just the other day I saw on the news that a
    house burned (actually exploded) because if a battery fire.

    Batteries made badly and sold out of spec.

    And/or badly abused during transport, installation, operation. Plus
    people with little knowledge of the technology tend to end up with
    the more exciting (read: fire-prone) lithium battery chemistries instead
    of with much safer ones e.g. LiFePo4. It doesn't help that often the
    cell chemistry is not explicitly mentioned, e.g. with battery packs
    for e-bikes and such.

    Kind regards,
    Alex.
    --
    "Opportunity is missed by most people because it is dressed in overalls and
    looks like work." -- Thomas A. Edison
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Alexander Schreiber@als@usenet.thangorodrim.de to comp.os.linux.misc,alt.folklore.computers on Fri Nov 13 21:49:02 2026
    From Newsgroup: comp.os.linux.misc

    Charlie Gibbs <cgibbs@kltpzyxm.invalid> wrote:
    On 2026-11-13, Peter Flass <Peter@Iron-Spring.com> wrote:

    On 11/13/26 00:31, Thomas Koenig wrote:

    Charlie Gibbs <cgibbs@kltpzyxm.invalid> schrieb:

    On 2026-11-12, Alexander Schreiber <als@usenet.thangorodrim.de> wrote: >>>>
    Charlie Gibbs <cgibbs@kltpzyxm.invalid> wrote:

    Many redundant aircraft systems are routed (or re-routed if
    a potential problem is later discovered) to try to prevent
    one incident from damaging them all.

    Indeed. At least aviation is pretty good at learning from previous failures.

    That's one of the things I like about aviation: it's one of
    the few remaining places where safety is taken seriously.

    I work in the chmemical industry, and we certainly take safety
    seriously there.

    But if you're interested in airline safety, I assume you read the
    Aviation Herald? One disturbing trend is the increasing number
    of lithium ion battery fires from things that passengers brought
    on board.

    I can't understand (well, I can...) why LiO2 batteries were ever
    approved for general use. Just the other day I saw on the news that a
    house burned (actually exploded) because if a battery fire.

    Years and years ago, Transport Canada mandated that all aircraft
    emergency locator transmitters (ELTs) must be equipped with lithium
    batteries because they were the only ones guaranteed to work in
    temperatures as low as -40 (Fahrenheit or Celsius, take your pick).

    "Lithium batteries" really doesn't narrow it down, there are at least
    a dozen very different chemistries and battery designs under that grossly simplifying label.

    Some digging indicates those were Lithium-Sulfurdioxide primary cells,
    and they are pressurized (!). I assume they didn't pick Lithium Thionyl- chloride because they were after higher cold temperature current pulses?

    (Never mind the fact that few aircraft - especially the puddle-jumpers
    in the general aviation fleet) were never taken that far north.)
    So we all dutifully complied - and then the reports started coming
    in of ELT batteries exploding, catching fire, or emitting harmful
    gases.

    Heaving sniffed sulfur dioxide (in low concentrations): oh yes, that
    stuff is nasty and stings for sure, do not inhale.

    So another directive came out mandating that all lithium
    batteries be replaced (with magnesium batteries IIRC). When I
    I pulled my ELT, I found corrosion around the connectors, from
    gases that had managed to leak from inside the case.

    Yeah, the Lithium-Sulfurdioxide cells apparently contain sulfur dioxide
    gas unter pressure (one source indicates 3 bar of pressure) and that
    gas is highly corrosive. All it takes is a slowly degrading seal and
    you've got a nasty mess on your hands.

    Kind regards,
    Alex.
    --
    "Opportunity is missed by most people because it is dressed in overalls and
    looks like work." -- Thomas A. Edison
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Johann 'Myrkraverk' Oskarsson@johann@myrkraverk.invalid to comp.os.linux.misc,alt.folklore.computers on Sat Nov 14 06:24:55 2026
    From Newsgroup: comp.os.linux.misc

    On 11/11/2026 12:46 PM, c186282 wrote:
    On 11/10/26 10:14, Scott Lurndal wrote:
    Pancho <Pancho.Jones@protonmail.com> writes:
    On 11/10/26 09:37, Alexander Schreiber wrote:

    No. The operating system and application software ran on the mainframe >>>> CPU. But - as described above - a key design aspect of mainframes back >>>> then was that I/O was not done by the CPU (like it was done on home
    computers were all I/O was entirely handled by the CPU), but fully off- >>>> loaded to peripherials that were essentially limited computers of their >>>
    "Fully offloaded" is a bit weasel worded. I thought PCs did offload some >>> duties to io devices, nic chip, wifi chip, gpu card. sound card. Also,
    even a mainframe will need service routines to handle communications
    with IO devices. Where to draw the line seems unclear.

    Fully offloaded is accurate.-a For the Burroughs mainframes, a single
    instruction -initiated- an I/O request to a channel.-a-a The IIO (later
    SPIO) instruction would instruct the I/O processor (in the late 70s
    the IOP used an 8085 uProcessor) to initiate an I/O by passing a
    6-digit (24-bit) opcode/variant word and the start and end addresses
    of the buffer.-a-a The IOP would pass the command to a Data-Link Processor >> (DLP) that handled communications to the device controller (magnetic
    tape, disk) or unit record device (printer, card reader, card punch)
    or external communications processor (RS232/20ma current loop) and wait
    for the controller to start passing data back to the IOP.-a The IOP
    would move the data to the appropriate memory buffer and, when complete,
    signal the processor that the I/O had completed with an interrupt.

    -a Esp in the Bad Old Days, "smart perips" were VERY
    -a important. The CPU just sent a LITTLE stuff to them.
    -a Very CPU/RAM/Cycle efficient.

    -a Flip - left you hostage to the periph "IQ", often
    -a no ways to tweak/enhance it all.


    Caveat: I don't have a clue about these things, it is just that what you >>> are saying doesn't sound quite right.

    -a Me ... SEEN it all - the plusses AND minuses.

    -a The somewhat older group members will have
    -a seen MORE of that.

    -a The newbies think NOTHING from the past can
    -a possibly inform them, pave the way.

    -a Horribly wrong ... but you'll never convince them.
    -a For them it's now all "AI" ... until it displaces
    -a them entirely.

    1970s "cloud computing" would be lugging around a VT-of-the-day, and a
    modem. Then you could connect it to the hotel's phone line and get some
    work done while far and away.

    Or at least, that's what I imagine it could have been; I lack the per-
    sonal experience of it.

    But no, no; that's all /new/. Cloud computing is hip and cool.


    -a Can ya say "Would You Like Fries With That ?" in
    -a three or four languages ???-a :-)

    Yes, as a matter of fact I can; English, Icelandic, and French. Proba-
    bly more languages in the future.


    Best wishes, and happy terminals!

    %! P.S. mine is VT520, so not a 70s vintage.
    --
    Johann | email: invalid -> com | http://www.myrkraverk.com/blog/
    I'm not from the Internet, I just work there. | via Easynews.com https://bsky.app/profile/myrkraverk.bsky.social | for ( ;; ) _:;
    Federated at https://fed.brid.gy/bsky/myrkraverk.bsky.social
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lawrence =?iso-8859-13?q?D=FFOliveiro?=@ldo@nz.invalid to comp.os.linux.misc,alt.folklore.computers on Sat Nov 14 01:09:57 2026
    From Newsgroup: comp.os.linux.misc

    On Fri, 13 Nov 2026 22:02:26 +0100, Alexander Schreiber wrote:

    ... lithium battery chemistries instead of with much safer ones e.g.
    LiFePo4.

    Polonium?!?
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From c186282@c186282@nnada.net to comp.os.linux.misc,alt.folklore.computers on Sat Nov 14 00:26:33 2026
    From Newsgroup: comp.os.linux.misc

    On 11/13/26 20:09, Lawrence DrCOOliveiro wrote:
    On Fri, 13 Nov 2026 22:02:26 +0100, Alexander Schreiber wrote:

    ... lithium battery chemistries instead of with much safer ones e.g.
    LiFePo4.

    Polonium?!?

    :-)

    Sometimes capitalization does count.

    DO remember when they promised everything would
    be run by atomic batteries that'd last 100 years.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From c186282@c186282@nnada.net to comp.os.linux.misc,alt.folklore.computers on Sat Nov 14 00:29:43 2026
    From Newsgroup: comp.os.linux.misc

    On 11/13/26 16:02, Alexander Schreiber wrote:
    The Natural Philosopher <tnp@invalid.invalid> wrote:
    On 13/11/2026 14:40, Peter Flass wrote:
    On 11/13/26 00:31, Thomas Koenig wrote:
    Charlie Gibbs <cgibbs@kltpzyxm.invalid> schrieb:
    On 2026-11-12, Alexander Schreiber <als@usenet.thangorodrim.de> wrote: >>>>>
    Charlie Gibbs <cgibbs@kltpzyxm.invalid> wrote:

    Many redundant aircraft systems are routed (or re-routed if
    a potential problem is later discovered) to try to prevent
    one incident from damaging them all.

    Indeed. At least aviation is pretty good at learning from previous >>>>>> failures.

    That's one of the things I like about aviation: it's one of
    the few remaining places where safety is taken seriously.

    I work in the chmemical industry, and we certainly take safety
    seriously there.

    But if you're interested in airline safety, I assume you read the
    Aviation Herald?-a One disturbing trend is the increasing number
    of lithium ion battery fires from things that passengers brought
    on board.


    I can't understand (well, I can...) why LiO2 batteries were ever
    approved for general use.

    There is no such battery. Lithium Ion or LIPO for lithium ion polymer or
    LiFePo for Lithium iron polymer

    Well, there are Lithium-Air batteries. But while various implementations
    of them have been demonstrated, AFAIK they are nowhere near commercialization.

    Just the other day I saw on the news that a
    house burned (actually exploded) because if a battery fire.

    Batteries made badly and sold out of spec.

    And/or badly abused during transport, installation, operation. Plus
    people with little knowledge of the technology tend to end up with
    the more exciting (read: fire-prone) lithium battery chemistries instead
    of with much safer ones e.g. LiFePo4. It doesn't help that often the
    cell chemistry is not explicitly mentioned, e.g. with battery packs
    for e-bikes and such.

    Kind regards,
    Alex.

    The best alt these days are sodium batteries ... already
    commercial, you can buy ones meant to start up automobiles
    on Amazon. Not quite the power density, but getting close
    and don't explode.

    As for 'quality batteries', just no WAY to know what
    you're really getting these days.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From c186282@c186282@nnada.net to comp.os.linux.misc,alt.folklore.computers on Sat Nov 14 00:33:28 2026
    From Newsgroup: comp.os.linux.misc

    On 11/13/26 17:24, Johann 'Myrkraverk' Oskarsson wrote:
    On 11/11/2026 12:46 PM, c186282 wrote:
    On 11/10/26 10:14, Scott Lurndal wrote:
    Pancho <Pancho.Jones@protonmail.com> writes:
    On 11/10/26 09:37, Alexander Schreiber wrote:

    No. The operating system and application software ran on the mainframe >>>>> CPU. But - as described above - a key design aspect of mainframes back >>>>> then was that I/O was not done by the CPU (like it was done on home
    computers were all I/O was entirely handled by the CPU), but fully
    off-
    loaded to peripherials that were essentially limited computers of
    their

    "Fully offloaded" is a bit weasel worded. I thought PCs did offload
    some
    duties to io devices, nic chip, wifi chip, gpu card. sound card. Also, >>>> even a mainframe will need service routines to handle communications
    with IO devices. Where to draw the line seems unclear.

    Fully offloaded is accurate.-a For the Burroughs mainframes, a single
    instruction -initiated- an I/O request to a channel.-a-a The IIO (later
    SPIO) instruction would instruct the I/O processor (in the late 70s
    the IOP used an 8085 uProcessor) to initiate an I/O by passing a
    6-digit (24-bit) opcode/variant word and the start and end addresses
    of the buffer.-a-a The IOP would pass the command to a Data-Link Processor >>> (DLP) that handled communications to the device controller (magnetic
    tape, disk) or unit record device (printer, card reader, card punch)
    or external communications processor (RS232/20ma current loop) and wait
    for the controller to start passing data back to the IOP.-a The IOP
    would move the data to the appropriate memory buffer and, when complete, >>> signal the processor that the I/O had completed with an interrupt.

    -a-a Esp in the Bad Old Days, "smart perips" were VERY
    -a-a important. The CPU just sent a LITTLE stuff to them.
    -a-a Very CPU/RAM/Cycle efficient.

    -a-a Flip - left you hostage to the periph "IQ", often
    -a-a no ways to tweak/enhance it all.


    Caveat: I don't have a clue about these things, it is just that what
    you
    are saying doesn't sound quite right.

    -a-a Me ... SEEN it all - the plusses AND minuses.

    -a-a The somewhat older group members will have
    -a-a seen MORE of that.

    -a-a The newbies think NOTHING from the past can
    -a-a possibly inform them, pave the way.

    -a-a Horribly wrong ... but you'll never convince them.
    -a-a For them it's now all "AI" ... until it displaces
    -a-a them entirely.

    1970s "cloud computing" would be lugging around a VT-of-the-day, and a modem.-a Then you could connect it to the hotel's phone line and get some work done while far and away.

    Or at least, that's what I imagine it could have been; I lack the per-
    sonal experience of it.

    But no, no; that's all /new/.-a Cloud computing is hip and cool.


    -a-a Can ya say "Would You Like Fries With That ?" in
    -a-a three or four languages ???-a :-)

    Yes, as a matter of fact I can; English, Icelandic, and French.-a Proba-
    bly more languages in the future.


    Best wishes, and happy terminals!

    %! P.S. mine is VT520, so not a 70s vintage.


    After all the prose, it has become clear that
    "mainframe" is very hard to define. There were
    many players, many models, everything a bit
    different, different emphasis on various
    aspects.




    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Alexander Schreiber@als@usenet.thangorodrim.de to comp.os.linux.misc,alt.folklore.computers on Sat Nov 14 09:48:45 2026
    From Newsgroup: comp.os.linux.misc

    Lawrence DrCOOliveiro <ldo@nz.invalid> wrote:
    On Fri, 13 Nov 2026 22:02:26 +0100, Alexander Schreiber wrote:

    ... lithium battery chemistries instead of with much safer ones e.g.
    LiFePo4.

    Polonium?!?

    Hehe, by special order from the FSB, maybe?

    No, that was supposed to be LiFePO4 - Lithium Iron Phosphate, just a
    single bit error.

