Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!uunet!mcsun!ukc!icdoc!dcw
From: dcw@doc.ic.ac.uk (Duncan C White)
Newsgroups: comp.lang.modula2
Subject: Re: re: SYSTEM
Message-ID: <1853@gould.doc.ic.ac.uk>
Date: 2 May 90 15:53:18 GMT
References: <4768.263B00CF@puddle.fidonet.org>
Reply-To: dcw@doc.ic.ac.uk (Duncan C White)
Organization: Dept. of Computing, Imperial College, London, UK.
Lines: 60

Hi all,

In article <4768.263B00CF@puddle.fidonet.org> Peter.M..Perchansky@f101.n273.z1.fidonet.org (Peter M. Perchansky) writes:
>Hello:
>
>    Wirth did not screw up in any fashion by not setting a standard for
>    libraries.  That is the job of "standard commitees."

That's not the whole story. If the initial designer of a language does a good
job (as Wirth did with Pascal, as Kernighan & Ritchie did with C) then after
defining the language proper, the designer should define a reasonable core set
of facilities - especially I/O, but also string handling, garbage collection
(if appropriate) and possibly a few other things - that users may expect to be
available with all compilers for that language.

This ensures at least a measure of basic portability of source code between
systems using a single language, during the decade or so that will typically
pass before any standards organisations get involved.

Some years later, a consensus may emerge that the whole language needs
re-examining, various undefined aspects needs defining, various non portable
assumptions need modifying in the light of experience, and that the "base
level" of portable facilities needs raising, and then a standards committee
gets formed, has many discussions and comes up (eventually) with a newly
"standardised" version of the language AND ITS LIBRARIES.

This seems like a very sensible way to manage change.
But it doesn't absolve the language designer from the responsibility of
providing a reasonable standard library at the beginning.

>    The C language did not have a "standard anything" until recently. 
>    Even though the ANSI standard for C has just come out, only 2 compilers
>    come close to 100% compliance.  Therefore, how portable is C???

Not in the limited sense of "an ANSI/ISO/BSI standard", but in reality, C has
had a well defined portable set of library routines with well understood (if
not defined!) semantics, for roughly the last 15 years.

Similarly, Pascal had an accepted and well understood IO system (read, write et
al) many years before there was any such thing as a "standard" Pascal.

But not Modula-2 - well, it has InOut et al, I suppose..

	Duncan

>--  
>uucp: uunet!m2xenix!puddle!273!101!Peter.M..Perchansky
>Internet: Peter.M..Perchansky@f101.n273.z1.fidonet.org

PS: I notice that no-one has as yet come up with a portable Modula-2 version
of that simple program I wrote in Pascal..

----------------------------------------------------------------------------
Duncan White,           |       "Once people start going about killing
Dept. Of Computing,     |       people, people have to take very special
Imperial College,       |       measures against certain people, don't
London SW7		|       people?  Even if it means people taking
England.		|	frightful risks"
			|		Lovejoy, "Gold from Gemini",
Email: dcw@doc.ic.ac.uk	|		written by Jonathan Gash