HeathField Strange Ideas!

[Richard Heathfield]

Chris Hills said:

I think you have just proved his point.

On the contrary. As you quoted him saying, "do whatever the hell you
want with C" - which is certainly not an attempt to use the standard to
restrict what people can do with C. On the contrary, the statement
embraces the rich fullness of C - anything that can be done can be done
in the C language, given appropriate library support. Lots of
newsgroups exist for discussing such library support. This newsgroup
exists for discussing using the language itself (including the standard
library).

That is for comp.std.c

No, that's for discussing the C Standard. I am given to understand that
you are on the ISO C Committee, so you ought to be able to tell the
difference between a language standard and a standard language.

<snip>

 

[Richard Bos]

That is for comp.std.c This group is for the use of C.

No, it's not, and since you read that group as well (or so you say) it
is disappointing that you should say so. comp.std.c is about the C
_Standard_, about the document itself, that is, and about the
Standardisation procedure. A question like "Does this sentence in the
ISO C Standard mean what it looks like it means" belongs there. A
question like "how can I use this feature of Standard C" belongs here.
Not that difficult a distinction, IYAM.

The "standard C" the pedants here refer to seems to encompass ANY
version of C including unpublished drafts but not the real world use of
C.

Ah, yes, the "if it isn't blinged up to the hilt with bells and whistles
it doesn't happen in the real world" argument. That was untrue last time
someone (Kenny or jacob, probably) trotted it out, and it's still
untrue.

Richard

 

[Richard]

No, it's not, and since you read that group as well (or so you say) it
is disappointing that you should say so. comp.std.c is about the C
_Standard_, about the document itself, that is, and about the
Standardisation procedure. A question like "Does this sentence in the
ISO C Standard mean what it looks like it means" belongs there. A
question like "how can I use this feature of Standard C" belongs here.
Not that difficult a distinction, IYAM.


Ah, yes, the "if it isn't blinged up to the hilt with bells and whistles
it doesn't happen in the real world" argument. That was untrue last time
someone (Kenny or jacob, probably) trotted it out, and it's still
untrue.

Richard

What are you talking about? This groups covers at least 2 standards and
one "beta" standard.

There is no "one" standard C as this group proves on a daily basis.

This group is for C, the language, not ONE single standard.

The rest is word games.

 

[Kelsey Bjarnason]

This guy 'Richard Heathfield' and a few others seem to be dictating
nonsense about the ISO standards etc and sending out the impression that
C is very restricted. For Example: There have been posts made asking for
help on basic things like simple keyboard input routines which are
quickly dismissed to be impossible to do with compliant code.

"Strictly conforming code", generally. Since C has no concept of
keyboards, it would be difficult indeed to provide keyboard functions
which used C and only C, as defined by the standard. Of course, C with
the addition of implementation-specific libraries makes it easy,
generally, but not portable.

The standards are on our side and there to help us. I trust most can see
why.

Yes; they provide a portable, reliable, predictable platform upon which to
build endless reams of code.

And I have also noted he has tried to copy and enlighten on, the much
respected, Kernighan & Ritchies works. Again who does this guy think he
is?

Probably he thinks he's someone moderately clueful about C. He's right.

I find utterly annoying that this person has had a book published
entitled 'C Unleashed'. This title implying that the book will teach a
programmer to write 'go anywhere code'.

Doesn't it? Oh, it may well contain things which might be hard to pull
off in the tiniest embedded systems, but apart from oddball edge cases,
doesn't it?

 

[Keith Thompson]

"Strictly conforming code", generally. Since C has no concept of
keyboards, it would be difficult indeed to provide keyboard functions
which used C and only C, as defined by the standard. Of course, C with
the addition of implementation-specific libraries makes it easy,
generally, but not portable.

[...]

No, strict conformance is an extremely narrow restriction. It
excludes unspecified, undefined, and implementation-defined behavior,
all of which can be topical here, though we tend to discuss the range
of possibilities rather than going into detail about what some
particularly implementation does.

