Article View: gnu.misc.discuss
Article #1332Re: Fortran Is Not Dead, Unfortunately
From: Alan Braggins
Date: Tue, 17 Nov 1992 10:53
Date: Tue, 17 Nov 1992 10:53
37 lines
2339 bytes
2339 bytes
>>>>> On Mon, 19 Oct 1992 21:29:47 GMT, tim@maths.tcd.ie (Timothy Murphy) said:
> hards4@evans.ee.adfa.oz.au (Brad Hards) writes:
>> P.S.: At MHC they're teaching intro-CompSci and data structures in
>> ANSI C now... but I just recently discovered that at least one
>> instructor is busily trying to turn C into Pascal, by doing such
>> horrible things as using "const" to declare constants instead of using
>> the pre-processor. Ugh!
> Are you sure this was in imitation of Pascal?
> There are obvious advantages in using const --
> for one thing, the compiler can check the type.
And you can have structure constants, and they show up in the debugger,
and you can pass the integral ones to Fortran by address (see topic title).
But they aren't compile time constants, so can't be used for dimensioning
arrays, or have constant folding done. Enums can, and can also be typechecked.
I'd like to see
extern const int foo = 5;
be legal, meaning "The space for foo is allocated elsewhere, but it is safe
to assume it will be 5". Having a definition which is not initialised to
the same value would be an error.
> Come on, admit it, the instructor was right ...
Depends whether he taught which should be used where and why, or just presented
a rule to be followed blindly...
--
Alan Braggins, alanb@sdl.mdcbbs.com, abraggins@cix.compulink.co.uk
Shape Data - A division of EDS-Scicon Limited. Cambridge, UK +44-223-316673
"Any technology distinguishable from magic is insufficiently advanced."
"My employer does not necessarily share my views - but I'm working on it."
Technical Details
Message-ID:
<ALANB.92Nov17115348@catalina.sdl.mdcbbs.com>Path:
arc1-ca.pugleaf.net!sparky!uunet!mdcbbs!swordfish!alanbReferences:
<FRIEDMAN.92Oct16200656@nutrimat.gnu.ai.mit.edu><STDKAP1.92Oct17132457@oedipus .spc.uchicago.edu><FRIEDMAN.92Oct18171824@nutrimat.gnu.ai.mit.edu><BwCBBn.8Dq@ mtholyoke.edu> <HARDS4.92Oct19105906@evans.ee.adfa.oz.au><1992Oct19.212947.20744@maths.tcd.ie>Show Full Headers
Complete Article Headers
Path: arc1-ca.pugleaf.net!sparky!uunet!mdcbbs!swordfish!alanb Newsgroups: gnu.misc.discuss Subject: Re: Fortran Is Not Dead, Unfortunately Message-ID: <ALANB.92Nov17115348@catalina.sdl.mdcbbs.com> From: alanb@sdl.mdcbbs.com (Alan Braggins) Date: 17 Nov 92 10:53:47 GMT References: <FRIEDMAN.92Oct16200656@nutrimat.gnu.ai.mit.edu><STDKAP1.92Oct17132457@oedipus .spc.uchicago.edu><FRIEDMAN.92Oct18171824@nutrimat.gnu.ai.mit.edu><BwCBBn.8Dq@ mtholyoke.edu> <HARDS4.92Oct19105906@evans.ee.adfa.oz.au><1992Oct19.212947.20744@maths.tcd.ie> Organization: Shape Data (Division of EDS), Cambridge UK Nntp-Posting-Host: 134.244.153.110 In-reply-to: tim@maths.tcd.ie's message of Mon, 19 Oct 1992 21:29:47 GMT Lines: 37