    Kind regards,
    Alex.
    --
    "Opportunity is missed by most people because it is dressed in overalls and
    looks like work." -- Thomas A. Edison
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Nuno Silva@nunojsilva@invalid.invalid to comp.os.linux.misc,alt.folklore.computers on Sat Nov 14 09:20:41 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-13, St|-phane CARPENTIER wrote:

    Le 12-11-2026, Alexander Schreiber <als@usenet.thangorodrim.de> a |-crit-a:

    The Windows key can be made useful by remapping it to Compose.

    Yes, but I don't consider it a remap.

    "Context menu" key here. It's funny how these "PC"-style keyboards come
    with so few modifiers, but if it's for GUI or Microsoft purposes, they
    get more keys :-P
    --
    Nuno Silva
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Nuno Silva@nunojsilva@invalid.invalid to comp.os.linux.misc,alt.folklore.computers on Sat Nov 14 10:02:47 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-13, John Ames wrote:

    On 13 Nov 2026 19:58:39 GMT
    St|-phane CARPENTIER <sc@fiat-linux.fr> wrote:

    I've never found a useful way to use a "Window Key" with a Microsoft
    computer. With a Linux computer, this key is very useful. It's used to
    manage my windows knowing that every configuration I make with this
    key won't mess with the shortcut of my programs.

    I use it for window-manager shortcuts for the same reason.

    Under X11, same here. Under WinNT, I tend to use it for some of the
    shortcuts, say, Win+R, Win+E.
    --
    Nuno Silva
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From The Natural Philosopher@tnp@invalid.invalid to comp.os.linux.misc,alt.folklore.computers on Sat Nov 14 12:14:20 2026
    From Newsgroup: comp.os.linux.misc

    On 13/11/2026 21:02, Alexander Schreiber wrote:
    The Natural Philosopher <tnp@invalid.invalid> wrote:
    On 13/11/2026 14:40, Peter Flass wrote:
    On 11/13/26 00:31, Thomas Koenig wrote:
    Charlie Gibbs <cgibbs@kltpzyxm.invalid> schrieb:
    On 2026-11-12, Alexander Schreiber <als@usenet.thangorodrim.de> wrote: >>>>>
    Charlie Gibbs <cgibbs@kltpzyxm.invalid> wrote:

    Many redundant aircraft systems are routed (or re-routed if
    a potential problem is later discovered) to try to prevent
    one incident from damaging them all.

    Indeed. At least aviation is pretty good at learning from previous >>>>>> failures.

    That's one of the things I like about aviation: it's one of
    the few remaining places where safety is taken seriously.

    I work in the chmemical industry, and we certainly take safety
    seriously there.

    But if you're interested in airline safety, I assume you read the
    Aviation Herald?-a One disturbing trend is the increasing number
    of lithium ion battery fires from things that passengers brought
    on board.


    I can't understand (well, I can...) why LiO2 batteries were ever
    approved for general use.

    There is no such battery. Lithium Ion or LIPO for lithium ion polymer or
    LiFePo for Lithium iron polymer

    Well, there are Lithium-Air batteries. But while various implementations
    of them have been demonstrated, AFAIK they are nowhere near commercialization.

    I did not claim the list was exhaustive, I am running some form of
    lithium primary cell with a 1.7V ish terminal voltage.

    There are 3V primaries too. Lthium manganese dioxide...

    Just the other day I saw on the news that a
    house burned (actually exploded) because if a battery fire.

    Batteries made badly and sold out of spec.

    And/or badly abused during transport, installation, operation. Plus
    people with little knowledge of the technology tend to end up with
    the more exciting (read: fire-prone) lithium battery chemistries instead
    of with much safer ones e.g. LiFePo4. It doesn't help that often the
    cell chemistry is not explicitly mentioned, e.g. with battery packs
    for e-bikes and such.

    Kind regards,
    Alex.
    All of the above.

    Look up the amusing video of 'my first ride on my E bike' which ends in
    a white, hot pile of ash.
    --
    "Nature does not give up the winter because people dislike the cold."

    rCo Confucius

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From The Natural Philosopher@tnp@invalid.invalid to comp.os.linux.misc,alt.folklore.computers on Sat Nov 14 12:16:12 2026
    From Newsgroup: comp.os.linux.misc

    On 14/11/2026 05:26, c186282 wrote:
    On 11/13/26 20:09, Lawrence DrCOOliveiro wrote:
    On Fri, 13 Nov 2026 22:02:26 +0100, Alexander Schreiber wrote:

    ... lithium battery chemistries instead of with much safer ones e.g.
    LiFePo4.

    Polonium?!?

    -a :-)

    -a Sometimes capitalization does count.

    Ooops, My bad. Its phosphate too, not polymer

    -a DO remember when they promised everything would
    -a be run by atomic batteries that'd last 100 years.

    Fir some reason battery manufacturers were dead set against that idea.
    --
    rCLIt is not the truth of Marxism that explains the willingness of intellectuals to believe it, but the power that it confers on
    intellectuals, in their attempts to control the world. And since...it is futile to reason someone out of a thing that he was not reasoned into,
    we can conclude that Marxism owes its remarkable power to survive every criticism to the fact that it is not a truth-directed but a
    power-directed system of thought.rCY
    Sir Roger Scruton

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From =?UTF-8?Q?St=C3=A9phane?= CARPENTIER@sc@fiat-linux.fr to comp.os.linux.misc,alt.folklore.computers on Sat Nov 14 13:44:27 2026
    From Newsgroup: comp.os.linux.misc

    Le 12-11-2026, Alexander Schreiber <als@usenet.thangorodrim.de> a |-crit-a:
    Lars Poulsen <lars@beagle-ears.com> wrote:

    On a transoceanic route, time in the air is quite significant.
    On a short "puddlejumper" route, a high-speed train (TGV in French) may
    well be faster.

    In fact the rollout of the TGV network largely killed off domestic flights
    in France, e.g. the Paris-Lyon TGV line opening in 1981 caused domestic
    air travel on that corridor to drop by 60% in one year.

    I never heard about anyone taking a plane from Paris to Lyon. I didn't
    knew that was possible, so I just checked and it still exists.

    And once you factor
    in all the overhead (not flying) time of air travel, taking a train can actually be faster on shorter (domestic) routes. It tends to be more convenient too.

    Now a TVG takes around 2h00 to go from Paris to Lyon. And a plane takes something like 1h10. But, as you said, you have to take the onboarding
    time which is far longer on a plane than on a train. And the Charles de
    Gaulle airport is not in Paris but is at something like 30km from it.
    And the Saint Exup|-ry airport is not in Lyon but is at something like
    30km from it.

    So, the plane is faster than the TGV, but it takes more time to go from
    Paris to Lyon on plane than on TGV. Maybe that's why I never heard about
    anyone taking the plane to go from Paris to Lyon. It's neither cheaper
    nor more comfortable. Maybe it can be taken by someone coming from far
    away, with no plane to Lyon and whose only possibility is to go to
    Paris. It can be an opportunity to take the plane directly from the
    airport but I'm not entirely convinced.
    --
    Si vous avez du temps |a perdre :
    https://scarpet42.gitlab.io
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lars Poulsen@lars@beagle-ears.com to comp.os.linux.misc,alt.folklore.computers on Sat Nov 14 05:49:52 2026
    From Newsgroup: comp.os.linux.misc

    On 11/13/2026 14:24, Johann 'Myrkraverk' Oskarsson wrote:
    1970s "cloud computing" would be lugging around a VT-of-the-day, and a modem.-a Then you could connect it to the hotel's phone line and get some work done while far and away.

    Or at least, that's what I imagine it could have been; I lack the per-
    sonal experience of it.

    But no, no; that's all /new/.-a Cloud computing is hip and cool.

    That is a far distant glimmer of a memory. Would not have been at a
    hotel ... but I might have lugged an ADM-3a to a field visit to a
    customer site. We did have dial-in lines when I was at the computer
    center in the early 70's, but I don't think anyone used them from mobile terminals. And when I went to the bespoke interfacing engineering
    company in 1975, we did NOT have any system with dial-in capability, nor
    did we have any networking with email or the like. FAX was the closest
    thing to email.

    In contrast, I distinctly remember that we had dial-in lines on the UNIX systems at ACC when I got to California. After a while (1981?), I got an LS-120 at home; at first I used an acoustic coupler, but soon after
    that, I got a modem; probably a 300 bps.
    ...

    %! P.S. mine is VT520, so not a 70s vintage.

    I see what you did there!
    --
    Lars Poulsen - an old geek in Santa Barbara, California
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From The Natural Philosopher@tnp@invalid.invalid to comp.os.linux.misc,alt.folklore.computers on Sat Nov 14 14:05:40 2026
    From Newsgroup: comp.os.linux.misc

    On 14/11/2026 13:44, St|-phane CARPENTIER wrote:
    So, the plane is faster than the TGV, but it takes more time to go from
    Paris to Lyon on plane than on TGV. Maybe that's why I never heard about anyone taking the plane to go from Paris to Lyon. It's neither cheaper
    nor more comfortable. Maybe it can be taken by someone coming from far
    away, with no plane to Lyon and whose only possibility is to go to
    Paris. It can be an opportunity to take the plane directly from the
    airport but I'm not entirely convinced.

    Typically a lot of short haul planes - at least in the UK - are ferries
    for people using the larger airports as staring points for international travel.

    But everybody does the sums. E,g a taxi ride from Glasgow City to the
    airport, an hour down to London Stansted and the express train to
    Liverpool street plus a short taxi ride will get you into a 10 a.m.
    meeting in the city from a 7:30 am start.

    OTOH it used to take me nearly as long to get to Stansted from where I
    live, go through all the BS and get on a plane to Brussels, as it did to
    drive 100 miles to the ferry and another 100 miles the other end...

    Borders, security checks, availability of ferries etc etc.It all bends
    you one way or another, as does cost and convenience
    --
    New Socialism consists essentially in being seen to have your heart in
    the right place whilst your head is in the clouds and your hand is in
    someone else's pocket.


    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Johann 'Myrkraverk' Oskarsson@johann@myrkraverk.invalid to comp.os.linux.misc,alt.folklore.computers on Sat Nov 14 23:12:34 2026
    From Newsgroup: comp.os.linux.misc

    On 11/14/2026 1:33 PM, c186282 wrote:
    On 11/13/26 17:24, Johann 'Myrkraverk' Oskarsson wrote:
    On 11/11/2026 12:46 PM, c186282 wrote:
    On 11/10/26 10:14, Scott Lurndal wrote:
    Pancho <Pancho.Jones@protonmail.com> writes:
    On 11/10/26 09:37, Alexander Schreiber wrote:

    No. The operating system and application software ran on the
    mainframe
    CPU. But - as described above - a key design aspect of mainframes >>>>>> back
    then was that I/O was not done by the CPU (like it was done on home >>>>>> computers were all I/O was entirely handled by the CPU), but fully >>>>>> off-
    loaded to peripherials that were essentially limited computers of >>>>>> their

    "Fully offloaded" is a bit weasel worded. I thought PCs did offload >>>>> some
    duties to io devices, nic chip, wifi chip, gpu card. sound card. Also, >>>>> even a mainframe will need service routines to handle communications >>>>> with IO devices. Where to draw the line seems unclear.

    Fully offloaded is accurate.-a For the Burroughs mainframes, a single
    instruction -initiated- an I/O request to a channel.-a-a The IIO (later >>>> SPIO) instruction would instruct the I/O processor (in the late 70s
    the IOP used an 8085 uProcessor) to initiate an I/O by passing a
    6-digit (24-bit) opcode/variant word and the start and end addresses
    of the buffer.-a-a The IOP would pass the command to a Data-Link
    Processor
    (DLP) that handled communications to the device controller (magnetic
    tape, disk) or unit record device (printer, card reader, card punch)
    or external communications processor (RS232/20ma current loop) and wait >>>> for the controller to start passing data back to the IOP.-a The IOP
    would move the data to the appropriate memory buffer and, when
    complete,
    signal the processor that the I/O had completed with an interrupt.

    -a-a Esp in the Bad Old Days, "smart perips" were VERY
    -a-a important. The CPU just sent a LITTLE stuff to them.
    -a-a Very CPU/RAM/Cycle efficient.

    -a-a Flip - left you hostage to the periph "IQ", often
    -a-a no ways to tweak/enhance it all.


    Caveat: I don't have a clue about these things, it is just that
    what you
    are saying doesn't sound quite right.

    -a-a Me ... SEEN it all - the plusses AND minuses.

    -a-a The somewhat older group members will have
    -a-a seen MORE of that.

    -a-a The newbies think NOTHING from the past can
    -a-a possibly inform them, pave the way.

    -a-a Horribly wrong ... but you'll never convince them.
    -a-a For them it's now all "AI" ... until it displaces
    -a-a them entirely.

    1970s "cloud computing" would be lugging around a VT-of-the-day, and a
    modem.-a Then you could connect it to the hotel's phone line and get some
    work done while far and away.

    Or at least, that's what I imagine it could have been; I lack the per-
    sonal experience of it.

    But no, no; that's all /new/.-a Cloud computing is hip and cool.


    -a-a Can ya say "Would You Like Fries With That ?" in
    -a-a three or four languages ???-a :-)

    Yes, as a matter of fact I can; English, Icelandic, and French.-a Proba-
    bly more languages in the future.


    Best wishes, and happy terminals!

    %! P.S. mine is VT520, so not a 70s vintage.


    -a After all the prose, it has become clear that
    -a "mainframe" is very hard to define. There were
    -a many players, many models, everything a bit
    -a different, different emphasis on various
    -a aspects.

    I read somewhere that /mainframe/ got its name from the -- wait for it
    -- main frame of the large vacuum tube computers of the 1950s, if not
    40s. I don't recall seeing pictures of these /main frames/, so I don't
    know what they looked like.

    It's just one of these computing terms that's not quite shrouded in mys-
    tery but few people know the etymology of it.


    Best wishes, and happy mainframes!
    --
    Johann | email: invalid -> com | http://www.myrkraverk.com/blog/
    I'm not from the Internet, I just work there. | via Easynews.com https://bsky.app/profile/myrkraverk.bsky.social | for ( ;; ) _:;
    Federated at https://fed.brid.gy/bsky/myrkraverk.bsky.social
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Thomas Koenig@tkoenig@netcologne.de to comp.os.linux.misc,alt.folklore.computers on Sat Nov 14 15:16:20 2026
    From Newsgroup: comp.os.linux.misc

    Lars Poulsen <lars@beagle-ears.com> schrieb:
    FAX was the closest
    thing to email.