 

[Keith Thompson]

That is for comp.std.c This group is for the use of C.
The "standard C" the pedants here refer to seems to encompass ANY
version of C including unpublished drafts but not the real world use
of C.

The actual scope of comp.std.c has been explained here many many times
(it's for discussion of the C standard document). Rather than just
asserting that it's for discussion of standard C (as opposed to
whatever you think comp.lang.c is for), please explain why you think
you're right and nearly everyone else is wrong on this point.

[...]

If all those who yell "OT!" every 5 minutes refrained from doing so I
think we would have a far better group and the signal to noise ratio
would drop because we would not have then constant "is it OT"
argument raging.

I have a counterproposal. Let everyone who complains about statements
that something is off-topic refrain from doing so. I believe this
would improve the signal to noise ratio even more. Alas, I doubt that
we'll ever find out.

Failing that, tell us what *you* think should be considered topical
here. Write up a set of proposed guidelines and post it here.

 

[Kelsey Bjarnason]

[snips]

No, strict conformance is an extremely narrow restriction. It
excludes unspecified, undefined, and implementation-defined behavior,
all of which can be topical here, though we tend to discuss the range
of possibilities rather than going into detail about what some
particularly implementation does.

Perhaps I wasn't clear; I meant that the bit about keyboards wasn't doable
in strictly conforming code, not that strictly conforming code is the only
topical flavour here.

 

[Al Balmer]

That is for comp.std.c

Backward. comp.std.c is for the discussion of the C Standard, not
standard C.

 

[Keith Thompson]

[snips]

No, strict conformance is an extremely narrow restriction. It
excludes unspecified, undefined, and implementation-defined behavior,
all of which can be topical here, though we tend to discuss the range
of possibilities rather than going into detail about what some
particularly implementation does.

Perhaps I wasn't clear; I meant that the bit about keyboards wasn't doable
in strictly conforming code, not that strictly conforming code is the only
topical flavour here.

The original (or unoriginal) troll said something about "compliant
code". It's true that you can't do certain kinds of keyboard input in
strictly conforming code -- but you can't print the value of INT_MAX
or the result of rand() in strictly conforming code either.

What's relevant here is that you can't do certain things in *portable*
code, i.e., without using extensions that are better discussed in
system-specific forums. Strict conformance doesn't enter into it.

It's not a huge deal, but the point is that I'm not sure why you
introduced strict conformance into the discussion.

 

[Kelsey Bjarnason]

[snips]

No, strict conformance is an extremely narrow restriction. It
excludes unspecified, undefined, and implementation-defined behavior,
all of which can be topical here, though we tend to discuss the range
of possibilities rather than going into detail about what some
particularly implementation does.

Perhaps I wasn't clear; I meant that the bit about keyboards wasn't doable
in strictly conforming code, not that strictly conforming code is the only
topical flavour here.

The original (or unoriginal) troll said something about "compliant
code". It's true that you can't do certain kinds of keyboard input in
strictly conforming code -- but you can't print the value of INT_MAX
or the result of rand() in strictly conforming code either.

What's relevant here is that you can't do certain things in *portable*
code, i.e., without using extensions that are better discussed in
system-specific forums. Strict conformance doesn't enter into it.

It's not a huge deal, but the point is that I'm not sure why you
introduced strict conformance into the discussion.

"There have been posts made asking for help on basic things like simple
keyboard input routines which are quickly dismissed to be impossible to
do with compliant code."

"Compliant" is meaningless. "Conforming" allows all sorts of nifty
goodness - including bringing in additional libraries if need be. Thus
keyboard input and the like are perfectly valid in a conforming program,
so whatever "compliant" means, it must be something other than conforming;
the only real options are non-conforming and strictly conforming.

If it were to mean "non-conforming" then the argument doesn't make sense,
as there is nothing about "non-conforming" that tells us you cannot use
keyboard support or anything else; this leaves only "strictly conforming".

And yes, Virginia, there is a lack of keyboard support in strictly
conforming programs, which means for his argument to make any sense at
all, we (he) must be discussing strictly conforming programs and thus the
observations to which he refers are correct - you can't get there from
here.

