From: Ned Ulbricht <nedu AT ee DOT washington DOT edu>
Newsgroups: comp.os.msdos.djgpp
Subject: Re: Suggestion: Portability section for libc docs
Date: Tue, 10 Feb 1998 06:20:39 -0800
Organization: University of Washington
Lines: 55
Message-ID: <34E06236.3D3B@ee.washington.edu>
References: <m0y2A7f-000S2cC AT inti DOT gov DOT ar>
NNTP-Posting-Host: cs239-4.student.washington.edu
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
To: djgpp AT delorie DOT com
DJ-Gateway: from newsgroup comp.os.msdos.djgpp
Precedence: bulk

Nate Eldredge <eldredge AT ap DOT net> wrote:

> Although DJGPP is of course the best DOS development environment in the
> known universe :), there is one feature I miss from Turbo C. There, the
> documentation for each function included a "Portability" section, indicating
> how possible it would be to use the function on other systems. I'd like to
> suggest something similar be added to DJGPP. Each function could say whether
> it is ANSI, POSIX, available on Unix or other DOS compilers, unique to
> DJGPP, etc.
> Before everyone gets upset, I am willing to work on this myself. However, I
> don't have the ANSI or POSIX standards and don't have much experience with
> other DOS compilers, other than a very old version of Turbo C. Perhaps
> someone can tell me where I can get (by e-mail or FTP only, please)
> summaries or lists of functions defined by the standards (or even the
> standards themselves, though I gather they cost lots of money :( ).
> 
> I think this would be a significant aid to people trying to write portable
> programs with DJGPP.


First, this idea would benefit not just the djgpp community, but the
whole gcc community. I'm completely in favor of it. 

Second, after thinking about it for a little bit, it seems to me that
the only way to completely avoid issues of copyright theft and
plagiarism is to create a validation stub for functions in the djgpp 
distribution and then attempt to compile/run them under a selection 
of other compilers/platforms. 

Obviously there's not a problem with relying on public domain
references when they're relevent and available. And certain functions
can just be declared non-portable (djgpp only or gcc only) by fiat. But
to list something as portable will usually require relying on sources
which are not in the public domain--unless we test it ourselves. 

I don't think there's a problem with using copyrighted references as
*references*. For instance "both Borland and Microsoft say that function
foo() is portable among MS-DOS platforms" so we *tested* it on them and
we think it is too. But if we don't test, then I think it ultimately
just comes down to just copying someone else's work. 

As a counterpoint: There's the line of reasoning, popular among
students, that if you copy something from three different places it's
not plagiarism. I don't think that really holds up in court all that
well, though.

Ultimately, this information should be in the public domain.  But
ANSI/ISO do support themselves by selling standards--and Microsoft,
Borland et al by selling compilers with documentation.  I would not be
surprised if one of them got upset by our changing the legal status of
a piece of information they thought they owned.

-- 
Ned Ulbricht
mailto:nedu AT ee DOT washington DOT edu