    When faxes were introduced, my father liked them a *lot*.

    He worked for a big chemical company, and sending letters in
    within the site took a long time (could be serval days) and was
    very unreliable because the internal mail serivce was... well,
    it could have been better.

    Sending a fax from one office to the other solved that problem, and
    it didn't even cost extra because on-site phone calls (and, of course,
    faxes) were free.

    In contrast, I distinctly remember that we had dial-in lines on the UNIX systems at ACC when I got to California. After a while (1981?), I got an LS-120 at home; at first I used an acoustic coupler, but soon after
    that, I got a modem; probably a 300 bps.

    You had free local calls, I presume? In Germany, it was much
    more expensive at the time.
    --
    This USENET posting was made without artificial intelligence,
    artificial impertinence, artificial arrogance, artificial stupidity,
    artificial flavorings or artificial colorants.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Johann 'Myrkraverk' Oskarsson@johann@myrkraverk.invalid to comp.os.linux.misc,alt.folklore.computers on Sat Nov 14 23:29:12 2026
    From Newsgroup: comp.os.linux.misc

    On 11/14/2026 9:44 PM, St|-phane CARPENTIER wrote:
    Le 12-11-2026, Alexander Schreiber <als@usenet.thangorodrim.de> a |-crit-a:
    Lars Poulsen <lars@beagle-ears.com> wrote:

    On a transoceanic route, time in the air is quite significant.
    On a short "puddlejumper" route, a high-speed train (TGV in French) may
    well be faster.

    In fact the rollout of the TGV network largely killed off domestic flights >> in France, e.g. the Paris-Lyon TGV line opening in 1981 caused domestic
    air travel on that corridor to drop by 60% in one year.

    I never heard about anyone taking a plane from Paris to Lyon. I didn't
    knew that was possible, so I just checked and it still exists.

    Based on personal experience, these flights serve the transfer passen-
    gers mostly [1]. I doubt people traveling from Paris to Lyon would
    choose flight over the TGV /just because/. But New York to Lyon is a
    different story.


    And once you factor
    in all the overhead (not flying) time of air travel, taking a train can
    actually be faster on shorter (domestic) routes. It tends to be more
    convenient too.

    Now a TVG takes around 2h00 to go from Paris to Lyon. And a plane takes something like 1h10. But, as you said, you have to take the onboarding
    time which is far longer on a plane than on a train. And the Charles de Gaulle airport is not in Paris but is at something like 30km from it.
    And the Saint Exup|-ry airport is not in Lyon but is at something like
    30km from it.

    So, the plane is faster than the TGV, but it takes more time to go from
    Paris to Lyon on plane than on TGV. Maybe that's why I never heard about anyone taking the plane to go from Paris to Lyon. It's neither cheaper
    nor more comfortable. Maybe it can be taken by someone coming from far
    away, with no plane to Lyon and whose only possibility is to go to
    Paris. It can be an opportunity to take the plane directly from the
    airport but I'm not entirely convinced.


    My personal experience is Aix-en-Provence to Reykjavik, Iceland. Well, technically the international airport is in a different municipality,
    but nobody knows /KEF/ nor /Keflavik/ anymore. It was a flight to Paris
    and from there to Iceland, if I remember correctly; and no, I don't rem-
    ember the departing airport, but probably /MRS/ based on recent search.

    I had reached Aix-en-Provence via the TGV, but didn't want the hassle of
    inner city Paris to the airport while going back.

    Keep in mind that when people travel with a lot of luggage, not having
    to check it out, and then take a taxi or whatever to the train station
    can be worth it. Then there are considerations such as the expense of
    a hotel for one night, plus all the food, for those not traveling in a
    hurry -- that's been a more recent experience of mine.

    But, for people not in hurry, and have the budget, it's is "better" for
    some value of better to stay one or more nights in a hotel and then move
    on to the next destination; it's less stressful.


    Best wishes, and happy traveling!

    [1] Well, that's somewhat an extrapolation of n=1, but rings true.
    --
    Johann | email: invalid -> com | http://www.myrkraverk.com/blog/
    I'm not from the Internet, I just work there. | via Easynews.com https://bsky.app/profile/myrkraverk.bsky.social | for ( ;; ) _:;
    Federated at https://fed.brid.gy/bsky/myrkraverk.bsky.social
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From The Natural Philosopher@tnp@invalid.invalid to comp.os.linux.misc,alt.folklore.computers on Sat Nov 14 17:51:31 2026
    From Newsgroup: comp.os.linux.misc

    On 14/11/2026 15:12, Johann 'Myrkraverk' Oskarsson wrote:
    I read somewhere that /mainframe/ got its name from the -- wait for it
    -- main frame of the large vacuum tube computers of the 1950s, if not
    40s.-a I don't recall seeing pictures of these /main frames/, so I don't
    know what they looked like.

    a computer built in multiple 19" racks?



    It's just one of these computing terms that's not quite shrouded in mys-
    tery but few people know the etymology of it.
    --
    It is the folly of too many to mistake the echo of a London coffee-house
    for the voice of the kingdom.

    Jonathan Swift


    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Charlie Gibbs@cgibbs@kltpzyxm.invalid to comp.os.linux.misc,alt.folklore.computers on Sat Nov 14 18:34:33 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-14, The Natural Philosopher <tnp@invalid.invalid> wrote:

    On 14/11/2026 15:12, Johann 'Myrkraverk' Oskarsson wrote:

    I read somewhere that /mainframe/ got its name from the -- wait for it
    -- main frame of the large vacuum tube computers of the 1950s, if not
    40s.-a I don't recall seeing pictures of these /main frames/, so I don't
    know what they looked like.

    a computer built in multiple 19" racks?

    What about all those DEC minis built in 19" racks?

    Many mainframes were built in bespoke racks, often with
    backplanes on swing-out frames. Much more exclusive and sexy.

    Whee, this is becoming fun...
    --
    /~\ Charlie Gibbs | In this world there are
    \ / <cgibbs@kltpzyxm.invalid> | two kinds of people:
    X I'm really at ac.dekanfrus | 1. Those who can extrapolate
    / \ if you read it the right way. | from incomplete data.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From John Levine@johnl@taugh.com to comp.os.linux.misc,alt.folklore.computers on Sat Nov 14 20:50:06 2026
    From Newsgroup: comp.os.linux.misc

    It appears that Charlie Gibbs <cgibbs@kltpzyxm.invalid> said:
    On 2026-11-14, The Natural Philosopher <tnp@invalid.invalid> wrote:

    On 14/11/2026 15:12, Johann 'Myrkraverk' Oskarsson wrote:

    I read somewhere that /mainframe/ got its name from the -- wait for it
    -- main frame of the large vacuum tube computers of the 1950s, if not
    40s.-a I don't recall seeing pictures of these /main frames/, so I don't >>> know what they looked like.

    a computer built in multiple 19" racks?

    Yes indeed, adapted from racks used in phone exchanges.

    What about all those DEC minis built in 19" racks?

    Well, sure, if it was too big to sit on a desk they might as well
    use (so to speak) off the shelf racks.

    Many mainframes were built in bespoke racks, often with
    backplanes on swing-out frames. Much more exclusive and sexy.

    I think that is also to minimize cable lengths, and probably has
    something to do with the piping for water cooling.
    --
    Regards,
    John Levine, johnl@taugh.com, Primary Perpetrator of "The Internet for Dummies",
    Please consider the environment before reading this e-mail. https://jl.ly
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From John Levine@johnl@taugh.com to comp.os.linux.misc,alt.folklore.computers on Sat Nov 14 20:58:47 2026
    From Newsgroup: comp.os.linux.misc

    According to Johann 'Myrkraverk' Oskarsson <johann@myrkraverk.invalid>:
    In fact the rollout of the TGV network largely killed off domestic flights >>> in France, e.g. the Paris-Lyon TGV line opening in 1981 caused domestic
    air travel on that corridor to drop by 60% in one year.

    I never heard about anyone taking a plane from Paris to Lyon. I didn't
    knew that was possible, so I just checked and it still exists.

    Based on personal experience, these flights serve the transfer passen-
    gers mostly [1]. I doubt people traveling from Paris to Lyon would
    choose flight over the TGV /just because/. But New York to Lyon is a >different story.

    Before the TGV, Paris to Lyon by train took over four hours, now it's about 2. Air France still has five flights a day from Lyon to CDG which take a little over an hour, I presume only for people connecting to long haul flights.

    If you're going city center to city center, the train is likely to be
    faster than the plane since there's no faffing about with air security
    and no long RER trip from CDG into the city.
    --
    Regards,
    John Levine, johnl@taugh.com, Primary Perpetrator of "The Internet for Dummies",
    Please consider the environment before reading this e-mail. https://jl.ly
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lawrence =?iso-8859-13?q?D=FFOliveiro?=@ldo@nz.invalid to comp.os.linux.misc,alt.folklore.computers on Sat Nov 14 22:08:01 2026
    From Newsgroup: comp.os.linux.misc

    On Sat, 14 Nov 2026 05:49:52 -0800, Lars Poulsen wrote:

    After a while (1981?), I got an LS-120 at home ...

    <https://en.wikipedia.org/wiki/LS-120>?

    <https://en.wikipedia.org/wiki/DECwriter>!
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Charlie Gibbs@cgibbs@kltpzyxm.invalid to comp.os.linux.misc,alt.folklore.computers on Sat Nov 14 23:19:29 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-14, John Levine <johnl@taugh.com> wrote:

    It appears that Charlie Gibbs <cgibbs@kltpzyxm.invalid> said:

    On 2026-11-14, The Natural Philosopher <tnp@invalid.invalid> wrote:

    On 14/11/2026 15:12, Johann 'Myrkraverk' Oskarsson wrote:

    I read somewhere that /mainframe/ got its name from the -- wait for it >>>> -- main frame of the large vacuum tube computers of the 1950s, if not
    40s.-a I don't recall seeing pictures of these /main frames/, so I don't >>>> know what they looked like.

    a computer built in multiple 19" racks?

    Yes indeed, adapted from racks used in phone exchanges.

    What about all those DEC minis built in 19" racks?

    Well, sure, if it was too big to sit on a desk they might as well
    use (so to speak) off the shelf racks.

    :-)

    Many mainframes were built in bespoke racks, often with
    backplanes on swing-out frames. Much more exclusive and sexy.

    I think that is also to minimize cable lengths, and probably has
    something to do with the piping for water cooling.

    I never worked on water-cooled machines, but the swing-out frames
    allowed a lot of circuitry to be packed into a box, and gave easy
    access to everything.
    --
    /~\ Charlie Gibbs | In this world there are
    \ / <cgibbs@kltpzyxm.invalid> | two kinds of people:
    X I'm really at ac.dekanfrus | 1. Those who can extrapolate
    / \ if you read it the right way. | from incomplete data.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From c186282@c186282@nnada.net to comp.os.linux.misc,alt.folklore.computers on Sat Nov 14 23:18:00 2026
    From Newsgroup: comp.os.linux.misc

    On 11/14/26 10:12, Johann 'Myrkraverk' Oskarsson wrote:
    On 11/14/2026 1:33 PM, c186282 wrote:
    On 11/13/26 17:24, Johann 'Myrkraverk' Oskarsson wrote:
    On 11/11/2026 12:46 PM, c186282 wrote:
    On 11/10/26 10:14, Scott Lurndal wrote:
    Pancho <Pancho.Jones@protonmail.com> writes:
    On 11/10/26 09:37, Alexander Schreiber wrote:

    No. The operating system and application software ran on the
    mainframe
    CPU. But - as described above - a key design aspect of mainframes >>>>>>> back
    then was that I/O was not done by the CPU (like it was done on home >>>>>>> computers were all I/O was entirely handled by the CPU), but
    fully off-
    loaded to peripherials that were essentially limited computers of >>>>>>> their

    "Fully offloaded" is a bit weasel worded. I thought PCs did
    offload some
    duties to io devices, nic chip, wifi chip, gpu card. sound card.
    Also,
    even a mainframe will need service routines to handle communications >>>>>> with IO devices. Where to draw the line seems unclear.

    Fully offloaded is accurate.-a For the Burroughs mainframes, a single >>>>> instruction -initiated- an I/O request to a channel.-a-a The IIO (later >>>>> SPIO) instruction would instruct the I/O processor (in the late 70s
    the IOP used an 8085 uProcessor) to initiate an I/O by passing a
    6-digit (24-bit) opcode/variant word and the start and end addresses >>>>> of the buffer.-a-a The IOP would pass the command to a Data-Link
    Processor
    (DLP) that handled communications to the device controller (magnetic >>>>> tape, disk) or unit record device (printer, card reader, card punch) >>>>> or external communications processor (RS232/20ma current loop) and
    wait
    for the controller to start passing data back to the IOP.-a The IOP
    would move the data to the appropriate memory buffer and, when
    complete,
    signal the processor that the I/O had completed with an interrupt.

    -a-a Esp in the Bad Old Days, "smart perips" were VERY
    -a-a important. The CPU just sent a LITTLE stuff to them.
    -a-a Very CPU/RAM/Cycle efficient.

    -a-a Flip - left you hostage to the periph "IQ", often
    -a-a no ways to tweak/enhance it all.


    Caveat: I don't have a clue about these things, it is just that
    what you
    are saying doesn't sound quite right.

    -a-a Me ... SEEN it all - the plusses AND minuses.

    -a-a The somewhat older group members will have
    -a-a seen MORE of that.

    -a-a The newbies think NOTHING from the past can
    -a-a possibly inform them, pave the way.

    -a-a Horribly wrong ... but you'll never convince them.
    -a-a For them it's now all "AI" ... until it displaces
    -a-a them entirely.

    1970s "cloud computing" would be lugging around a VT-of-the-day, and a
    modem.-a Then you could connect it to the hotel's phone line and get some >>> work done while far and away.

    Or at least, that's what I imagine it could have been; I lack the per-
    sonal experience of it.

    But no, no; that's all /new/.-a Cloud computing is hip and cool.


    -a-a Can ya say "Would You Like Fries With That ?" in
    -a-a three or four languages ???-a :-)

    Yes, as a matter of fact I can; English, Icelandic, and French.-a Proba- >>> bly more languages in the future.


    Best wishes, and happy terminals!

    %! P.S. mine is VT520, so not a 70s vintage.


