Fortran - Fotran 2003 Compiler

Author

Fotran 2003 Compiler

Anthony Iannetti

2004-06-25, 7:43 pm

Dear Fortran Newsgroup:

Does anybody know who makes a Fortran 2003 compilant compiler? Yes, I
know the standard has not been ratified. I am looking for something to
play around with.

Thanks,
Tony

beliavsky@aol.com

2004-06-25, 7:43 pm

Anthony Iannetti <Anthony.C.Iannetti@nasa.gov> wrote:
>Dear Fortran Newsgroup:
>
>Does anybody know who makes a Fortran 2003 compilant compiler? Yes, I
>know the standard has not been ratified. I am looking for something to

>play around with.
>
>Thanks,
>Tony

No Fortran 2003 compiler is available yet, but I think NAG has the most F2003
features -- see http://www.nag.co.uk/nagware/NP/NP50_announcement.asp for
details. Lahey/Fujitsu Fortran for Windows .NET has added some object-oriented
features of F2003 -- see http://www.lahey.com/lf71/netwtpr1.htm .

----== Posted via Newsfeed.Com - Unlimited-Uncensored-Secure Usenet News==----
http://www.newsfeed.com The #1 Newsgroup Service in the World! >100,000 Newsgroups
---= 19 East/West-Coast Specialized Servers - Total Privacy via Encryption =---

Richard Maine

2004-06-25, 7:43 pm

Anthony Iannetti <Anthony.C.Iannetti@nasa.gov> writes:

> Does anybody know who makes a Fortran 2003 compilant compiler?

Yes, I know who. Nobody is who. Nor would I expect one for
some time, where the eact value of some is TBD. Compilers
with some f2003 features are one thing. Compilers that
come even close to compliance are another thing entirely.

As Beliavsky said, NAG has a few of the OO features...but not
enough to do much interesting with IMO. (They have
inheritance, but neither type-bound procedures nor procedure
pointers).

Several compilers have the TRs implemented; one could consider
those to be f2003 features.

C interop is probably coming soon to compilers near you. (Some
apparently have it now, though none that I've used).

If you are looking for specific features, ask about them.
If you are looking for compliance, you are way too early.

--
Richard Maine | Good judgment comes from experience;
email: my first.last at org.domain | experience comes from bad judgment.
org: nasa, domain: gov | -- Mark Twain

Anthony Iannetti

2004-06-25, 7:43 pm

Richard,

I am looking for object oriented features. Unfortunately, a few
usually is not enough to stop wrinting in F95 "Modular Programming". I
think there would be real benefits of programming in an object oriented
language, for what I do.

Thanks,
Tony

Richard Maine

2004-06-25, 7:43 pm

Anthony Iannetti <Anthony.C.Iannetti@nasa.gov> writes:

> I am looking for object oriented features.

Then NAG is probably as close as there is right now. I don't
find it close enough to do much OO stuff yet though, as I mentioned
before, so I wouldn't rush out and buy it for that right away.
Still, it wouldn't take *TOO* much more to start being useful
for that.

But then it isn't a bad f95 compiler at all, so if you want to
buy it as an f95 compiler and see what you might be able to do
with what OO stuff it has, that's reasonable.

--
Richard Maine
email: my last name at domain
domain: summertriangle dot net

William Clodius

2004-06-25, 7:43 pm

Anthony Iannetti <Anthony.C.Iannetti@nasa.gov> wrote in message news:<casvr1$ke6$2@sulawesi-fi.lerc.nasa.gov>...
> Dear Fortran Newsgroup:
>
> Does anybody know who makes a Fortran 2003 compilant compiler? Yes, I
> know the standard has not been ratified. I am looking for something to
> play around with.
>
> Thanks,
> Tony
Is there even an F95 "compliant" compiler available? I know of no
official definition of compliance for F90. The one language I know
of that has compliance measures, Ada, has a standard defined test
suite. There are commercial est suites available for F95, and I expect
at least one to be available soon after official publication of the
standard, but succh test suites usually test only the simples, most
easilly isolated parts of the language. Compilers are complex programs
and invariably have errors in them that cause problems with edge
cases. Also programming language standards are complex documents and
that invariably have ambiguities or errors that need to be resolved
before you can truly check compliance. There is an interpretation
process for resolving these ambiguities that takes several years. This
is not unique to Fortran, even now C++ compilers often have problems
with namespace issues which would not normally be considered edge
cases.

The F03 standard is a significant increase in the complexity of the
language comparable to F90 in its impact. F90 can therefore be used as
a very rough guide to what might be expected for F03. F90's draft was
distributed for review in mid-late 1990. After minor corrections
prompted by the review it was published as a standard in mid-91. One
or two month's later NAG released its compiler to K&R C, which was the
first attempt at a compliant compiler. Shortly thereafter a couple of
"F90" compilers were announced that translated to F77 + libraries.
These "F77" based compilers were nowhere near conformance. About a
year later, NAG published a revision that was significantly closer to
true conformance, and one vendor adopted their compiler as the basis
of their Windows compiler. In mid 93 Lahey announced their compiler,
then in late 93 several system vendors(Cray and IBM) announced their
compilers. By late 94 a large variety of compilers were available,
with NAG having the best compliance, but several others having
acceptable compliance.

There are differences now from F90. With a significant existing user
base, NAG has a reputation to maintain, a useful population for beta
testing, and a requirement for backwards compatibility. I would
therefore expect them to be slower about providing an official F03
release. Maybe their F03 compiler will be in Beta testing soon, but I
would be surprised to see a release of an official F03 compiler within
the next few months. On the other hand part of the delay for the
other vendors was due to several causes that no longer apply: a
failure to take F90 seriously until very late in the process and F77
compilers that were poorly structured for transformation into F90
compilers, partly because F77 was a small enough language that good
structure was not critical, and partly because they incorporated a
large number of ad hoc extensions whose interaction with F90 features
were problematic. The remaining vendors have structured their
compilers so that they can be more easily upgraded, they have taken
F03 more seriously, they have eliminated some of the more problematic
extensions with their F90 compilers, and they have more clearly
defined the remaining extensions so their interaction with F03
features are better understood. While I still expect them to lag NAG,
I expect the relative delay to be less than the 2+ years that occurred
with F90.

Sponsored Links