Cross-platform: Coloured text, Networking, Multithreading

[Nick Keighley]

Martin Ambuhl said:

I find it surprising that so many people who witnessed that discussion and
should therefore be aware of [Tomás Ó hÉilidhe's] racist attitude continue
to enter into
further technical discussions with him despite the absence of any apology
or retraction on his behalf. (But then I find many things surprising.)

unfortunatly my killfile is stored in memory (my memory!) and is
hence unreliable

 

[Chris McDonald]

The groupname field was not designed to be an all encompassing
definition of its purpose.

Take the groups ending in .tv - related to audiovisual entertainment, or
cross-dressing, or the state of Tuvalu, or what? There are alt. groups
that have highly cryptic names on purpose. Do you propose to go through
the entire hierarchy and enforce descriptive names? Won't work. For one
thing, the groupname is too short. For another, defining the group is
the charter or equivalent is for. For a third, if you lurk for a few
days the topicality will be clear. If you still can't figure out the
topicality, you can always ask.

Some servers seem to support a description field, but I suspect few new
users bother to read those.

All true, but no-one proposed reviewing and changing all group names.
My comment was simply that the original suggestion was not redundant.

Not even the most staunch of c.l.c. regulars could deny that the name
comp.lang.iso-c would be a name better reflecting the desired discussion
of greatest interest to those same regulars.

It's also less likely that first-time posters wanting help with basic C
would choose comp.lang.iso-c over c.l.c.

 

[CBFalconer]

(not trolling) but the suggestion is *not* redundant.

It comes with the suggestion that the name comp.lang.iso-c would
provide a clearer label for those ill-informed about the current
use of comp.lang.c

Why? The C language is defined by the C standard, and has been
previously defined by C95, C90, C89, and the K&R books, in order of
age. Those definitions are well known, and pretty generally
available.

Extensions etc. are born of the abilities, needs, and biases of
individual developers. They do NOT define things that can be
counted on, outside of use on those particular systems.
comp.lang.c simply avoids those unknown and variable extensions,
and deals with the facets that are constant and known.

This definition of topicality here is well known, even though some
trolls attempt to deride it.

 

[Keith Thompson]

[...]
Not even the most staunch of c.l.c. regulars could deny that the name
comp.lang.iso-c would be a name better reflecting the desired discussion
of greatest interest to those same regulars.

Since C is *defined* by various generations of ISO documents,
a better proposal would be comp.lang.c and comp.lang.pidgin-c or
comp.lang.c-variants or comp.lang.c-ish.

It's also less likely that first-time posters wanting help with basic C
would choose comp.lang.iso-c over c.l.c.

The first-time posters are not the problem. The trouble is
with the repeat offenders, those who know better but choose worse.

In fact, I have no real problem with first-timers posting off-topic
questions. This isn't a good place to get detailed information about,
say, Win32 or POSIX, but we do tend to be pretty good at advising
people where find the actual experts on what they're asking about.

The idea that redirecting a poster to a more appropriate newsgroup is
somehow rude is, of course, absurd. I'm sure there are examples of
people doing it in a rude manner, but in my experience they're the
exception.

 

[Antoninus Twink]

Not even the most staunch of c.l.c. regulars could deny that the name
comp.lang.iso-c would be a name better reflecting the desired
discussion of greatest interest to those same regulars.

True - but of course, their particular view on "desirable discussions"
carries no more weight than the next man's. This is not a moderated
group, and there is no charter - each person must decide for themselves
what to post. There is no absolute topicality, no matter how often and
how loudly and how aggressively Heathfield & friends say there is.

It's also less likely that first-time posters wanting help with basic
C would choose comp.lang.iso-c over c.l.c.

Exactly. Very few people care about the minutiae of ISO C, whereas lots
and lots of people want to learn C. The "regs" would never support a
name change, because they know that c.l.iso-c would become a ghetto,
while clc would be an active forum for exchange of expertise.

They're perfectly happy with the status quo, where they can fuel their
egos by trying to bully newbies into accepting *their* definition of
"topicality".

 

[Chris McDonald]

Chris McDonald wrote:

K&R C is on topic here too.

Yes and, as would be confirmed by many regular readers and posters here,
each of whom's viewpoint carries equal validity, so could be the typical
approaches to supporting coloured text, networking, and multithreading
from within the C language. Hence the (never-ending) discussion.

 

[James Kuyper]

Chris McDonald wrote:
....

Not even the most staunch of c.l.c. regulars could deny that the name
comp.lang.iso-c would be a name better reflecting the desired discussion
of greatest interest to those same regulars.

I'm afraid you're wrong about that. The proposed name incorrectly
suggests that ISO C is some weird variant of C that's probably of
negligible interest to general C programmers. ISO C is C, and should in
fact be the primary concern of any C programmer who's not made a
deliberate well-thought-out decision that his code should be written in
a way that restricts its portability to a limited set of platforms.

It's also less likely that first-time posters wanting help with basic C
would choose comp.lang.iso-c over c.l.c.

First-time posters wanting help with basic C questions should be posting
to the group you want to call comp.lang.iso-c. As RH has already said
"ISO C is basic C". Actual questions about C are precisely what that
group should be restricted to. Take those away, and it becomes pointless.