    -a-a After all the prose, it has become clear that
    -a-a "mainframe" is very hard to define. There were
    -a-a many players, many models, everything a bit
    -a-a different, different emphasis on various
    -a-a aspects.

    I read somewhere that /mainframe/ got its name from the -- wait for it
    -- main frame of the large vacuum tube computers of the 1950s, if not
    40s.-a I don't recall seeing pictures of these /main frames/, so I don't
    know what they looked like.

    Yep - big "frames" of low-capacity stuff :-)

    The first UNIVACs are the model for this.

    It's just one of these computing terms that's not quite shrouded in mys-
    tery but few people know the etymology of it.


    Best wishes, and happy mainframes!

    As said, we all have our own idea of a "mainframe",
    what it is, what it does, how it does it.

    All the olde-tyme makers - as said, many models,
    many points of focus, many methods. "Mainframe"
    is more an IDEA than a real THING.

    These days, a 'mainframe' probably means a CLUSTER
    computer set-up.

    At minimum, the term refers to the biggest/fastest
    shit around.

    IBM still sells "mainframes" for big biz/banking/science.
    Seen pix, big black box(es) packed with the fastest boards
    to be had.

    "Peripherials" are no longer an emphasis - it's vast
    net connectivity and parallel tasks instead.

    The IBM spin of RHEL is the favorite OS for its
    'mainframes'. Utils/drivers for LARGE scale
    cluster/distributed processing. "AI" does not
    do EVERYTHING well yet.

    Note "Plan-9" has been ported to the IBM boxes.
    It was/is a good cluster/distributed OS with
    a UNIX-ish look and feel. P9 is *wasted* on a
    single PC/mini/mainframe, that's not what it
    was writ for.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lawrence =?iso-8859-13?q?D=FFOliveiro?=@ldo@nz.invalid to comp.os.linux.misc,alt.folklore.computers on Sun Nov 15 05:57:04 2026
    From Newsgroup: comp.os.linux.misc

    On Fri, 13 Nov 2026 15:05:00 +0100, Alexander Schreiber wrote:

    Lawrence DrCOOliveiro <ldo@nz.invalid> wrote:

    Perception doesnrCOt count. ThatrCOs reality. You canrCOt *pretend* to
    offer interactivity: you really have to have support for
    interactivity built into your architecture from the lowest levels
    up. You canrCOt build per-character response on top of block mode.

    You absolutely can and the interactive web proves it.

    The interactive web is not run on mainframe-style block-mode
    terminals.

    WebSocket connections are asynchronous full-duplex connections --
    thatrCOs not rCLRPCrCY by any stretch.

    Of course it is. RPCs come in various kinds, from the classic
    one-shot (establish connection, send request, wait for and read
    response, close connection) to streaming RPCs.

    rCLRPCrCY stands for rCLRemote Procedure CallrCY. ItrCOs emulating a procedure call over a network connection. You know how a procedure call works,
    donrCOt you? The caller transfers control to the callee, waits for the
    callee to do its thing and return a result, whereupon the caller
    resumes execution again.

    And nobody ever described block-mode mainframe terminals as running
    any kind of rCLRPCrCY either. The rCLRPCrCY concept was only invented after mainframes had been around for a few decades.

    Before getting even deeper into your love for RPCs and seeing every
    kind of network protocol as RPCs, maybe you should read this <https://www.cs.vu.nl/~ast/Publications/Papers/euteco-1988.pdf>.

    Even HTTP, which is not exactly the most efficient RPC method, came
    up with streaming RPCs: HTTP 1.1 pipelining, where multiple requests
    are handled within the same TCP connection.

    Those donrCOt work for RPCs, though.

    And "asynchronous" just means you don't do blocking reads and
    instead use the framework equivalent of select(2) for a more fluid interaction.

    Well, maybe not select(2), maybe something more modern?
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From rbowman@bowman@montana.com to comp.os.linux.misc,alt.folklore.computers on Sun Nov 15 06:44:13 2026
    From Newsgroup: comp.os.linux.misc

    On Sun, 15 Nov 2026 05:57:04 -0000 (UTC), Lawrence DrCOOliveiro wrote:

    And nobody ever described block-mode mainframe terminals as running any
    kind of rCLRPCrCY either. The rCLRPCrCY concept was only invented after mainframes had been around for a few decades.

    I don't have a timeline but I mostly worked with ONC-RPC, which was a Sun protocol from the '80s. XDR had its oddities but was fairly straight
    forward. In later years using network byte order on x86 only systems was slightly inefficient.

    Next was XML-RPC that grew into SOAP and was supported by the wondrous
    Windows Communication Framework. imnsho it was a steaming pile of shit and
    if I never see another wsdl I'll be happy.

    Then we moved on to an Angular SPA and I learned the joys of protobufs.
    From my viewpoint that was almost back to ONC-RPC without all the
    portmapper complexity. We never used the word gRPC but it was probably
    close.

    Then there is the who DCE/RPC family that Microsoft derived MSRPC from and used for COM and DCOM.

    Many strange flavors of RPC.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From The Natural Philosopher@tnp@invalid.invalid to comp.os.linux.misc,alt.folklore.computers on Sun Nov 15 09:52:56 2026
    From Newsgroup: comp.os.linux.misc

    On 14/11/2026 18:34, Charlie Gibbs wrote:
    On 2026-11-14, The Natural Philosopher <tnp@invalid.invalid> wrote:

    On 14/11/2026 15:12, Johann 'Myrkraverk' Oskarsson wrote:

    I read somewhere that /mainframe/ got its name from the -- wait for it
    -- main frame of the large vacuum tube computers of the 1950s, if not
    40s.-a I don't recall seeing pictures of these /main frames/, so I don't >>> know what they looked like.

    a computer built in multiple 19" racks?

    What about all those DEC minis built in 19" racks?

    Many mainframes were built in bespoke racks, often with
    backplanes on swing-out frames. Much more exclusive and sexy.

    Whee, this is becoming fun...

    Indeed. Originally it was what it was, but it became

    1. A computer able to handled thousands of user sessions
    or
    2. A computer with massive storage and the ability to do payroll
    batches in time for payday.

    Naturally these are very large organisation solutions, so not many
    people were exposed to them. Which is why so many people think these
    machines no longer exist
    --
    ItrCOs easier to fool people than to convince them that they have been fooled. Mark Twain



    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From The Natural Philosopher@tnp@invalid.invalid to comp.os.linux.misc,alt.folklore.computers on Sun Nov 15 09:58:14 2026
    From Newsgroup: comp.os.linux.misc

    On 14/11/2026 23:19, Charlie Gibbs wrote:
    I never worked on water-cooled machines, but the swing-out frames
    allowed a lot of circuitry to be packed into a box, and gave easy
    access to everything.

    At the expense of needing a fair amount of cooling..
    There were not many MIPS per Watt in tjhe days of TTL...
    --
    rCLPuritanism: The haunting fear that someone, somewhere, may be happy.rCY

    H.L. Mencken, A Mencken Chrestomathy

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Johann 'Myrkraverk' Oskarsson@johann@myrkraverk.invalid to comp.os.linux.misc,alt.folklore.computers on Sun Nov 15 18:04:33 2026
    From Newsgroup: comp.os.linux.misc

    On 11/15/2026 2:44 PM, rbowman wrote:
    On Sun, 15 Nov 2026 05:57:04 -0000 (UTC), Lawrence DrCOOliveiro wrote:

    And nobody ever described block-mode mainframe terminals as running any
    kind of rCLRPCrCY either. The rCLRPCrCY concept was only invented after
    mainframes had been around for a few decades.

    I don't have a timeline but I mostly worked with ONC-RPC, which was a Sun protocol from the '80s. XDR had its oddities but was fairly straight
    forward. In later years using network byte order on x86 only systems was slightly inefficient.

    Next was XML-RPC that grew into SOAP and was supported by the wondrous Windows Communication Framework. imnsho it was a steaming pile of shit and
    if I never see another wsdl I'll be happy.

    Then we moved on to an Angular SPA and I learned the joys of protobufs.
    From my viewpoint that was almost back to ONC-RPC without all the
    portmapper complexity. We never used the word gRPC but it was probably
    close.

    Then there is the who DCE/RPC family that Microsoft derived MSRPC from and used for COM and DCOM.

    Many strange flavors of RPC.

    I have worked with the following subsets of flavours [sic].

    * SQL
    * Websocket
    * HTTP with XML and JSON
    * ASCII over TCP

    And possibly a few others that don't make the recollection. It's always
    hard to remember anything beyond the first three -- something about how
    the mind works. I have never professionally done ASCII over serial, but
    the hardware was said to support that too.

    [Encryption such as SSL doesn't change the protocol, and just adds a
    step so I don't include that, even when certificates are used.]

    JSON is probably the most popular today. I personally don't like it for
    being wasteful.
    --
    Johann | email: invalid -> com | http://www.myrkraverk.com/blog/
    I'm not from the Internet, I just work there. | via Easynews.com https://bsky.app/profile/myrkraverk.bsky.social | for ( ;; ) _:;
    Federated at https://fed.brid.gy/bsky/myrkraverk.bsky.social
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Johann 'Myrkraverk' Oskarsson@johann@myrkraverk.invalid to comp.os.linux.misc,alt.folklore.computers on Sun Nov 15 18:33:34 2026
    From Newsgroup: comp.os.linux.misc

    On 11/15/2026 5:52 PM, The Natural Philosopher wrote:
    On 14/11/2026 18:34, Charlie Gibbs wrote:
    On 2026-11-14, The Natural Philosopher <tnp@invalid.invalid> wrote:

    On 14/11/2026 15:12, Johann 'Myrkraverk' Oskarsson wrote:

    I read somewhere that /mainframe/ got its name from the -- wait for it >>>> -- main frame of the large vacuum tube computers of the 1950s, if not
    40s.-a I don't recall seeing pictures of these /main frames/, so I don't >>>> know what they looked like.

    a computer built in multiple 19" racks?

    What about all those DEC minis built in 19" racks?

    Many mainframes were built in bespoke racks, often with
    backplanes on swing-out frames.-a Much more exclusive and sexy.

    Whee, this is becoming fun...

    Indeed. Originally it was what it was, but it became

    1. A computer able to handled thousands of user sessions
    or
    2.-a A computer with massive storage and the ability to do payroll
    batches in time for payday.

    Naturally these are very large organisation solutions, so not many
    people were exposed to them. Which is why so many people think these machines no longer exist



    The problem of /how do we get $amount into every employee's bank ac-
    count/ with thousands of employees and perhaps a single payroll account
    for the company does not sound like a solution waiting for a cluster of
    Linux boxes.
    --
    Johann | email: invalid -> com | http://www.myrkraverk.com/blog/
    I'm not from the Internet, I just work there. | via Easynews.com https://bsky.app/profile/myrkraverk.bsky.social | for ( ;; ) _:;
    Federated at https://fed.brid.gy/bsky/myrkraverk.bsky.social
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From The Natural Philosopher@tnp@invalid.invalid to comp.os.linux.misc,alt.folklore.computers on Sun Nov 15 10:48:39 2026
    From Newsgroup: comp.os.linux.misc

    On 15/11/2026 10:04, Johann 'Myrkraverk' Oskarsson wrote:
    JSON is probably the most popular today.-a I personally don't like it for being wasteful.

    There are enough bugs in JavaScript without introducing yet another layer...
    I use the minimum pure JavaScript sufficient to get the job done
    --
    There is something fascinating about science. One gets such wholesale
    returns of conjecture out of such a trifling investment of fact.

    Mark Twain

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From The Natural Philosopher@tnp@invalid.invalid to comp.os.linux.misc,alt.folklore.computers on Sun Nov 15 10:56:12 2026
    From Newsgroup: comp.os.linux.misc

    On 15/11/2026 10:33, Johann 'Myrkraverk' Oskarsson wrote:
    On 11/15/2026 5:52 PM, The Natural Philosopher wrote:
    On 14/11/2026 18:34, Charlie Gibbs wrote:
    On 2026-11-14, The Natural Philosopher <tnp@invalid.invalid> wrote:

    On 14/11/2026 15:12, Johann 'Myrkraverk' Oskarsson wrote:

    I read somewhere that /mainframe/ got its name from the -- wait for it >>>>> -- main frame of the large vacuum tube computers of the 1950s, if not >>>>> 40s.-a I don't recall seeing pictures of these /main frames/, so I
    don't
    know what they looked like.

    a computer built in multiple 19" racks?

    What about all those DEC minis built in 19" racks?

    Many mainframes were built in bespoke racks, often with
    backplanes on swing-out frames.-a Much more exclusive and sexy.

    Whee, this is becoming fun...

    Indeed. Originally it was what it was, but it became

    1. A computer able to handled thousands of user sessions
    or
    2.-a A computer with massive storage and the ability to do payroll
    batches in time for payday.

    Naturally these are very large organisation solutions, so not many
    people were exposed to them. Which is why so many people think these
    machines no longer exist



    The problem of /how do we get $amount into every employee's bank ac-
    count/ with thousands of employees and perhaps a single payroll account
    for the company does not sound like a solution waiting for a cluster of
    Linux boxes.

    That was essentially my point.

    Neither is necessarily running a global airline booking application.

    We have a range of tools at our disposal, from a Raspberry Pi Pico to an
    IBM mainframe, and none of them are suitable for *all* applications.

    In the middle, a linux *86 style machine is a very useful and cost
    effective machine, but not at the spectrum edges.

    But by definition, apart from wintrolls, it is what most of us will be
    using in *this* NG (comp.os.linux.misc)
    --
    There is something fascinating about science. One gets such wholesale
    returns of conjecture out of such a trifling investment of fact.

    Mark Twain

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Ian@${send-direct-email-to-news1021-at-jusme-dot-com-if-you-must}@jusme.com to comp.os.linux.misc,alt.folklore.computers on Sun Nov 15 12:22:30 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-15, Johann 'Myrkraverk' Oskarsson <johann@myrkraverk.invalid> wrote:
    On 11/15/2026 5:52 PM, The Natural Philosopher wrote:
    On 14/11/2026 18:34, Charlie Gibbs wrote:
    On 2026-11-14, The Natural Philosopher <tnp@invalid.invalid> wrote:

    On 14/11/2026 15:12, Johann 'Myrkraverk' Oskarsson wrote:

    I read somewhere that /mainframe/ got its name from the -- wait for it >>>>> -- main frame of the large vacuum tube computers of the 1950s, if not >>>>> 40s.-a I don't recall seeing pictures of these /main frames/, so I don't >>>>> know what they looked like.

    a computer built in multiple 19" racks?

