Hi,

mainly for curiosity, I set up the following webpage to see which
Fortran 2003 and Fortran 2008 are most requested and which are already
in use. Additionally, I ask some questions about used parallelization
techniques and compilers. (Sorry, the number of questions is quite
long.)

http://physik.fu-berlin.de/~tburnus/F2003-polling/index.php

The input might also be used by compiler writers (esp. gfortran) to
help deciding which feature should be implemented first.

Tobias

Tobias Burnus wrote:

Wow.

That is an excellent page. I won't be able to get to it this week, but I will study the
questions on the train rides to/from work for action next week.

I have to say that it is a nice distillation of the new features.

cheers,

paulv

I think ,of course,most wanted is oo.

On Jun 9, 11:21 am, Tobias Burnus wrote:

The responses are interesting, someone thinks enumerations are "very
important". I fear that some responses will be made without adequate
understanding of the features (certainly, I don't understand all the
features adequately).

"Tobias Burnus" wrote in message
news:b43df39c-63ea-4708-9fd4-d6b1edb30c51@26g2000hsk.googlegroups.com...

Interesting survey. I think there should be an opportunity for
respondents to pass on any question. Certainly for my part I see
many features I haven't had the chance to assess, and for the
object-oriented stuff the worth of any feature or combination of
features requires a little practice to see whether it adds
capabilities that I would consider useful.

"James Van Buskirk" comcast.net> wrote in message
news:DNmdnYy0tpey59DVnZ2dnUVZ_qLinZ2d@comcast.com...

Looking further at bugzilla, I see that these PRs are marked as
fixed. Does that mean I am going to have to come up with further
examples to stress gfortran's handling of specification
expressions?

James Van Buskirk wrote:
(snip)

Wow, I hadn't noticed yet that they were all missing
from the previous standards.

I was checking not so long ago which calculators
can do asin(2) (with complex mode active), but I hadn't
checked recently for Fortran.

A complex data type was added to C only in C99 along with,
as far as I can tell, complex versions of the inverse,
hyperbolic, and inverse hyperbolic functions.
(Many added for real arguments only in C99.)

It seems that the relatively current PL/I systems have
more than Fortran 2003, but not all of those listed above.

-- glen

On 9 juin, 19:22, hes selex sohu.com> wrote:

Not me ! OO is just "useful" no more ...

The ISO_C_BINDING on the contrary is very important for my job. This
is also the case for OpenMP, MPI. For coarrays I wait ...

On 9 juin, 18:21, Tobias Burnus wrote:

Just a remark Tobias : is it possible to explain a little bit more the
new features, for instance via a "help" button ? I had sometimes
doubts about the true meaning of the subject.

fj wrote:

Me too, but I note that it is keyed to John Reid's paper (see the link
near the top). Still, I had filled out a bit before I found myself
stumped by what something could mean. I might have answered one or two
based on a wrong guess.

I'm betting that some other people did simillarly. I mean 2 people
thought that "complex constants" was "very important?" While I agree
that complex constants are important, I find it hard to imagine the
particular aspect of complex constants in question as making anyone's
"very important" list. It makes me suspicious that some people didn't
actually check beyond reading the somewhat nondescriptive heading (yes,
I see that you copied it from John's paper) "complex constants".

But then perhaps, hard as I find it to imagine, someone really does
think that particular enhancement to complex constant syntax is "very
important" instead of just a minor nicety.