 

[Old Wolf]

It would also expand the cause of using portable c (where appropriate)
at the moment most who drop in and ask an "OT" question usually leave
very quickly thinking that the group is full of a pedant net nannines
who are not connected to the real world. That helps no one.

Seems to me that it is very helpful, if it causes OT
posters to leave quickly.

 

[Keith Thompson]

"There have been posts made asking for help on basic things like simple
keyboard input routines which are quickly dismissed to be impossible to
do with compliant code."

"Compliant" is meaningless. "Conforming" allows all sorts of nifty
goodness - including bringing in additional libraries if need be. Thus
keyboard input and the like are perfectly valid in a conforming program,
so whatever "compliant" means, it must be something other than conforming;
the only real options are non-conforming and strictly conforming.

If it were to mean "non-conforming" then the argument doesn't make sense,
as there is nothing about "non-conforming" that tells us you cannot use
keyboard support or anything else; this leaves only "strictly conforming".

And yes, Virginia, there is a lack of keyboard support in strictly
conforming programs, which means for his argument to make any sense at
all, we (he) must be discussing strictly conforming programs and thus the
observations to which he refers are correct - you can't get there from
here.

I think you picked "strictly conforming" because it's one of the
categories defined by the standard, but it's still far too narrow for
the current discussion. The troll whined about being told we can't
discuss keyboard input here, but we discuss non-strictly conforming
code all the time.

Yes, "compliant" is meaningless unless it's defined, and "conforming"
(a term defined in the standard) is too loose. But we're not limited
to categories defined by the standard.

The set of programs that are topical in comp.lang.c is not a category
for which the standard defines a term -- but we can. I've seen the
term "clc-compliant", which is by definition exactly the right term,
but the trick is defining what it means. "Portable" is a good
approximation.

The relevant statement is that you can't do certain kinds of keyboard
input in portable code.

 

[Flash Gordon]

Keith Thompson wrote, On 30/08/07 00:05:

Kelsey Bjarnason <[email protected]> writes:

Yes, "compliant" is meaningless unless it's defined, and "conforming"
(a term defined in the standard) is too loose. But we're not limited
to categories defined by the standard.

The set of programs that are topical in comp.lang.c is not a category
for which the standard defines a term -- but we can. I've seen the
term "clc-compliant", which is by definition exactly the right term,
but the trick is defining what it means. "Portable" is a good
approximation.

How about, "uses only facilities documented by the C standard." I might
have used "defined" or "specified" but that would be confusing as the
standard defined "undefined behaviour" and "unspecified behaviour". I'm
sure there are holes, but I think it comes closer than "portable".

The relevant statement is that you can't do certain kinds of keyboard
input in portable code.

True.

 

[pete]

Keith Thompson wrote, On 30/08/07 00:05:

How about, "uses only facilities documented by the C standard."
I might
have used "defined" or "specified" but that would be confusing as the
standard defined "undefined behaviour" and "unspecified behaviour".

C99 introduces a term that I like: "correct program"

ISO/IEC 9899:1999 (E)
4. Conformance

3 A program that is correct in all other aspects,
operating on correct data, containing unspecified behavior
shall be a correct program and act in accordance with 5.1.2.3.

 

[Chris Hills]

Seems to me that it is very helpful, if it causes OT
posters to leave quickly.

They are only OT in the eyes of a small but vocal group who are slowly
killing this NG.

At one time c.l.c used to be the place to go for information on C . Now
there are many other forums and c.l.c is getting reputation for
unhelpful pedantry

 

[Richard]

They are only OT in the eyes of a small but vocal group who are
slowly killing this NG.

At one time c.l.c used to be the place to go for information on C
. Now there are many other forums and c.l.c is getting reputation for
unhelpful pedantry

It would appear that a few have placed themselves in some sort of Dan
Pop copy cat mode. They only have the rudeness bit down to a T
unfortunately.

 

[jacob navia]

They are only OT in the eyes of a small but vocal group who are slowly
killing this NG.

At one time c.l.c used to be the place to go for information on C . Now
there are many other forums and c.l.c is getting reputation for
unhelpful pedantry

I would second that. Well said.