    What about all those DEC minis built in 19" racks?

    Many mainframes were built in bespoke racks, often with
    backplanes on swing-out frames.-a Much more exclusive and sexy.

    Whee, this is becoming fun...

    Indeed. Originally it was what it was, but it became

    1. A computer able to handled thousands of user sessions
    or
    2.-a A computer with massive storage and the ability to do payroll
    batches in time for payday.

    Naturally these are very large organisation solutions, so not many
    people were exposed to them. Which is why so many people think these
    machines no longer exist



    The problem of /how do we get $amount into every employee's bank ac-
    count/ with thousands of employees and perhaps a single payroll account
    for the company does not sound like a solution waiting for a cluster of
    Linux boxes.

    The problem is actually how do we get the *right* amount into every
    employee's bank account (or charge the *right* amount of tax, or give
    the *right* amount of benefits...), where *right* is defined by years
    or decades of ad-hoc business logic and legislation that is embedded into ancient software (COBOL/JCL et. al.) that can only run on the "mainframe"
    it was developed on.

    Too often the shiny new solution doesn't get the right answers, despite
    being infinitely faster. So what do you do - keep paying your mainframe supplier and not have any risk, or switch to a modern solution and risk
    your job if thousands of people complain that you've got the sums wrong?

    Yes, eventually the modern solution wins, but IMO this is the primary
    reason "mainframes" still exist.

    So today, "mainframe" = "System capable of running software developed
    before Linux or Windows was ever conceived" :)
    --
    Ian

    "Tamahome!!!" - "Miaka!!!"
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Kerr-Mudd, John@admin@127.0.0.1 to comp.os.linux.misc,alt.folklore.computers on Sun Nov 15 12:21:48 2026
    From Newsgroup: comp.os.linux.misc

    On Sat, 14 Nov 2026 00:29:43 -0500
    c186282 <c186282@nnada.net> wrote:

    On 11/13/26 16:02, Alexander Schreiber wrote:
    The Natural Philosopher <tnp@invalid.invalid> wrote:
    On 13/11/2026 14:40, Peter Flass wrote:
    []

    I can't understand (well, I can...) why LiO2 batteries were ever
    approved for general use.

    There is no such battery. Lithium Ion or LIPO for lithium ion polymer or >> LiFePo for Lithium iron polymer

    Well, there are Lithium-Air batteries. But while various implementations
    of them have been demonstrated, AFAIK they are nowhere near commercialization.

    Just the other day I saw on the news that a
    house burned (actually exploded) because if a battery fire.

    Batteries made badly and sold out of spec.

    And/or badly abused during transport, installation, operation. Plus
    people with little knowledge of the technology tend to end up with
    the more exciting (read: fire-prone) lithium battery chemistries instead
    of with much safer ones e.g. LiFePo4. It doesn't help that often the
    cell chemistry is not explicitly mentioned, e.g. with battery packs
    for e-bikes and such.

    []
    The best alt these days are sodium batteries ... already
    commercial, you can buy ones meant to start up automobiles
    on Amazon. Not quite the power density, but getting close
    and don't explode.

    As for 'quality batteries', just no WAY to know what
    you're really getting these days.

    Whatever happened to Iron-Air batteries?

    Or glass ?
    https://en.wikipedia.org/wiki/Glass_battery

    https://en.wikipedia.org/wiki/Solid-state_battery

    Lots of "promise", when do we get them in the shops ^w online?
    --
    Bah, and indeed Humbug.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Johann 'Myrkraverk' Oskarsson@johann@myrkraverk.invalid to comp.os.linux.misc,alt.folklore.computers on Sun Nov 15 20:43:53 2026
    From Newsgroup: comp.os.linux.misc

    On 11/15/2026 8:22 PM, Ian wrote:
    On 2026-11-15, Johann 'Myrkraverk' Oskarsson <johann@myrkraverk.invalid> wrote:
    On 11/15/2026 5:52 PM, The Natural Philosopher wrote:
    On 14/11/2026 18:34, Charlie Gibbs wrote:
    On 2026-11-14, The Natural Philosopher <tnp@invalid.invalid> wrote:

    On 14/11/2026 15:12, Johann 'Myrkraverk' Oskarsson wrote:

    I read somewhere that /mainframe/ got its name from the -- wait for it >>>>>> -- main frame of the large vacuum tube computers of the 1950s, if not >>>>>> 40s.-a I don't recall seeing pictures of these /main frames/, so I don't >>>>>> know what they looked like.

    a computer built in multiple 19" racks?

    What about all those DEC minis built in 19" racks?

    Many mainframes were built in bespoke racks, often with
    backplanes on swing-out frames.-a Much more exclusive and sexy.

    Whee, this is becoming fun...

    Indeed. Originally it was what it was, but it became

    1. A computer able to handled thousands of user sessions
    or
    2.-a A computer with massive storage and the ability to do payroll
    batches in time for payday.

    Naturally these are very large organisation solutions, so not many
    people were exposed to them. Which is why so many people think these
    machines no longer exist



    The problem of /how do we get $amount into every employee's bank ac-
    count/ with thousands of employees and perhaps a single payroll account
    for the company does not sound like a solution waiting for a cluster of
    Linux boxes.

    The problem is actually how do we get the *right* amount into every employee's bank account (or charge the *right* amount of tax, or give
    the *right* amount of benefits...), where *right* is defined by years
    or decades of ad-hoc business logic and legislation that is embedded into ancient software (COBOL/JCL et. al.) that can only run on the "mainframe"
    it was developed on.

    Too often the shiny new solution doesn't get the right answers, despite
    being infinitely faster. So what do you do - keep paying your mainframe supplier and not have any risk, or switch to a modern solution and risk
    your job if thousands of people complain that you've got the sums wrong?

    Yes, eventually the modern solution wins, but IMO this is the primary
    reason "mainframes" still exist.

    So today, "mainframe" = "System capable of running software developed
    before Linux or Windows was ever conceived" :)



    Or you define an Oxide computer as a /mainframe/ and then you have /n/
    virtual machines running on it, be it Linux, Illumos, or Windows.

    No, I'm not sure I would define an Oxide computer as a mainframe, but
    then I don't really care about the terminology as it's most often a
    marketing effort.

    People working at Oxide can -- potentially -- get offended if someone
    labels their spiffy new computery systems as /mainframes/.

    And if the outside box is not important, a Talos II workstation might
    qualify. [Unavailable last time I looked at the website, but some
    machines must still exist.]

    Bottom line, do we care if we run the Linux on a mainframe?
    --
    Johann | email: invalid -> com | http://www.myrkraverk.com/blog/
    I'm not from the Internet, I just work there. | via Easynews.com https://bsky.app/profile/myrkraverk.bsky.social | for ( ;; ) _:;
    Federated at https://fed.brid.gy/bsky/myrkraverk.bsky.social
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Alexander Schreiber@als@usenet.thangorodrim.de to comp.os.linux.misc,alt.folklore.computers on Sun Nov 15 13:15:02 2026
    From Newsgroup: comp.os.linux.misc

    Lawrence DrCOOliveiro <ldo@nz.invalid> wrote:
    On Fri, 13 Nov 2026 15:05:00 +0100, Alexander Schreiber wrote:

    Lawrence DrCOOliveiro <ldo@nz.invalid> wrote:

    Perception doesnrCOt count. ThatrCOs reality. You canrCOt *pretend* to
    offer interactivity: you really have to have support for
    interactivity built into your architecture from the lowest levels
    up. You canrCOt build per-character response on top of block mode.

    You absolutely can and the interactive web proves it.

    The interactive web is not run on mainframe-style block-mode
    terminals.

    I did not claim so. But the basic concepts of:
    - smart local terminal
    - data transfer between terminal and server not keystroke-by-keystroke
    but in functional blocks of data
    underlies both, hence one could argue that they share underlying
    functional ideas, however massively they differ in the user experience.

    WebSocket connections are asynchronous full-duplex connections --
    thatrCOs not rCLRPCrCY by any stretch.

    Of course it is. RPCs come in various kinds, from the classic
    one-shot (establish connection, send request, wait for and read
    response, close connection) to streaming RPCs.

    rCLRPCrCY stands for rCLRemote Procedure CallrCY. ItrCOs emulating a procedure
    call over a network connection. You know how a procedure call works,
    donrCOt you? The caller transfers control to the callee, waits for the
    callee to do its thing and return a result, whereupon the caller
    resumes execution again.

    Thanks for the refresher, so local procedure calls still work the same
    as back then when I wrote code that involved the occasional function
    and/or procedure call (yes, there are languages where there is a
    difference)? That's a relief.

    But that's not how RPCs work. Most importantly, there is no transfer
    of control, the local system can continue executing code while writing to
    or reading from the remote system. In the most simple case:
    - local code connects to remote system
    - writes to it to invoke a defined endpoint there, optionally writing
    arguments to that endpoint
    - reads back the results
    - disconnects

    One can do that as a full straight run of blocking network I/O. More
    commonly, the network I/O is mixed with ongoing local execution and
    calls to select(2) or epoll(7)/kqueue(2) to not block the caller waiting
    for the callee or the network, plus one wants timeouts to avoid waiting forever.

    And nobody ever described block-mode mainframe terminals as running
    any kind of rCLRPCrCY either. The rCLRPCrCY concept was only invented after mainframes had been around for a few decades.

    Because the concept, as you pointed out, was not invented when block
    mode terminals where the new hot stuff. In retrospect, there are
    functional similarities between those block mode terminals and web
    services, which might be more obvious if one uses a terminal browser
    like links for the web (which of course limits one to sites that don't
    rely on JavaScript and lots of graphics for their function).

    Before getting even deeper into your love for RPCs and seeing every
    kind of network protocol as RPCs, maybe you should read this

    Professional bias. I work in an environment where every in-house binary
    is always both an RPC server and an RPC client - RPC traffic is everywhere.
    So one tends to recognize the pattern. Yes, proper designed-as-such RPC
    systems have defined endpoint naming structure (e.g. Service.Method), well defined serialization (such as XDR or protobufs) and proper builtin error handling and signaling mechanisms. That never stopped anyone from not doing things that way and e.g. just pushing raw bytes with no metadata (type, size, etc) over the ether (or worse, Python pickles) ...

    <https://www.cs.vu.nl/~ast/Publications/Papers/euteco-1988.pdf>.

    Interesting read, but it is painfully obvious from the text (even without peeking at the date) that this was written in the very early days of
    networked services (and the dig at the ISO model is also a nice tell).
    A lot has happened in computer science in the last almost 40y. In modern networked service infrastructure, the pain points the paper raises have
    pretty much all been dealt with in various ways. It does neatly demonstrate
    how painful and primitive (from todays perspective) networked service
    were at times back then, though.

    Even HTTP, which is not exactly the most efficient RPC method, came
    up with streaming RPCs: HTTP 1.1 pipelining, where multiple requests
    are handled within the same TCP connection.

    Those donrCOt work for RPCs, though.

    Running multiple RPC requests over HTTP 1.1 pipelining is going to be problematic at best and for obvious reasons, sure. But it does work when
    you consider it a primitive (half-duplex) streaming RPC with multiple sub-requests.

    Again: I see HTTP functionally as _a_ RPC system, which works reasonably
    well in the constrained functional space HTTP was designed for and has
    evolved in. Should you use it as a generic RPC system? Of course not, that would be silly (and I write that as someone who used ssh as primitive RPC system before).

    And "asynchronous" just means you don't do blocking reads and
    instead use the framework equivalent of select(2) for a more fluid
    interaction.

    Well, maybe not select(2), maybe something more modern?

    Sure, hence "functional equivalent", e.g. epoll or kqueue - or whatever
    higher level abstraction (such as Promises/Futures in Java) is offered
    by whatever language or framework one uses.

    Kind regards,
    Alex.
    --
    "Opportunity is missed by most people because it is dressed in overalls and
    looks like work." -- Thomas A. Edison
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From The Natural Philosopher@tnp@invalid.invalid to comp.os.linux.misc,alt.folklore.computers on Sun Nov 15 13:55:00 2026
    From Newsgroup: comp.os.linux.misc

    On 15/11/2026 12:22, Ian wrote:
    Yes, eventually the modern solution wins, but IMO this is the primary
    reason "mainframes" still exist.

    So today, "mainframe" = "System capable of running software developed
    before Linux or Windows was ever conceived" EfOe

    Simply not true

    All trad OS and languages run on linux machines of the standard sort -
    Fortran - COBOL - RPG...SQL.

    It's not software, it's hardware.
    Uber large storage, +4ber large user numbers.

    Not always a huge amount of CPU.
    --
    Climate Change: Socialism wearing a lab coat.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From The Natural Philosopher@tnp@invalid.invalid to comp.os.linux.misc,alt.folklore.computers on Sun Nov 15 13:56:34 2026
    From Newsgroup: comp.os.linux.misc

    On 15/11/2026 12:43, Johann 'Myrkraverk' Oskarsson wrote:
    Bottom line, do we care if we run the Linux on a mainframe?

    Nope. But it probably would need to be optimised for it
    Mainframes today tend to run just one or two applications and the
    operating system is whatever works with those
    --
    "It was a lot more fun being 20 in the 70's that it is being 70 in the 20's" Joew Walsh

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From The Natural Philosopher@tnp@invalid.invalid to comp.os.linux.misc,alt.folklore.computers on Sun Nov 15 14:01:23 2026
    From Newsgroup: comp.os.linux.misc

    On 15/11/2026 12:15, Alexander Schreiber wrote:
    In retrospect, there are
    functional similarities between those block mode terminals and web
    services, which might be more obvious if one uses a terminal browser
    like links for the web (which of course limits one to sites that don't
    rely on JavaScript and lots of graphics for their function).

    Indeed. the web browser is a thick client and so is a 3270 terminal.,

    Whether something is an RPC or an AJAX call is really semantics.
    You ask or send, you wait, you receive.

    But not a whole screens worth of data. So transactions are lighter and smaller

    My AJAX calls generally send/return less than 1k of data
    --
    "It was a lot more fun being 20 in the 70's that it is being 70 in the 20's" Joew Walsh

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Thomas Koenig@tkoenig@netcologne.de to comp.os.linux.misc,alt.folklore.computers on Sun Nov 15 17:07:29 2026
    From Newsgroup: comp.os.linux.misc

    Johann 'Myrkraverk' Oskarsson <johann@myrkraverk.invalid> schrieb:

    The problem of /how do we get $amount into every employee's bank ac-
    count/ with thousands of employees and perhaps a single payroll account
    for the company does not sound like a solution waiting for a cluster of
    Linux boxes.

