From Newsgroup: alt.folklore.computers

On 8/20/26 07:53, John Ames wrote:

On Thu, 20 Aug 2026 14:12:09 +0800
Johann 'Myrkraverk' Oskarsson <johann@myrkraverk.invalid> wrote:

My /white tower/ of blinkenlights has the fan blinking, and the two
RAM sticks. The /graphical prognostication unit/ also blinks an
arrow and a Gigabyte.

It's a veritable /industrial computer/!

AFAIK those lights don't *tell* you anything, though, other than that
the power's on. I know the BeBox blinkenlights would indicate CPU load
in either of the dual processors, but I don't think anyone has done an equivalent for modern motherboards and their surfeit of LED pulsers...?

Even in the elder days their use was fading. Sometimes a light pattern
might indicate a loop. I can recall an IBMer debugging on a 360/30 using console switches and lights, but they were gone by the System/370.
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: alt.folklore.computers

On 20/08/2026 10:53 PM, John Ames wrote:

On Thu, 20 Aug 2026 14:12:09 +0800
Johann 'Myrkraverk' Oskarsson <johann@myrkraverk.invalid> wrote:

My /white tower/ of blinkenlights has the fan blinking, and the two
RAM sticks. The /graphical prognostication unit/ also blinks an
arrow and a Gigabyte.

It's a veritable /industrial computer/!

AFAIK those lights don't *tell* you anything, though, other than that
the power's on. I know the BeBox blinkenlights would indicate CPU load
in either of the dual processors, but I don't think anyone has done an equivalent for modern motherboards and their surfeit of LED pulsers...?

No, and now that I think about it, that's /weird/. Like really weird.
Why not use the R.A.M. stick blinkenlights to tell me the memory module
is working? Of course, putting E.C.C. D.R.A.M. in a consumer tower is /anathema/, so I don't know how one would implement it. At a guess, my motherboard won't accept any.

And no, I didn't bother to try to find out if my memory sticks have
E.C.C. before writing this followup. If it is, it'll be a pleasant
surprise for the future.

I only have a single socket C.P.U., and I believe it has eight cores,
and there are no good blinkenlights to tell me which core is busy with
what.

The /graphical prognostication unit/ just blinks. It won't tell me any-
thing about how many polygons the OpenGL is drawing. Not to mention it
has no visible blinkenlight for P.H.I.G.S.

All of this is just outright bizarre.
--
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 ( ;; ) _:;
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: alt.folklore.computers

On 2026-08-20, Peter Flass <Peter@Iron-Spring.com> wrote:

On 8/20/26 07:53, John Ames wrote:

On Thu, 20 Aug 2026 14:12:09 +0800
Johann 'Myrkraverk' Oskarsson <johann@myrkraverk.invalid> wrote:

My /white tower/ of blinkenlights has the fan blinking, and the two
RAM sticks. The /graphical prognostication unit/ also blinks an
arrow and a Gigabyte.

It's a veritable /industrial computer/!

AFAIK those lights don't *tell* you anything, though, other than that
the power's on. I know the BeBox blinkenlights would indicate CPU load
in either of the dual processors, but I don't think anyone has done an
equivalent for modern motherboards and their surfeit of LED pulsers...?

Yeah, it's all just bling, existing solely to win pissing contests.
The last thing I want is pulsating rainbows everywhere - especially
on the keyboard. :-p

Even in the elder days their use was fading. Sometimes a light pattern
might indicate a loop. I can recall an IBMer debugging on a 360/30 using console switches and lights, but they were gone by the System/370.

The Univac 90/30s on which I mostly worked from 1975 to 1985 were
among the last machines with a couple of rows of LEDs with rollers
by which you could select various internal signals. I found one
display where the internal disk controller would display the
currently-active drive number and the length of the current seek;
it was useful for detecting disk thrashing, especially after the
new drives came out that had opaque covers so you could no longer
see the head assembly and its movements.

The earlier 9300 would by default display the first two bytes of
the currently-executing instruction. Due to the inertia of the
incandescent lamps the most recent instructions were effectively
ORed together. However, you could tell whether a program was CPU-
or I/O-bound (and if I/O-bound, which device it was waiting for).
Naturally I wrote some hacks which would display strange things,
like a 4-bit binary counter that would cycle repeatedly. Since this
machine treated 0x00 as a NOP, I wrote a program that cleared most of
memory, put a return at the bottom, and called it while waiting for
the card reader to complete. It would run the reader at full speed
with all of the lights out, something which was normally impossible.
I titled the program "Read Cards in the Dark".
--
/~\ 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 Newsgroup: alt.folklore.computers

Charlie Gibbs <cgibbs@kltpzyxm.invalid> writes:

On 2026-08-20, Peter Flass <Peter@Iron-Spring.com> wrote:

Even in the elder days their use was fading. Sometimes a light pattern
might indicate a loop. I can recall an IBMer debugging on a 360/30 using
console switches and lights, but they were gone by the System/370.

The Univac 90/30s on which I mostly worked from 1975 to 1985 were
among the last machines with a couple of rows of LEDs with rollers
by which you could select various internal signals. I found one
display where the internal disk controller would display the
currently-active drive number and the length of the current seek;
it was useful for detecting disk thrashing, especially after the
new drives came out that had opaque covers so you could no longer
see the head assembly and its movements.

The Burroughs B6800 was famous for displaying the
big B (B in a circle) logo on the console light
matrix during the idle loop.

We had the engineering prototype B4800 in the software
lab and had replaced the metal skins (190cm tall) with
smoked glass. The processor logic 15cm tall cards were racked
in the cabinets in trays, with the non-backplane
end of each card loaded with signal LEDs.

Halt/Loading the system in the dark was impressive to watch.

The most useful diagnostics were the channel lights, which
blinked for every I/O operation initiated to the channel;
we actually built a portable version of the channel lights
to plug into the subsequent B4900 I/O processor to help debugging the
MCP.
--- Synchronet 3.22a-Linux NewsLink 1.2