 

[Chris Hills]

It would appear that a few have placed themselves in some sort of Dan
Pop copy cat mode. They only have the rudeness bit down to a T
unfortunately.

That is unfair... they are not rude. A bit abrupt maybe with the "OT"
but not rude. They may not be quite as knowledgeable as Dan but they are
still largely knowledgeable and experienced.

My only real gripe is the pedantry and restriction as to the purpose of
this NG. I would like it to be what it should be. Discussions on C in
general.

As all of the worlds C compilers have extensions to the C standard and
or do not fully implement the C standard (which is why the pedants
include in their definition all and any obsolete or unpublished
versions. Drafts are not published standards) The are trying to restrict
this NG to an artificial subset of the language.

I think we need to be able to discuss C as it is used.

A discussion on accessing the screen and or keyboard using C is a valid
one here. There is no portable way of doing it but there are some
"standard" ways of doing it to make it more portable. EG Curses is
available for many platforms.

Also a discussion of this sort will prompt programmers to look for ways
of doing it more portably or at least contain their non portable stuff
to one area. It will help to highlight the non standard parts of their
code.

Simply sending people to Linux and Windows NG's removes them from this
one. They will probably end up on a Windows or Gcc/Linux group without
ever realising that MS-C is GCC is NOT the standard.

As it is many drop in here and leave never to return after the "OT"
response they get. At one time ClC used to be the centre for C
discussion world wide. Now it is one of the minor backwaters.

 

[Richard]

That is unfair... they are not rude. A bit abrupt maybe with the "OT"

No, they can be.

but not rude. They may not be quite as knowledgeable as Dan but they
are still largely knowledgeable and experienced.

But make this relatively useless some of the time in the interminable
word games and attempts to keep everything on what they see as "On
topic".

My only real gripe is the pedantry and restriction as to the purpose
of this NG. I would like it to be what it should be. Discussions on
C in general.

Agreed. As most do.

As all of the worlds C compilers have extensions to the C standard and
or do not fully implement the C standard (which is why the pedants
include in their definition all and any obsolete or unpublished
versions. Drafts are not published standards) The are trying to
restrict this NG to an artificial subset of the language.
Agreed.


I think we need to be able to discuss C as it is used.
Yup.


A discussion on accessing the screen and or keyboard using C is a
valid one here. There is no portable way of doing it but there are
some "standard" ways of doing it to make it more portable. EG Curses
is available for many platforms.

Yup. And it's as easy to mention that than scream "off topic". There is
no other C group that I am aware of. People should be able to discuss
such things in a C manner looking to best utilise such features using
C. Frequently these libraries ARE portable across 3 or 4 platforms.

Also a discussion of this sort will prompt programmers to look for
ways of doing it more portably or at least contain their non portable
stuff to one area. It will help to highlight the non standard parts
of their code.

Agreed again.

Simply sending people to Linux and Windows NG's removes them from this
one. They will probably end up on a Windows or Gcc/Linux group
without ever realising that MS-C is GCC is NOT the standard.

As it is many drop in here and leave never to return after the "OT"
response they get. At one time ClC used to be the centre for C
discussion world wide. Now it is one of the minor backwaters.

And a core of about 5 people are responsible from what I have gathered.

 

[Richard Heathfield]

Chris Hills said:

writes


That is unfair...

Chris, you're arguing with a troll. What's more, you're arguing with a
troll who broadly agrees with you. When you find yourself in broad
agreement with a troll, you need to think very carefully about why that
is.

My only real gripe is the pedantry and restriction as to the purpose
of this NG. I would like it to be what it should be. Discussions on
C in general.

That is already what it is - discussions about C.

I think we need to be able to discuss C as it is used.

Agreed. And we do.

A discussion on accessing the screen and or keyboard using C is a
valid one here.

Show me. Show me how you can access the screen or keyboard using C.

There is no portable way of doing it but there are some
"standard" ways of doing it to make it more portable. EG Curses is
available for many platforms.

Show me. Show me this "standard" way for accessing the screen or
keyboard in C. This I have to see.