    SAP moved away from mainframes in the early 1990s with SAP R3, and
    they do run payrolls.

    And SAP S4/HANA runs on... you guessed it, Linux, exclusively.
    It does not even run on mainframes any more - it's Intel (not AMD)
    or IBM POWER.
    --
    This USENET posting was made without artificial intelligence,
    artificial impertinence, artificial arrogance, artificial stupidity,
    artificial flavorings or artificial colorants.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Ian@${send-direct-email-to-news1021-at-jusme-dot-com-if-you-must}@jusme.com to comp.os.linux.misc,alt.folklore.computers on Sun Nov 15 17:41:52 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-15, The Natural Philosopher <tnp@invalid.invalid> wrote:
    On 15/11/2026 12:22, Ian wrote:
    Yes, eventually the modern solution wins, but IMO this is the primary
    reason "mainframes" still exist.

    So today, "mainframe" = "System capable of running software developed
    before Linux or Windows was ever conceived" :)

    Simply not true

    All trad OS and languages run on linux machines of the standard sort - Fortran - COBOL - RPG...SQL.

    It simply is true, and it pays my wages. Yes, of course COBOL et. el. are available on modern OS's, but not the dialects in which all that business
    logic is encoded. And the business logic is a lot more than a single
    COBOL program, it's literally hundreds of them, all knitted together in
    strange and interesting ways by proprietary job control language scripting,
    and all the data is in legacy proprietary databases, not nice modern
    relational ones. Unravelling that is a huge effort, and organisations have spent millions trying and failing to do it, so they're still paying for
    the mainframe OS (which is now running on Linux, of course!)

    It won't last forever, but it's still doing the job and paying the bills, literally.
    --
    Ian

    "Tamahome!!!" - "Miaka!!!"
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lawrence =?iso-8859-13?q?D=FFOliveiro?=@ldo@nz.invalid to comp.os.linux.misc,alt.folklore.computers on Sun Nov 15 18:39:26 2026
    From Newsgroup: comp.os.linux.misc

    On Sun, 15 Nov 2026 12:22:30 -0000 (UTC), Ian wrote:

    The problem is actually how do we get the *right* amount into every employee's bank account (or charge the *right* amount of tax, or
    give the *right* amount of benefits...), where *right* is defined by
    years or decades of ad-hoc business logic and legislation that is
    embedded into ancient software (COBOL/JCL et. al.) that can only run
    on the "mainframe" it was developed on.

    Too often the shiny new solution doesn't get the right answers,
    despite being infinitely faster. So what do you do - keep paying
    your mainframe supplier and not have any risk, or switch to a modern
    solution and risk your job if thousands of people complain that
    you've got the sums wrong?

    Salaries donrCOt stand still; benefits and deductions donrCOt stand still; taxes donrCOt stand still. Keep running old code, and you find it harder
    and harder to keep it up to date with all the new laws and
    regulations.

    Remember BrooksrCO rCLten lines per dayrCY rule? 10 lines of COBOL code doesnrCOt do much; if your programmers can only produce that much per
    day, can they keep up with all the changes that need to be made?

    So, one way or another, the old code will remove itself from
    circulation. The only option the company has is to go down with it or
    not.

    Yes, eventually the modern solution wins, but IMO this is the primary
    reason "mainframes" still exist.

    So today, "mainframe" = "System capable of running software developed
    before Linux or Windows was ever conceived" :)

    Given that a lot of IBMrCOs rCLmainframesrCY are running Linux now, I
    suspect even they are now just using it as a marketing term ...
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Johann 'Myrkraverk' Oskarsson@johann@myrkraverk.invalid to comp.os.linux.misc,alt.folklore.computers on Mon Nov 16 02:44:00 2026
    From Newsgroup: comp.os.linux.misc

    On 11/16/2026 1:07 AM, Thomas Koenig wrote:
    Johann 'Myrkraverk' Oskarsson <johann@myrkraverk.invalid> schrieb:

    The problem of /how do we get $amount into every employee's bank ac-
    count/ with thousands of employees and perhaps a single payroll account
    for the company does not sound like a solution waiting for a cluster of
    Linux boxes.

    SAP moved away from mainframes in the early 1990s with SAP R3, and
    they do run payrolls.

    And SAP S4/HANA runs on... you guessed it, Linux, exclusively.
    It does not even run on mainframes any more - it's Intel (not AMD)
    or IBM POWER.


    It's not exactly clear from https://en.wikipedia.org/wiki/SAP_R/3 but
    wouldn't the payroll system still run on a single machine, with perhaps
    the database on another, and different parts of the application on even
    more nodes? Would that machine qualify as a /mainframe/?
    --
    Johann | email: invalid -> com | http://www.myrkraverk.com/blog/
    I'm not from the Internet, I just work there. | via Easynews.com https://bsky.app/profile/myrkraverk.bsky.social | for ( ;; ) _:;
    Federated at https://fed.brid.gy/bsky/myrkraverk.bsky.social
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From rbowman@bowman@montana.com to comp.os.linux.misc,alt.folklore.computers on Sun Nov 15 20:01:23 2026
    From Newsgroup: comp.os.linux.misc

    On Sun, 15 Nov 2026 18:04:33 +0800, Johann 'Myrkraverk' Oskarsson wrote:

    JSON is probably the most popular today. I personally don't like it for being wasteful.

    I think that record goes to XML. After 9/11 it became painfully apparent
    that the various agencies couldn't communicate with each other. The DOJ
    set out to fix that and XML was the shiny new technology.

    https://en.wikipedia.org/wiki/NIEMOpen

    Prior to the fix a return from a NCIC query was a reasonably well
    formatted plain text document you could dump to a screen as is. It took a decade before the XML schema was completely adopted. Now the return is a
    huge XML document that requires XSLT to extract any human readable
    information from. Progress.

    It doesn't help that the various schemae change. A tag like <Id/> can be almost anything depending on its ancestry. The clients, of course, wanted
    the return to be formatted exactly like the plain text returns they had
    been getting for years. It kept me in a job, I guess.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Rich@rich@example.invalid to comp.os.linux.misc on Sun Nov 15 20:19:19 2026
    From Newsgroup: comp.os.linux.misc

    In comp.os.linux.misc Charlie Gibbs <cgibbs@kltpzyxm.invalid> wrote:
    On 2026-11-11, The Natural Philosopher <tnp@invalid.invalid> wrote:

    With rising fuel costs I somehow doubt supersonic airliners would ever
    be profitable again

    Besides, no matter how fast the plane flies you're still
    going to be waiting for your baggage at the other end.

    Actual air time is only part of the total travel time -
    once you add in the trip to the airport, check-in time,
    baggage retrieval, and transport from the destination
    terminal to wherever you're really going, air time becomes
    a relatively small portion of the total time.

    And, for USA travelers, TSA security theater line wait times, which can
    be substantial sometimes.

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From The Natural Philosopher@tnp@invalid.invalid to comp.os.linux.misc on Sun Nov 15 20:23:03 2026
    From Newsgroup: comp.os.linux.misc

    On 15/11/2026 20:19, Rich wrote:
    In comp.os.linux.misc Charlie Gibbs <cgibbs@kltpzyxm.invalid> wrote:
    On 2026-11-11, The Natural Philosopher <tnp@invalid.invalid> wrote:

    With rising fuel costs I somehow doubt supersonic airliners would ever
    be profitable again

    Besides, no matter how fast the plane flies you're still
    going to be waiting for your baggage at the other end.

    Actual air time is only part of the total travel time -
    once you add in the trip to the airport, check-in time,
    baggage retrieval, and transport from the destination
    terminal to wherever you're really going, air time becomes
    a relatively small portion of the total time.

    And, for USA travelers, TSA security theater line wait times, which can
    be substantial sometimes.

    Indeed. the fact is sadly that the bearded bastards like to destroy shit
    and planes are very handy and very vulnerable
    --
    Renewable energy: Expensive solutions that don't work to a problem that doesn't exist instituted by self legalising protection rackets that
    don't protect, masquerading as public servants who don't serve the public.


    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From rbowman@bowman@montana.com to comp.os.linux.misc,alt.folklore.computers on Sun Nov 15 20:57:11 2026
    From Newsgroup: comp.os.linux.misc

    On Sun, 15 Nov 2026 10:48:39 +0000, The Natural Philosopher wrote:

    On 15/11/2026 10:04, Johann 'Myrkraverk' Oskarsson wrote:
    JSON is probably the most popular today.-a I personally don't like it
    for being wasteful.

    There are enough bugs in JavaScript without introducing yet another
    layer...
    I use the minimum pure JavaScript sufficient to get the job done

    JSON wa derived from JavaScript but is widely used outside of JavaScript.
    My fitness tracker exports mmy routes in GPX format, which was developed
    when XML was at the height of its popularity.

    https://en.wikipedia.org/wiki/GPS_Exchange_Format

    In order to display the information on a map I need GeoJSON:

    https://en.wikipedia.org/wiki/GeoJSON

    I do the conversion with a C program, no JavaScript involved. I could have done it in Python using the gpxpy module to parse the gpx files rather
    than the expat library in C. Still no JS.

    MongoDB uses BSON internally but that is a binary representation of JSON,
    sort of like a protobuf.

    I also use JSON files with C# for configuration. They are pretty much self explanatory. That's good since JSON doesn't support comments. Another
    slight annoyance is while comma separated lists are used, the last item
    cannot be followed by a comma. That means you have to keep track of the
    last item.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Alexander Schreiber@als@usenet.thangorodrim.de to comp.os.linux.misc,alt.folklore.computers on Sun Nov 15 22:02:41 2026
    From Newsgroup: comp.os.linux.misc

    Ian <${send-direct-email-to-news1021-at-jusme-dot-com-if-you-must}@jusme.com> wrote:
    On 2026-11-15, Johann 'Myrkraverk' Oskarsson <johann@myrkraverk.invalid> wrote:
    On 11/15/2026 5:52 PM, The Natural Philosopher wrote:
    On 14/11/2026 18:34, Charlie Gibbs wrote:
    On 2026-11-14, The Natural Philosopher <tnp@invalid.invalid> wrote:

    On 14/11/2026 15:12, Johann 'Myrkraverk' Oskarsson wrote:

    I read somewhere that /mainframe/ got its name from the -- wait for it >>>>>> -- main frame of the large vacuum tube computers of the 1950s, if not >>>>>> 40s.-a I don't recall seeing pictures of these /main frames/, so I don't >>>>>> know what they looked like.

    a computer built in multiple 19" racks?

    What about all those DEC minis built in 19" racks?

    Many mainframes were built in bespoke racks, often with
    backplanes on swing-out frames.-a Much more exclusive and sexy.

    Whee, this is becoming fun...

    Indeed. Originally it was what it was, but it became

    1. A computer able to handled thousands of user sessions
    or
    2.-a A computer with massive storage and the ability to do payroll
    batches in time for payday.

    Naturally these are very large organisation solutions, so not many
    people were exposed to them. Which is why so many people think these
    machines no longer exist



    The problem of /how do we get $amount into every employee's bank ac-
    count/ with thousands of employees and perhaps a single payroll account
    for the company does not sound like a solution waiting for a cluster of
    Linux boxes.

    The problem is actually how do we get the *right* amount into every employee's bank account (or charge the *right* amount of tax, or give
    the *right* amount of benefits...), where *right* is defined by years
    or decades of ad-hoc business logic and legislation that is embedded into ancient software (COBOL/JCL et. al.) that can only run on the "mainframe"
    it was developed on.

    And that! It's not that an amd64 box running Linux doesn't have the power
    to run payroll, _that_ part is easy. The real difficulty is having the financial rules engine and continuously updated rules (because e.g.
    governments and insurance companies love to change those) for things like taxes, benefits, insurance contributions and so many more details.

    If you have a working system on a mainframe that gets these correct,
    the difficulty in porting this to e.g. Linux on amd64 is to make sure
    the new implementation is _also_ correct, including for "this rule only
    applies once a year to specific employees under specific circumstances"
    and other "fun" like that. And since that typically also tends to
    involve a change of implementation language (e.g. COBOL to Java) there
    are lots of opportunities to mess things up. Which will not be appreciated
    by those whose pay gets messed up.

    Too often the shiny new solution doesn't get the right answers, despite
    being infinitely faster. So what do you do - keep paying your mainframe supplier and not have any risk, or switch to a modern solution and risk
    your job if thousands of people complain that you've got the sums wrong?

    Indeed. People may be annoyed if Facebook doesn't load, but mess with
    their pay and they will be (understandably) ANGRY. "Move fast and break
    things" is not the way to do this ;-)

    Yes, eventually the modern solution wins, but IMO this is the primary
    reason "mainframes" still exist.

    I suspect so as well, yes.

    So today, "mainframe" = "System capable of running software developed
    before Linux or Windows was ever conceived" :)

    Well, I guess that is one definition ;-)

    Kind regards,
    Alex.
    --
    "Opportunity is missed by most people because it is dressed in overalls and
    looks like work." -- Thomas A. Edison
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Alexander Schreiber@als@usenet.thangorodrim.de to comp.os.linux.misc,alt.folklore.computers on Sun Nov 15 21:27:49 2026
    From Newsgroup: comp.os.linux.misc

    Johann 'Myrkraverk' Oskarsson <johann@myrkraverk.invalid> wrote:
    On 11/15/2026 5:52 PM, The Natural Philosopher wrote:
    On 14/11/2026 18:34, Charlie Gibbs wrote:
    On 2026-11-14, The Natural Philosopher <tnp@invalid.invalid> wrote:

    On 14/11/2026 15:12, Johann 'Myrkraverk' Oskarsson wrote:

    I read somewhere that /mainframe/ got its name from the -- wait for it >>>>> -- main frame of the large vacuum tube computers of the 1950s, if not >>>>> 40s.-a I don't recall seeing pictures of these /main frames/, so I don't >>>>> know what they looked like.

    a computer built in multiple 19" racks?

    What about all those DEC minis built in 19" racks?

    Many mainframes were built in bespoke racks, often with
    backplanes on swing-out frames.-a Much more exclusive and sexy.

    Whee, this is becoming fun...

