From Newsgroup: comp.lang.c

by skinnest c i mean such c 'skin' that would be much more thin than
the present (also its strictly not skinnest but a compromise of thinning
and elegance)

i think on it form time to time on span of years, but its a hard task

i may show my recent thoughts:

i get the idea that int x you turn into _x (where _ is like "declare
int" so _x is :declare int x"

then i think assigment i would do x_ 1 (the other candidate for
assigment would be x 1, maybe better but it blocks a lot of syntax like argiment list

imo thui definitelly looks somewhat good, it also may be used to
function calls

_x _y foo a b
mean "int x, y = foo(a,b)"

x_ y_ foo a b
mean " x, y = foo(a,b)"

this is somewhat good becouse it shows very cheap which arguments are
in and which are out (in are those without underscore out are those with underscore

it also has advantage that result could be seen as a vector/tuple for immediate use ..this hovever still has some problems (though i like it generally)

= would be not used for assigments, it would be used only for
comparisons (comparsions?)

a=3?buum x y //(if(a==3) buum(x,y)

else probably should be used some other sign like -+

a=4?buum -+zoom

and i think may be used for populating ? like

300?a<200? boom

where or need a sign im not sure is space would sufice

a=2 a=3 a=4? zoom

loops also needs some sign temporalely take ??

a<10?? boom(a++) // while(a<10) boom(a++)

hard problem is how function definition should look like and i have it
not resolved yet

_x _y foo a_ b_ {} //?? maybe this could work but im not sure

im also not sure how to introduce types like float
simply maybe special unicode signs would be needed (something like _
with smal f above it)





--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.lang.c

fir pisze:

by skinnest c i mean such c 'skin' that would be much more thin than
the present (also its strictly not skinnest but a compromise of thinning
and elegance)

i think on it form time to time on span of years, but its a hard task

i may show my recent thoughts:

i get the idea that int x you turn into _x (where _ is like "declare
int" so _x is :declare int x"

then i think assigment i would do x_ 1 (the other candidate for
assigment would be x 1, maybe better but it blocks a lot of syntax like argiment list

imo thui definitelly looks somewhat good, it also may be used to
function calls

_x _y foo a b
-amean "int x, y =-a foo(a,b)"

x_ y_ foo a b
mean " x, y =-a foo(a,b)"

this is somewhat good becouse it shows very cheap which arguments are
in and which are out (in are those without underscore out are those with underscore

it also has advantage that result could be seen as a vector/tuple for immediate use ..this hovever still has some problems (though i like it generally)

= would be not used for assigments, it would be used only for
comparisons (comparsions?)

a=3?buum x y-a //(if(a==3) buum(x,y)

else probably should be used some other sign like -+

a=4?buum -+zoom

and i think may be used for populating ? like

300?a<200? boom

where or need a sign im not sure is space would sufice

a=2 a=3 a=4? zoom

loops also needs some sign temporalely take ??

a<10?? boom(a++) // while(a<10) boom(a++)

hard problem is how function definition should look like and i have it
not resolved yet

_x _y foo a_ b_ {} //?? maybe this could work but im not sure

im also not sure how to introduce types like float
simply maybe special unicode signs would be needed (something like _
with smal f above it)

as to tuple/vector arithmetic im not sure but maybe just must use
aprenthesis

(a b c)_ (1 2 3) // tuple assignment

(a b c) = (1 2 3) ? //tuple comparison

it could be done thinner but overally vector arithmentic is probably not
the priority..i find finally "," good as separator

so

a b c = 1 2 3 //i would find probably as a 5 element tuple

'akjankjan' would be an hoc enums and only enums i think
"shbshjbs" would be strings "A" would be chars

overally as i once said there is probably a good idea ti reach for
unicode signs and use them..i think much more elegance could be obtained
with unicode signs than reusing only old ascii

most problematic is still the function header thing

--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.lang.c

fir pisze:

fir pisze:

by skinnest c i mean such c 'skin' that would be much more thin than
the present (also its strictly not skinnest but a compromise of
thinning and elegance)

i think on it form time to time on span of years, but its a hard task


hard problem is how function definition should look like and i have it
not resolved yet

