NetNews Offline 2/NetNews Offline Volume 2.iso/news/comp/lang/c++-part2/13589

Internet Message Format | 1996-08-05 | 3.8 KB

Path: mail2news.demon.co.uk!genesis.demon.co.uk
From: Lawrence Kirby <[e-mail]>
Newsgroups: comp.lang.c,comp.lang.c++,comp.object,comp.software-eng
Subject: Re: Beware of "C" Hackers -- A rebuttal to Bertrand Meyer
Date: Tue, 26 Mar 96 12:27:13 GMT
Organization: none
Message-ID: <[e-mail]>
References: <[e-mail]> <[e-mail]> <[e-mail]> <[e-mail]> <[e-mail]> <[e-mail]>
Reply-To: [e-mail]
X-NNTP-Posting-Host: genesis.demon.co.uk
X-Newsreader: Demon Internet Simple News v1.27
X-Mail2News-Path: genesis.demon.co.uk

In article <[e-mail]> [e-mail] "Larry Weiss" writes:

>Kazimir Kylheku wrote:
> >
> > >Kazimir Kylheku wrote:
> > > >  this I think is a major flaw in the otherwise great book.
> >
> > The book is a classic work, so it might as well be left alone, with its few
> > minor warts and all.
>
>So, which is it?  "major" or "minor" ?

IMHO minor because it is not leading you on to write poor or non-conforming
C code. By comparison the bad description of scanf and invalid redefinition
of standard library functions in examples are much more serious. I've
criticised Schildt in the past for leading by bad example; where K&R do the
same they deserve the same criticism.

>I don't want to allow this to pass so casually.   K&R are certainly selling
>a bunch of books to people on the premise that they have expertise in the
>subject.

There is no doubt that they do. The problem is in assuming that they are
perfect which, as far as standard C is concerned, they are clearly not.
However they make few mistakes compared to most other authors.

>   If they make fundamental errors in their text (and I'm starting
>to believe that they are, in fact, misrepresenting the language C with respect
>to the standard library) then we should shed some light on these mistakes
>and maybe influence the next revision.

I suspect that the full ANSI version of K&R2 was rushed. I'm sure K&R like
everybody else understand the implications of the ANSI standard much more
now than they did then. There are certainly errors in K&R2 that warrant
revision.

>This is exactly why I started this discussion.  I am now even more convinced
>that most programmers do not even yet understand the implications of some of
>the most important decisions of the C Standardization effort.

There are 3 categories of function you can call:

1. Functions that are well defined by the C language because you have
   the source code

2. Functions that are defined by the language itself i.e. standard library
   functions.

3. 'Other library' functions. Calling these results in undefined behaviour
   as far as the C language is concerned (or you had already invoked
   undefined behaviour by including a non-standard header file).

The only things special about standard library functions are that
they are guaranteed to exist with well-defined semantics in any conforming
hosted implementation, and that certain identifiers are reserved. So long
as you understand the rules it doesn't make a practical difference whether
you consider the standard library to be part of the language or not -
indeed a (freestanding) C language implementation can exist without the
standard library.

However saying that the standard library is not a part of the language is
certainly going against the grain and I don't think it is a helpful. And of
course it is a simple position to take that anything defined in ISO/IEC
9899-1990 is part of the C language. It guess it boils down to what you
mean by "the C language" (which is quite important to a newsgroup
called comp.lang.c!)

--
-----------------------------------------
Lawrence Kirby | [e-mail]
Wilts, England | [e-mail]
-----------------------------------------