    Indeed. Originally it was what it was, but it became

    1. A computer able to handled thousands of user sessions
    or
    2.-a A computer with massive storage and the ability to do payroll
    batches in time for payday.

    Naturally these are very large organisation solutions, so not many
    people were exposed to them. Which is why so many people think these
    machines no longer exist



    The problem of /how do we get $amount into every employee's bank ac-
    count/ with thousands of employees and perhaps a single payroll account
    for the company does not sound like a solution waiting for a cluster of
    Linux boxes.

    The actual payroll processing itself is a small problem on modern
    machines, you could probably stuff that in a small corner of a modern
    amd64 Linux server. Unless, of course, you use SAP for that, in which
    case you need at least a rack full of the largest amd64 machines available
    on the market ...

    Far more critical is the integration with the bank to ensure the payments
    get sent out correctly and on time.

    Kind regards,
    Alex.
    --
    "Opportunity is missed by most people because it is dressed in overalls and
    looks like work." -- Thomas A. Edison
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Alexander Schreiber@als@usenet.thangorodrim.de to comp.os.linux.misc,alt.folklore.computers on Sun Nov 15 21:46:54 2026
    From Newsgroup: comp.os.linux.misc

    The Natural Philosopher <tnp@invalid.invalid> wrote:
    On 15/11/2026 10:33, Johann 'Myrkraverk' Oskarsson wrote:
    On 11/15/2026 5:52 PM, The Natural Philosopher wrote:
    On 14/11/2026 18:34, Charlie Gibbs wrote:
    On 2026-11-14, The Natural Philosopher <tnp@invalid.invalid> wrote:

    On 14/11/2026 15:12, Johann 'Myrkraverk' Oskarsson wrote:

    I read somewhere that /mainframe/ got its name from the -- wait for it >>>>>> -- main frame of the large vacuum tube computers of the 1950s, if not >>>>>> 40s.-a I don't recall seeing pictures of these /main frames/, so I >>>>>> don't
    know what they looked like.

    a computer built in multiple 19" racks?

    What about all those DEC minis built in 19" racks?

    Many mainframes were built in bespoke racks, often with
    backplanes on swing-out frames.-a Much more exclusive and sexy.

    Whee, this is becoming fun...

    Indeed. Originally it was what it was, but it became

    1. A computer able to handled thousands of user sessions
    or
    2.-a A computer with massive storage and the ability to do payroll
    batches in time for payday.

    Naturally these are very large organisation solutions, so not many
    people were exposed to them. Which is why so many people think these
    machines no longer exist



    The problem of /how do we get $amount into every employee's bank ac-
    count/ with thousands of employees and perhaps a single payroll account
    for the company does not sound like a solution waiting for a cluster of
    Linux boxes.

    That was essentially my point.

    Neither is necessarily running a global airline booking application.

    Do keep up with the industry, please. E.g. Amadeus - in their own words "WerCOve freed ourselves from legacy mainframe systems and have adapted and evolved our existing IT framework to todayrCOs cloud environment."[0]

    We have a range of tools at our disposal, from a Raspberry Pi Pico to an
    IBM mainframe, and none of them are suitable for *all* applications.

    So you pick the machines that fit the job, of course.

    In the middle, a linux *86 style machine is a very useful and cost
    effective machine, but not at the spectrum edges.

    That "middle" these days is reaching up all the way to 16 sockets (so easily thousands of cores depending one the CPUs used) and 32 TiB of RAM.

    E.g. one of the certified platforms for SAP HANA[1] is the Google Cloud x4-1920-32t-metal VM instance: 32 TiB RAM, 1920 vCPUs, running various
    flavors of e.g. Redhat or SLES Linux.

    But then, SAP has always been an epic resource hog and SAP HANA even
    more so.

    And the usual supercomputers run variants of Linux these days too,
    e.g. the top two: Lineshine (China) and El Capitan (USA) both run
    custom Linux distributions.

    Kind regards,
    Alex.
    [0] https://amadeus.com/en/blog/articles/five-things-to-know-amadeus-travel-platform
    [1] https://www.sap.com/dmc/exp/2014-09-02-hana-hardware/enEN/
    --
    "Opportunity is missed by most people because it is dressed in overalls and
    looks like work." -- Thomas A. Edison
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Johann 'Myrkraverk' Oskarsson@johann@myrkraverk.invalid to comp.os.linux.misc,alt.folklore.computers on Mon Nov 16 05:33:38 2026
    From Newsgroup: comp.os.linux.misc

    On 11/16/2026 4:46 AM, Alexander Schreiber wrote:
    The Natural Philosopher <tnp@invalid.invalid> wrote:
    On 15/11/2026 10:33, Johann 'Myrkraverk' Oskarsson wrote:
    On 11/15/2026 5:52 PM, The Natural Philosopher wrote:
    On 14/11/2026 18:34, Charlie Gibbs wrote:
    On 2026-11-14, The Natural Philosopher <tnp@invalid.invalid> wrote:

    On 14/11/2026 15:12, Johann 'Myrkraverk' Oskarsson wrote:

    I read somewhere that /mainframe/ got its name from the -- wait for it >>>>>>> -- main frame of the large vacuum tube computers of the 1950s, if not >>>>>>> 40s.-a I don't recall seeing pictures of these /main frames/, so I >>>>>>> don't
    know what they looked like.

    a computer built in multiple 19" racks?

    What about all those DEC minis built in 19" racks?

    Many mainframes were built in bespoke racks, often with
    backplanes on swing-out frames.-a Much more exclusive and sexy.

    Whee, this is becoming fun...

    Indeed. Originally it was what it was, but it became

    1. A computer able to handled thousands of user sessions
    or
    2.-a A computer with massive storage and the ability to do payroll
    batches in time for payday.

    Naturally these are very large organisation solutions, so not many
    people were exposed to them. Which is why so many people think these
    machines no longer exist



    The problem of /how do we get $amount into every employee's bank ac-
    count/ with thousands of employees and perhaps a single payroll account
    for the company does not sound like a solution waiting for a cluster of
    Linux boxes.

    That was essentially my point.

    Neither is necessarily running a global airline booking application.

    Do keep up with the industry, please. E.g. Amadeus - in their own words "WerCOve freed ourselves from legacy mainframe systems and have adapted and evolved our existing IT framework to todayrCOs cloud environment."[0]

    We have a range of tools at our disposal, from a Raspberry Pi Pico to an
    IBM mainframe, and none of them are suitable for *all* applications.

    So you pick the machines that fit the job, of course.

    In the middle, a linux *86 style machine is a very useful and cost
    effective machine, but not at the spectrum edges.

    That "middle" these days is reaching up all the way to 16 sockets (so easily thousands of cores depending one the CPUs used) and 32 TiB of RAM.

    E.g. one of the certified platforms for SAP HANA[1] is the Google Cloud x4-1920-32t-metal VM instance: 32 TiB RAM, 1920 vCPUs, running various flavors of e.g. Redhat or SLES Linux.

    But then, SAP has always been an epic resource hog and SAP HANA even
    more so.

    And the usual supercomputers run variants of Linux these days too,
    e.g. the top two: Lineshine (China) and El Capitan (USA) both run
    custom Linux distributions.

    Kind regards,
    Alex.
    [0] https://amadeus.com/en/blog/articles/five-things-to-know-amadeus-travel-platform
    [1] https://www.sap.com/dmc/exp/2014-09-02-hana-hardware/enEN/

    I'm not sure that's too relevant to the discussion at hand. I mean,
    we're casually thinking what /mainframe/ means in this day and age,
    plus the past, and just because a machine is running the x86 instruction
    set doesn't disqualify it of the label.

    I know a lot of organizations are running "on premise" cloud infrastruc-
    ture, such as with the Oxide [1] machines I mentioned previously. And
    then using that machine to turn off and on various virtual machines to
    run the workload as needed.

    Unless you want /mainframe/ to mean an IBM computer?

    [1] Not that I care too much about Oxide, but theirs is the only com-
    puter system I know which you can just buy and run a local cloud on.
    --
    Johann | email: invalid -> com | http://www.myrkraverk.com/blog/
    I'm not from the Internet, I just work there. | via Easynews.com https://bsky.app/profile/myrkraverk.bsky.social | for ( ;; ) _:;
    Federated at https://fed.brid.gy/bsky/myrkraverk.bsky.social
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Alexander Schreiber@als@usenet.thangorodrim.de to comp.os.linux.misc,alt.folklore.computers on Sun Nov 15 22:47:06 2026
    From Newsgroup: comp.os.linux.misc

    Ian <${send-direct-email-to-news1021-at-jusme-dot-com-if-you-must}@jusme.com> wrote:
    On 2026-11-15, The Natural Philosopher <tnp@invalid.invalid> wrote:
    On 15/11/2026 12:22, Ian wrote:
    Yes, eventually the modern solution wins, but IMO this is the primary
    reason "mainframes" still exist.

    So today, "mainframe" = "System capable of running software developed
    before Linux or Windows was ever conceived" :)

    Simply not true

    All trad OS and languages run on linux machines of the standard sort -
    Fortran - COBOL - RPG...SQL.

    It simply is true, and it pays my wages. Yes, of course COBOL et. el. are available on modern OS's, but not the dialects in which all that business logic is encoded. And the business logic is a lot more than a single
    COBOL program, it's literally hundreds of them, all knitted together in strange and interesting ways by proprietary job control language scripting, and all the data is in legacy proprietary databases, not nice modern relational ones. Unravelling that is a huge effort, and organisations have spent millions trying and failing to do it, so they're still paying for
    the mainframe OS (which is now running on Linux, of course!)

    Porting a codebase that:
    - has grown over decades (as a lof of these "business logic" systems on
    mainframes did)
    - represents complex business logic (because legal frameworks,
    contractual agreements etc - all which was/is being added to and
    changed over time) in code and config
    - has been running with a high level of correctness for many years
    - probably isn't exactly documented in every detail (including _why_
    certain design decisions where made 5, 10, 25y ago) that you really
    would like to know and understand for your porting effort

    is going to be a massive effort, especially if you change the implementation language (as these efforts usually do) as well. Getting this right pretty
    much inevitably costs a lot more time, money and effort than anyone initially predicts. And then you still might find gremlins in various corner cases, especially if they are rarely exercised (but often critical).

    It won't last forever, but it's still doing the job and paying the bills, literally.

    Indeed.

    Kind regards,
    Alex.
    --
    "Opportunity is missed by most people because it is dressed in overalls and
    looks like work." -- Thomas A. Edison
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Alexander Schreiber@als@usenet.thangorodrim.de to comp.os.linux.misc,alt.folklore.computers on Sun Nov 15 22:55:54 2026
    From Newsgroup: comp.os.linux.misc

    Kerr-Mudd, John <admin@127.0.0.1> wrote:
    On Sat, 14 Nov 2026 00:29:43 -0500
    c186282 <c186282@nnada.net> wrote:

    On 11/13/26 16:02, Alexander Schreiber wrote:
    The Natural Philosopher <tnp@invalid.invalid> wrote:
    On 13/11/2026 14:40, Peter Flass wrote:
    []

    I can't understand (well, I can...) why LiO2 batteries were ever
    approved for general use.

    There is no such battery. Lithium Ion or LIPO for lithium ion polymer or >> >> LiFePo for Lithium iron polymer

    Well, there are Lithium-Air batteries. But while various implementations >> > of them have been demonstrated, AFAIK they are nowhere near
    commercialization.

    Just the other day I saw on the news that a
    house burned (actually exploded) because if a battery fire.

    Batteries made badly and sold out of spec.

    And/or badly abused during transport, installation, operation. Plus
    people with little knowledge of the technology tend to end up with
    the more exciting (read: fire-prone) lithium battery chemistries instead >> > of with much safer ones e.g. LiFePo4. It doesn't help that often the
    cell chemistry is not explicitly mentioned, e.g. with battery packs
    for e-bikes and such.

    []
    The best alt these days are sodium batteries ... already
    commercial, you can buy ones meant to start up automobiles
    on Amazon. Not quite the power density, but getting close
    and don't explode.

    As for 'quality batteries', just no WAY to know what
    you're really getting these days.

    Whatever happened to Iron-Air batteries?

    Being deployed for grid-scale applications (think plot of land full
    of shipping container sized battery modules). Right now they seem to
    mostly target this market (which is a lot less sensitive to battery
    capacity per weight or volume).

    https://en.wikipedia.org/wiki/Glass_battery

    Ongoing research.

    https://en.wikipedia.org/wiki/Solid-state_battery

    Lots of research, some early applications.

    Lots of "promise", when do we get them in the shops ^w online?

    Some of those are going to ... need a much larger pocket to put them
    into. ;-)

    Kind regards,
    Alex.
    --
    "Opportunity is missed by most people because it is dressed in overalls and
    looks like work." -- Thomas A. Edison
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Alexander Schreiber@als@usenet.thangorodrim.de to comp.os.linux.misc,alt.folklore.computers on Sun Nov 15 22:40:31 2026
    From Newsgroup: comp.os.linux.misc

    The Natural Philosopher <tnp@invalid.invalid> wrote:
    On 15/11/2026 12:43, Johann 'Myrkraverk' Oskarsson wrote:
    Bottom line, do we care if we run the Linux on a mainframe?

    Nope. But it probably would need to be optimised for it
    Mainframes today tend to run just one or two applications and the
    operating system is whatever works with those

    Linux has been running on IBM "son of S/390" mainframes for many years
    now. Not on the bare metal[1], because that would require a lot of duplicate effort for h/w support at the very least, but inside suitable virtual
    machines. There is an IBM Redbook[0] about Linux on S/390 from 2000.

    There was even a time when IBM actively promoted their mainframes as
    HA hosts for running lots of Linux VMs.

    Kind regards,
    Alex.
    [0] https://www.redbooks.ibm.com/redbooks/pdfs/sg244987.pdf
    [1] Well, also in LPARs which apparently is as close to the "bare metal"
    as an OS can get on these machines.
    --
    "Opportunity is missed by most people because it is dressed in overalls and
    looks like work." -- Thomas A. Edison
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Nuno Silva@nunojsilva@invalid.invalid to comp.os.linux.misc,alt.folklore.computers on Sun Nov 15 23:45:16 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-15, Lawrence DrCOOliveiro wrote:

    On Sun, 15 Nov 2026 12:22:30 -0000 (UTC), Ian wrote:

    The problem is actually how do we get the *right* amount into every
    employee's bank account (or charge the *right* amount of tax, or
    give the *right* amount of benefits...), where *right* is defined by
    years or decades of ad-hoc business logic and legislation that is
    embedded into ancient software (COBOL/JCL et. al.) that can only run
    on the "mainframe" it was developed on.