_x _y foo a_ b_ {} //?? maybe this could work but im not sure

i also think

foo ->x ->y _a _b { x-> a+b, y->a-b }

_a for "input int a"

x for "output int x"

here _x a+b would be "x=a+b" and x-> a+b would be just the same, but
maybe version for assigning output variables for clarity

it seems reasonably good, not to say there is no something better

--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.lang.c

fir pisze:

as to tuple/vector arithmetic im not sure but maybe just must use
aprenthesis

(a b c)_ (1 2 3)-a // tuple assignment

(a b c) = (1 2 3) ?-a //tuple comparison

as to this im not sure hoveverv as it looks bad
there is also to consider

a_b_c_ 2 3 4 //assign a tuple

a b c = 1 2 3 //meybe this one after all (for tuple comparsions)

the last one need to be checked what would be more usable/
what is more important, tuple comparsion or local comparsion in tuples
(the second one seem somewhat doubtfull, as honestly the value of
comparsion as an element in vector is kinda doubtfull)



--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.lang.c

fir pisze:

-ai also think

foo ->x ->y _a _b-a-a-a {-a x-> a+b,-a y->a-b-a }

ye this look not bad, my design still somewhat 'shines' sometimes,
despite im old and sick

(not to say the problem is to introdue other types and ot spoil it,
sothis is more example of new b than new c)
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.lang.c

fir pisze:

loops also needs some sign temporalely take ??

a<10?? boom(a++) // while(a<10) boom(a++)

thet wouled be while but there is still an option for for loop


10x boom

//ten times boom

10x20x

//double loop

thsi could be done especially if rake x from unicode there should be
something

if _10 would be an int[10] evector type im not sure how double array
would be _10 _20 cant be as its two veectors maybe _10x20 could be..but
still im not sure bcouse the syntax _a (thus _N) is locked thus _NxM cant be

__N could eventually but the construction _n as n is name not counter is locked too

_x is int x

_6x would be eventually vector of 6 ints

_*y could be eventually and unknown length array? then **y two
dimensional (?)

need to search some things here yet

--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.lang.c

fir pisze:

fir pisze:

loops also needs some sign temporalely take ??

a<10?? boom(a++) // while(a<10) boom(a++)

thet wouled be while but there is still an option for for loop

10x boom

//ten times boom

10x20x

//double loop

thsi could be done especially if rake x from unicode there should be something

if _10 would be an int[10] evector type im not sure how double array
would be _10 _20 cant be as its two veectors-a maybe _10x20 could be..but still im not sure bcouse the syntax _a (thus _N) is locked thus _NxM
cant be

__N could eventually but the construction _n as n is name not counter is locked too

_x is int x

_6x would be eventually vector of 6 ints

_*y could be eventually and unknown length array? then **y two
dimensional (?)

need to search some things here yet

well maybe.. i remember i had some problem
i wanted to aksess arrays with . like a.1,
a.2 but i thought cant do a.b when b is variable
of value say 2..bet here it seems possible

a.b (a is an array, b is variable) so it
probably means its at least partially good idea


_*a //declare undefined array of ints (?)

_b 2 //int b =2;

a.b //take element 2
or *a.b (?)

a.b _ a.b+1
or a_.b a.b+1 (?)

still this assigment operator looks a bit weird but at this moment im
quire sure "=" cant stay

"=" is over

--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.lang.c

fir pisze:

fir pisze:

fir pisze:

loops also needs some sign temporalely take ??

a<10?? boom(a++) // while(a<10) boom(a++)

thet wouled be while but there is still an option for for loop


10x boom

//ten times boom

10x20x

//double loop

thsi could be done especially if rake x from unicode there should be
something

if _10 would be an int[10] evector type im not sure how double array
would be _10 _20 cant be as its two veectors-a maybe _10x20 could
be..but still im not sure bcouse the syntax _a (thus _N) is locked
thus _NxM cant be

__N could eventually but the construction _n as n is name not counter
is locked too

_x is int x

_6x would be eventually vector of 6 ints

_*y could be eventually and unknown length array? then **y two
dimensional (?)

need to search some things here yet

-awell maybe.. i remember i had some problem
-a i wanted to aksess arrays with . like a.1,
-a a.2 but i thought cant-a do a.b when b is variable
-a of value say 2..bet here it seems possible

-a a.b (a is an array, b is variable) so it
-a probably means its at least partially good idea

-a _*a-a //declare undefined array of ints (?)

-a _b 2 //int b =2;

-a a.b //take element 2
-aor *a.b (?)