Questions about the application domain, rather than the C code that that
the application is written in, should be directed at domain experts, not
C experts.

Questions about a particular implementation of C should be directed to
forums specific to that implementation, and similarly for questions
about C-like non-implementations of C, such as GNU-C.

OS-specific questions should go to OS-specific forums.

Questions about the C standard itself, rather than about the language
defined by that standard, should go to comp.std.c

However, questions that are about basic C itself most certainly should
go here (or to the place you want to call comp.lang.iso-c, should it be
created).

 

[Kenny McCormack]

The idea that redirecting a poster to a more appropriate newsgroup is
somehow rude is, of course, absurd. I'm sure there are examples of
people doing it in a rude manner, but in my experience they're the
exception.

You are hallucinating again.

 

[Chris McDonald]

First-time posters wanting help with basic C questions should be posting
to the group you want to call comp.lang.iso-c......

Actually, if you took the courtesy to read what I wrote, you'd appreciate
that I didn't make any statements about what I want.

But I will (providing you with some potential fodder) - I think that a
(new) group named comp.programming.c would accurately suggest where
questions about programming in C, and its dialects, and compiler-specific
extensions, and cross-platform APIs, could be posted.

 

[Phil Carmody]

Not even the most staunch of c.l.c. regulars could deny that the name
comp.lang.iso-c would be a name better reflecting the desired discussion
of greatest interest to those same regulars.

I deny that. C is C. C is ISO C. ISO C is C. End of.

If you think there's a need for alternatives that aren't adequately
supported elsewhere in the tree of groups, then I'd suggest something
like comp.lang.c.${VARIANT} for the bastardised versions.

Phil

 

[Eric Sosman]

Actually, if you took the courtesy to read what I wrote, you'd appreciate
that I didn't make any statements about what I want.

But I will (providing you with some potential fodder) - I think that a
(new) group named comp.programming.c would accurately suggest where
questions about programming in C, and its dialects, and compiler-specific
extensions, and cross-platform APIs, could be posted.

The suggestion has merit. There's a procedure somewhere
to ballot for a new newsgroup: I'm not familiar with how it
works, but if you can formulate a proposal along these lines
I'd probably vote for it.

 

[James Kuyper]

Actually, if you took the courtesy to read what I wrote, you'd appreciate
that I didn't make any statements about what I want.

You're correct. You expressed a strong opinion that this name was better
than the existing one, at least for certain purposes, an opinion you
hold so strongly that you suggested that no one could deny the validity
of that opinion. However, I can see now that I went overboard in
assuming that you wanted that name to be used - you didn't actually say so.

But I will (providing you with some potential fodder) - I think that a
(new) group named comp.programming.c would accurately suggest where
questions about programming in C, and its dialects, and compiler-specific
extensions, and cross-platform APIs, could be posted.

That sounds like a reasonable choice. The distinction between
"comp.lang.c" and "comp.programming.c" isn't as clear as I'd like from
the group names, but no group name of reasonable length could possibly
fully clarify that distinction.

 

[s0suk3]

I'm currently writing a program and I've got in mind to keep it as
portable as possible. In particular I want it to run on Linux and
Windows, but I'm also keeping an open mind to any machine that has a
screen and is capable of Ethernet networking. The program requires
three things that aren't available in the C Standard:
    1) Coloured text (for a console application)
    2) Raw socket networking
    3) Multithreading

For number 1, I've already written my own tiny little cross-platform
library (it uses ANSI escape sequences for Linux, and Win32 API
functions for Windows).

For number 2, I've again written my own tiny little cross-platform
library (it uses Berkeley Sockets for Linux, and winpcap for Windows
-- I would have used Winsock for Windows but Winsock no longer
supports raw sockets).

For number 3, well I'm just about to delve into it now. I see that
somebody has already ported the "pthreads" library to Win32, so that
looks like a very attractive option. Just before I go down that road
though, I'd like to ask if anyone has a better idea than to use
pthreads?

Perhaps too big for your purposes, but it's very portable:

http://www.mozilla.org/projects/nspr/

I'm not sure if it provides console text coloring. There's a newsgroup
for it:

mozilla.dev.tech.nspr

Sebastian

 

[Nick Keighley]

All true, but no-one proposed reviewing and changing all group names.
My comment was simply that the original suggestion was not redundant.

Not even the most staunch of c.l.c. regulars could deny that the name
comp.lang.iso-c  would be a name better reflecting the desired discussion
of greatest interest to those same regulars.

I would, I'm not sure if I qualify if I'm "staunch" or not.

clc also discusses K&R. What would your definition of
"The C Programming Language" if you ddin't use a standard?

 

[CBFalconer]

Chris McDonald wrote:
.... snip ...


That sounds like a reasonable choice. The distinction between
"comp.lang.c" and "comp.programming.c" isn't as clear as I'd like
from the group names, but no group name of reasonable length could
possibly fully clarify that distinction.

I disagree. comp.programming already exists, does not worry too
much about topicality (and is infested by Nilges/spinoza).
Strictly speaking it deals with algorithms, not languages. A
subgroup dealing with C will only confuse users.