Compilers - Fortran without its legacy.

Author

Fortran without its legacy.

Toon Moene

2004-03-27, 12:29 am

Gentle Compiler Folk,

Recently, http://www.fortranstatement.com popped up. It petitions the
ANSI J3 committee to retire FORTRAN (I don't think we're allowed to do
so - however, we *do* have to vote for the continued existence of
previous Fortran Standards occasionally).

A more interesting question is (implicit in this site's "discussion"):

Would Fortran exist without its legacy code ?

I.e., would the current Fortran language, as defined by its Standard
(loosely called "Fortran 95") exist on its own merits ?

Please, discuss.
--
Toon Moene - mailto:toon@moene.indiv.nluug.nl - phoneto: +31 346 214290
Saturnushof 14, 3738 XG Maartensdijk, The Netherlands
Maintainer, GNU Fortran 77: http://gcc.gnu.org/onlinedocs/g77_news.html
GNU Fortran 95: http://gcc.gnu.org/fortran/ (under construction)
[I think the first response on the response page pretty much sums up why
Fortran is still an important language, and wishing otherwise won't make
it go away. It's still the only language for seriously high performance
numerical software. -John]

Joachim Durchholz

2004-03-27, 12:29 am

Our esteemed moderator wrote:
> [I think the first response on the response page pretty much sums up
> why Fortran is still an important language, and wishing otherwise
> won't make it go away. It's still the only language for seriously
> high performance numerical software. -John]

I think that "informed response" is a bit narrow-sighted - he's
considering only those languages that are currently in the mainstream.
With that, he's probably right (though I don't know Fortran well enough
to be sure).

If you look at the future, things look better. Better languages than C,
C++, or Java exist: less than 100 pages for a language reference, clean
syntax, clean semantics. ("Clean" meaning "free of legacy quirks and
subtle points".)
Of course, it's a question whether these languages will gain enough
momentum to actually replace Fortran; that could happen anytime between
next year and never...

Regards,
Jo
--
Currently looking for a new job.
[But none of those address what has always been Fortran's uniqe design
point, the ability to generate blindingly fast numeric code. C isn't
terrible, but unless you carefully declare all of your pointers to be
restricted, it doesn't come close. Fortran has ugly semantics,
particularly in respect to storage management in large part because
they make it possible to do optimizations you can't do if you have to
preserve cleaner semantics. -John]

Sponsored Links