-a a.b _ a.b+1
or a_.b a.b+1-a (?)

possibly a_.b c syntax is better coz a.b_ 3 could be used naturally for
"b=3; a[b];"

so

foo _*a _i _c
{
a_.i c //analog for a[i]=c, same length as in old c
}

foo a i c

-astill this assigment operator looks a bit weird but at this moment im quire sure "=" cant stay

-a"=" is over

--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.lang.c

fir pisze:

fir pisze:

fir pisze:

so

-a foo _*a _i _c
-a {
-a-a-a a_.i c-a-a //analog for a[i]=c, same length as in old c
-a }

well a is input and output - input is a as an entry output is a contents
though im not sure it would be proper

foo ->*_a _i _c {a->.i c }

thsi goes down into more details (?)
--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.lang.c

fir pisze:

fir pisze:

fir pisze:

fir pisze:

so

-a-a foo _*a _i _c
-a-a {
-a-a-a-a a_.i c-a-a //analog for a[i]=c, same length as in old c
-a-a }

well a is input and output - input is a as an entry output is a contents though im not sure it would be proper

foo-a ->*_a _i _c {a->.i c }

thsi goes down into more details (?)

well meybe this is good?

i mean you eventually can even do foo ->_*->_a _i _c
where ->_a would mean input output a array as a structore
(pointer and boundary, it also mena acces to indexing it but no acces to
its contents) where ->_* could mean inpout and output of its contents

those are further details so maybe it even should be yet thinked


--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.lang.c

fir pisze:

im also not sure how to introduce types like float
simply maybe special unicode signs would be needed (something like _
with smal f above it)

it could be done with subscripts and superscripts but scandalously
unicode dont have all basic letters there afaik

b|ax 20 //float x = 20

xreu 10 // x = (float) 20, if there would be subscript f i put k instead

though for assigments the type is not needed rpobably so it could be

x_ 10 //if x is float, but subscripts are not bad imo


--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.lang.c

fir pisze:

fir pisze:

im also not sure how to introduce types like float
simply maybe special unicode signs would be needed (something like _
with smal f above it)

it could be done with subscripts and superscripts but scandalously
unicode dont have all basic letters there afaik

b|ax 20 //float x = 20

xreu 10 // x = (float) 20, if there would be subscript f i put k instead

though for assigments the type is not needed rpobably so it could be

x_ 10 //if x is float, but subscripts are not bad imo

those prev functions for floats would look like ?:

foo b|a*a b|oi reuc {ab|a.i c }

but it seems * should also be subscripted - noting that array a adress
is for read, float contents are for write (there is no subscript * in
unicode)

wventually it could be maybe

foo *b|aa b|oi reuc {ab|a.i c }


*b|aa means array of output float values

though then

7b|a can be a ector of 7 floats? maybe

--- Synchronet 3.22a-Linux NewsLink 1.2

From Newsgroup: comp.lang.c

fir pisze:

fir pisze:

fir pisze:

im also not sure how to introduce types like float
simply maybe special unicode signs would be needed (something like _
with smal f above it)

it could be done with subscripts and superscripts but scandalously
unicode dont have all basic letters there afaik

b|ax 20 //float x = 20

xreu 10 // x = (float) 20, if there would be subscript f i put k instead

though for assigments the type is not needed rpobably so it could be

x_ 10 //if x is float, but subscripts are not bad imo

those prev functions for floats would look like ?:

-a-a foo b|a*a b|oi reuc {ab|a.i c }

but it seems * should also be subscripted - noting that array a adress
is for read, float contents are for write (there is no subscript * in unicode)

wventually it could be maybe

-a-a foo *b|aa b|oi reuc {ab|a.i c }

this i know dont look god and need to be something other
but the version with dashes/over/underscores looks nice and
generally im quite happy for some conclusions here

for float shortcut i could eventually would take something like


foo -Uf -|x -|y //floats
{
f-Ux+y //f_x+y also allowed for floats?
}

foo ->i _x _y //ints
{
i->x+y //i_ x+y allowed?
}

though it seems maybe a bit too fragile
still there would be a need fo shorts/long chars doubles

and finally for defined complex types (no idea)

*b|aa means array of output float values

though then

-a7b|a can be a ector of 7 floats? maybe

--- Synchronet 3.22a-Linux NewsLink 1.2