    Too often the shiny new solution doesn't get the right answers,
    despite being infinitely faster. So what do you do - keep paying
    your mainframe supplier and not have any risk, or switch to a modern
    solution and risk your job if thousands of people complain that
    you've got the sums wrong?

    Salaries donrCOt stand still; benefits and deductions donrCOt stand still; taxes donrCOt stand still. Keep running old code, and you find it harder
    and harder to keep it up to date with all the new laws and
    regulations.

    The very same argument applies to everything but the new laws and
    regulations, old code at least has these covered. Extending that code is
    the least error-prone way of the two?
    --
    Nuno Silva
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Johann 'Myrkraverk' Oskarsson@johann@myrkraverk.invalid to comp.os.linux.misc,alt.folklore.computers on Mon Nov 16 07:47:03 2026
    From Newsgroup: comp.os.linux.misc

    On 11/16/2026 4:01 AM, rbowman wrote:
    On Sun, 15 Nov 2026 18:04:33 +0800, Johann 'Myrkraverk' Oskarsson wrote:

    JSON is probably the most popular today. I personally don't like it for
    being wasteful.

    I think that record goes to XML. After 9/11 it became painfully apparent
    that the various agencies couldn't communicate with each other. The DOJ
    set out to fix that and XML was the shiny new technology.

    https://en.wikipedia.org/wiki/NIEMOpen

    Prior to the fix a return from a NCIC query was a reasonably well
    formatted plain text document you could dump to a screen as is. It took a decade before the XML schema was completely adopted. Now the return is a
    huge XML document that requires XSLT to extract any human readable information from. Progress.

    It doesn't help that the various schemae change. A tag like <Id/> can be almost anything depending on its ancestry. The clients, of course, wanted
    the return to be formatted exactly like the plain text returns they had
    been getting for years. It kept me in a job, I guess.

    I don't know how things are decided at the government scale, but most of
    the time when XML, JSON, etc., are decided upon, it's the developer who
    is deciding on a /library/ he's familiar with, rather than the underly-
    ing technology. Which means a lot of /shit decisions/ are made for peo-
    ple who need to interact with the resulting mess.

    JSON can be just as complicated. There's no rule that an "id": in a sub
    object is the same as the "id": in the parent object. The one thing XML
    has above JSON is the ability to -- in theory -- unambiguously embed one
    XML document in another.

    Most people don't need that, and can safely use JSON or similar instead.

    You can even use Lua for that matter; its syntax isn't that much differ-
    ent from JSON -- well, a passive subset I mean.


    Best wishes, and happy textual transaction messages!
    --
    Johann | email: invalid -> com | http://www.myrkraverk.com/blog/
    I'm not from the Internet, I just work there. | via Easynews.com https://bsky.app/profile/myrkraverk.bsky.social | for ( ;; ) _:;
    Federated at https://fed.brid.gy/bsky/myrkraverk.bsky.social
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Lawrence =?iso-8859-13?q?D=FFOliveiro?=@ldo@nz.invalid to comp.os.linux.misc,alt.folklore.computers on Mon Nov 16 01:20:03 2026
    From Newsgroup: comp.os.linux.misc

    On Sun, 15 Nov 2026 22:40:31 +0100, Alexander Schreiber wrote:

    There was even a time when IBM actively promoted their mainframes as
    HA hosts for running lots of Linux VMs.

    Why not any more?
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From rbowman@bowman@montana.com to comp.os.linux.misc on Mon Nov 16 03:12:20 2026
    From Newsgroup: comp.os.linux.misc

    On Sun, 15 Nov 2026 20:23:03 +0000, The Natural Philosopher wrote:

    On 15/11/2026 20:19, Rich wrote:
    In comp.os.linux.misc Charlie Gibbs <cgibbs@kltpzyxm.invalid> wrote:
    On 2026-11-11, The Natural Philosopher <tnp@invalid.invalid> wrote:

    With rising fuel costs I somehow doubt supersonic airliners would
    ever be profitable again

    Besides, no matter how fast the plane flies you're still going to be
    waiting for your baggage at the other end.

    Actual air time is only part of the total travel time -
    once you add in the trip to the airport, check-in time,
    baggage retrieval, and transport from the destination terminal to
    wherever you're really going, air time becomes a relatively small
    portion of the total time.

    And, for USA travelers, TSA security theater line wait times, which can
    be substantial sometimes.

    Indeed. the fact is sadly that the bearded bastards like to destroy shit
    and planes are very handy and very vulnerable

    Right.. My cousin, who was in her 80s and brought to the boarding gate in
    a wheel chair since she can't walk far, was hassled by TSA who confiscated
    her jar of hand cream since it was larger than the 0.3 fl oz allowed.
    Typical terrorist profile.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From c186282@c186282@nnada.net to comp.os.linux.misc,alt.folklore.computers on Sun Nov 15 22:28:29 2026
    From Newsgroup: comp.os.linux.misc

    On 11/15/26 07:21, Kerr-Mudd, John wrote:
    On Sat, 14 Nov 2026 00:29:43 -0500
    c186282 <c186282@nnada.net> wrote:

    On 11/13/26 16:02, Alexander Schreiber wrote:
    The Natural Philosopher <tnp@invalid.invalid> wrote:
    On 13/11/2026 14:40, Peter Flass wrote:
    []

    I can't understand (well, I can...) why LiO2 batteries were ever
    approved for general use.

    There is no such battery. Lithium Ion or LIPO for lithium ion polymer or >>>> LiFePo for Lithium iron polymer

    Well, there are Lithium-Air batteries. But while various implementations >>> of them have been demonstrated, AFAIK they are nowhere near
    commercialization.

    Just the other day I saw on the news that a
    house burned (actually exploded) because if a battery fire.

    Batteries made badly and sold out of spec.

    And/or badly abused during transport, installation, operation. Plus
    people with little knowledge of the technology tend to end up with
    the more exciting (read: fire-prone) lithium battery chemistries instead >>> of with much safer ones e.g. LiFePo4. It doesn't help that often the
    cell chemistry is not explicitly mentioned, e.g. with battery packs
    for e-bikes and such.

    []
    The best alt these days are sodium batteries ... already
    commercial, you can buy ones meant to start up automobiles
    on Amazon. Not quite the power density, but getting close
    and don't explode.

    As for 'quality batteries', just no WAY to know what
    you're really getting these days.

    Whatever happened to Iron-Air batteries?

    Or glass ?
    https://en.wikipedia.org/wiki/Glass_battery

    https://en.wikipedia.org/wiki/Solid-state_battery

    Lots of "promise", when do we get them in the shops ^w online?

    Note some of the technologies CAN provide power
    for long periods - just not MUCH power in any
    SHORT period. Iron/air and 'glass' and such ...
    you're not gonna run EVs, likely not even phones,
    on them. Edison batteries are kinda good, but
    a maint nightmare - potassium hydroxide solution
    for the electrolyte.

    For EVs ... it's either lithium or sodium at
    this point. Nothing else will deliver. Due
    to the fire/explosion issue, strongly rec
    sodium whenever possible now. We keep hearing
    promises of more - but it never makes it off
    the lab bench.

    And DO have a little scar from when a LiPo
    battery mysteriously went up in a big red
    flame ... just barely TOUCHED the thing ....

    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Charlie Gibbs@cgibbs@kltpzyxm.invalid to comp.os.linux.misc,alt.folklore.computers on Mon Nov 16 05:03:59 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-15, Johann 'Myrkraverk' Oskarsson <johann@myrkraverk.invalid> wrote:

    JSON can be just as complicated. There's no rule that an "id": in a sub object is the same as the "id": in the parent object. The one thing XML
    has above JSON is the ability to -- in theory -- unambiguously embed one
    XML document in another.

    Ugh. That headache hasn't hit me yet; the sources I parse are all
    JSON. I do JSON data collection from various sources; I wrote a JSON
    parser and added it to my private library. Its arguments include a
    table whose elements comprise a parent and sub ID pair and a pointer
    to where to store the corresponding data value (plus a maximum length).
    One of our data sources uses the same ID in a sub-object as another
    source uses as a parent object. As long as data from the two aren't
    mixed things work for me. So far, anyway.
    --
    /~\ Charlie Gibbs | In this world there are
    \ / <cgibbs@kltpzyxm.invalid> | two kinds of people:
    X I'm really at ac.dekanfrus | 1. Those who can extrapolate
    / \ if you read it the right way. | from incomplete data.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Charlie Gibbs@cgibbs@kltpzyxm.invalid to comp.os.linux.misc on Mon Nov 16 05:04:00 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-16, rbowman <bowman@montana.com> wrote:
    On Sun, 15 Nov 2026 20:23:03 +0000, The Natural Philosopher wrote:

    On 15/11/2026 20:19, Rich wrote:

    And, for USA travelers, TSA security theater line wait times, which can
    be substantial sometimes.

    Indeed. the fact is sadly that the bearded bastards like to destroy shit
    and planes are very handy and very vulnerable

    Right.. My cousin, who was in her 80s and brought to the boarding gate in
    a wheel chair since she can't walk far, was hassled by TSA who confiscated her jar of hand cream since it was larger than the 0.3 fl oz allowed. Typical terrorist profile.

    I've heard that another little old lady had her knitting needles
    confiscated because they were afraid she might knit an Afghan.
    --
    /~\ Charlie Gibbs | In this world there are
    \ / <cgibbs@kltpzyxm.invalid> | two kinds of people:
    X I'm really at ac.dekanfrus | 1. Those who can extrapolate
    / \ if you read it the right way. | from incomplete data.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Thomas Koenig@tkoenig@netcologne.de to comp.os.linux.misc,alt.folklore.computers on Mon Nov 16 09:29:49 2026
    From Newsgroup: comp.os.linux.misc

    Johann 'Myrkraverk' Oskarsson <johann@myrkraverk.invalid> schrieb:
    On 11/16/2026 1:07 AM, Thomas Koenig wrote:
    Johann 'Myrkraverk' Oskarsson <johann@myrkraverk.invalid> schrieb:

    The problem of /how do we get $amount into every employee's bank ac-
    count/ with thousands of employees and perhaps a single payroll account
    for the company does not sound like a solution waiting for a cluster of
    Linux boxes.

    SAP moved away from mainframes in the early 1990s with SAP R3, and
    they do run payrolls.

    And SAP S4/HANA runs on... you guessed it, Linux, exclusively.
    It does not even run on mainframes any more - it's Intel (not AMD)
    or IBM POWER.


    It's not exactly clear from https://en.wikipedia.org/wiki/SAP_R/3 but wouldn't the payroll system still run on a single machine, with perhaps
    the database on another, and different parts of the application on even
    more nodes? Would that machine qualify as a /mainframe/?

    Conceptually, SAP is centered around a single database ("single source
    of truth"). Physically, this will probably be mirrored, different
    parts may be on different machines, ...
    --
    This USENET posting was made without artificial intelligence,
    artificial impertinence, artificial arrogance, artificial stupidity,
    artificial flavorings or artificial colorants.
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Carlos E.R.@robin_listas@es.invalid to comp.os.linux.misc on Mon Nov 16 11:22:16 2026
    From Newsgroup: comp.os.linux.misc

    On 2026-11-16 06:04, Charlie Gibbs wrote:
    On 2026-11-16, rbowman <bowman@montana.com> wrote:
    On Sun, 15 Nov 2026 20:23:03 +0000, The Natural Philosopher wrote:

    On 15/11/2026 20:19, Rich wrote:

    And, for USA travelers, TSA security theater line wait times, which can >>>> be substantial sometimes.

    Indeed. the fact is sadly that the bearded bastards like to destroy shit >>> and planes are very handy and very vulnerable

    Right.. My cousin, who was in her 80s and brought to the boarding gate in >> a wheel chair since she can't walk far, was hassled by TSA who confiscated >> her jar of hand cream since it was larger than the 0.3 fl oz allowed.
    Typical terrorist profile.

    I've heard that another little old lady had her knitting needles
    confiscated because they were afraid she might knit an Afghan.


    I carry (and I avoid the USA if I can) a water bottle that I fill in a
    toilet tap, or a drinking fountain if I find it. But I must remember to
    empty it before arriving at security, of I have to drink it all. I am
    not allowed to empty it somewhere, nor drink half of it to prove it is drinking water and not explosives. And that is in Spain.
    --
    Cheers, Carlos.
    ESEfc-Efc+, EUEfc-Efc|;
    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From The Natural Philosopher@tnp@invalid.invalid to comp.os.linux.misc on Mon Nov 16 11:26:50 2026
    From Newsgroup: comp.os.linux.misc

    On 16/11/2026 03:12, rbowman wrote:
    And, for USA travelers, TSA security theater line wait times, which can
    be substantial sometimes.

    Indeed. the fact is sadly that the bearded bastards like to destroy shit
    and planes are very handy and very vulnerable
    Right.. My cousin, who was in her 80s and brought to the boarding gate in
    a wheel chair since she can't walk far, was hassled by TSA who confiscated her jar of hand cream since it was larger than the 0.3 fl oz allowed.
    Typical terrorist profile.

    There is a difference between the justification for the 'service' and
    how the amoeba at the sharp end uses it to be an asshole.
    --
    "And if the blind lead the blind, both shall fall into the ditch".

    Gospel of St. Mathew 15:14


    --- Synchronet 3.22a-Linux NewsLink 1.2
  • From Alexander Schreiber@als@usenet.thangorodrim.de to comp.os.linux.misc,alt.folklore.computers on Mon Nov 16 13:39:20 2026
    From Newsgroup: comp.os.linux.misc

    Lawrence DrCOOliveiro <ldo@nz.invalid> wrote:
    On Sun, 15 Nov 2026 22:40:31 +0100, Alexander Schreiber wrote:

    There was even a time when IBM actively promoted their mainframes as
    HA hosts for running lots of Linux VMs.

    Why not any more?

    I haven't really paid attention to it, maybe they are still doing it?
    No idea.

    I mean, technically they are of course correct, but using a mainframe to
    run Linux VMs might not be the most cost effective approach ...

    Kind regards,
    Alex.
    --
    "Opportunity is missed by most people because it is dressed in overalls and
    looks like work." -- Thomas A. Edison
    --- Synchronet 3.22a-Linux NewsLink 1.2