Well, I dug the CD out of the file cabinet and You're right-
it also says it includes Fortran 95 improvements!
As one of my friends pointed out recently, I'm getting old.
Gene Wheeler
Herff Professor of Structural Mechanics
Civil Engineering Department
The University of Memphis
[log in to unmask]
----- Original Message -----
From: "Richard E Maine" <[log in to unmask]>
To: <[log in to unmask]>
Sent: Friday, March 04, 2005 10:10 AM
Subject: Re: Has Fortran 90 been superseded by Fortran 95?
> On Mar 4, 2005, at 6:39 AM, Orville E. Wheeler wrote:
>
> > I'm tight fisted so I still have Microsoft Powerstation
90
>
> There is no such thing. I suspect you mean Powerstation
4.0.
>
> > It, of course, has gone through several vendors
>
> If you do mean Powerstation 4.0, then this is not true,
although it is
> a widely repeated error. The Powerstation 4.0 product was
just
> discontinued - period. Microsoft made a recommendation
about a compiler
> to use as a replacement, but that was a completely
different compiler
> from a completely different development line. No,
Microsoft did not
> sell the compiler to Digital; that statement has often
been repeated,
> but is simply false. The developers from Digital
(including some who I
> know personally - so I'm not just depending on some
"official company
> line") make this quite clear and explicit.
>
> > It's a perfectly good compiler with a good interface
...
>
> Our judgments differ there. I never could get any of my
code running on
> it. I worked around a large number of internal compiler
errors (even on
> simple f77-ish things) and eventually managed to make an
executable...
> that didn't work correctly. Somewhere around there, I gave
up on it, as
> other compilers for the platform were working fine for me.
>
> On fairly rare occasion, I still run into people using
that compiler.
> If they are asking for my help with it, I tend to tell
them that I've
> already invested more than enough of my time fighting with
it to pay
> for many compiler licenses, and that I decline to spend
more such time.
> It is pretty close to the bottom on my personal list of
compilers that
> I've had experience with. I could probably name one from a
decade or so
> ago that was lower on the list (but it was never a
significant
> commercial success and I haven't run into people using
it).
>
> On John's original question...
>
> I waffle about the same question myself. I stuck to f90
for a long
> time. I still do know some people who at least have f90
compilers
> installed. One of the users I support here has an old
version of PGI's
> f90 compiler that doesn't do f95. But then he also has
other compilers
> on the system. The PGI compiler was purchased for a
particular
> application (which needed both f90 and Cray pointers - a
combination
> that limited the available choices - the PGI compiler was
specifically
> listed as supported by the application) and is not the
"usual" compiler
> used on the system. (That's part of why we haven't
purchased upgrades
> for it - the one target application still works fine).
>
> I also recently ran into someone using V 1.0 of Sun's f90
compiler.
> This came up for reasons unrelated to f95, but I do know
that the
> compiler was f90-only (I forget the exact dates, but it
probably
> pre-dated the f95 standard even being released).
Fortunately, the user
> found that current compiler versions were also available
to him once he
> asked in more depth (V 1.0 would not have worked for my
application).
>
> But for new applications, I personally think I'd go ahead
with f95
> constructs now, even for code to be widely distributed. I
do still
> consider it a close call; I'd have given a different
answer not to
> awfully long ago.
>
> --
> Richard Maine | Good judgment comes from
experience;
> [log in to unmask] | experience comes from bad
judgment.
> | -- Mark Twain
|