Re: [photran] Where to focus development efforts

[Michel DEVEL <michel.devel@xxxxxxxxxxxxxx>]

Hi folks,

What functionality is most conspicuously missing from Photran?
What existing features are sub-par?
What can we do to make Photran more appealing and approachable to new
    

First and above all for me : a working debugger on windows with tool chain for gfortran and g95 !

Also, I would appreciate a way to convert between Managed make to standard make project by freezing the files created by Managed make and letting the manual changes made by a user not be erased automatically. At the moment I can do that by creating a new standard make project and reloading all the files but it would be great to simply convert.

To my very humble opinion a single distribution for
installation of *everything* would be great.
Here *everything* should be selectable from a menu,
including (for Windows-people...):

- eclipse
- MinGW
- g95
- gfortran
- msys
- gdb
- any other utility that is necessary...
  

Full install version 3.0 was great to begin with !!!
Please do not forget a script to set environment variables for either Mingw or Cygwin
(However, as I use both, I sometimes have problems to set my PATH correctly !)

In case someone already has some of these, the install should suggest
updates where applicable.
A set of default settings should integrate all installed packages to a
working lot, without an immediate
need to fill out hundreds of forms for paths, keywords, libraries etc
before you can "go".
In this way someone with a good knowledge of FORTRAN, but with no
experience in Photran,
who wants to try out Photran, will not be disappointed by the large number
of packages
he/she has to find and install.
  

The problem is that it is sometimes quite tricky to mix latest versions of all these things (including PTP, CDT,...). You really have to keep information about which combinations of versions work. In this respect Cygwin is much more convenient to update that Mingw.

I think that a "bare" version would get sufficient attention. What I mean
is an IDE (integrated
development environment) that allows for the coding of "just fortran" in
the medieval
meaning, i.e. intended for the development of programs that you call from a
commandline and that return to
the commandline after having made some awfully difficult calculations
(justifying the use
of FORTRAN!), of which the output has been written to some files.
Without all the user interfacing, widgets, OOP and graphics that you can do
today (and which to my
opinion would automatically lead to a re-thinking of which programming
language to use!),
FORTRAN-coding in the classical sense can use the boost from a decent IDE,
especially from the (visual)
integration with a debugger, without a need for the user to learn all the
interfacing commands with that
debugger.
  

I entirely agree with that, all the more that I precisely only code and use those kinds of "console codes"

Maybe an (automatic?) "update" key that checks all the above packages for
newer versions?
  

Once again, I agree.

To summarize my opinion, I would say that improving the ease of use of current photran (improved Help, Tutorials, etc... are part of it) is much more important for me than adding new functionalities.
Photran is still not at  the level Visual Fortran for Windows was years ago.
As I teach numerical methods for physicists and chemists, I wish to use Photran for that and be able to transition gradually from Windows to Linux, but in practice we still use an old Visual fortran for that, just because of the ease of use of the debugger.
I understand that it may not be compatible with new PHD thesis subjects and much less interesting for people doing the job, but I really feel this focus on usability is necessary for the community of users.
Photran is a very nice piece of work but still not ready for production in my opinion.
The problem is that if you hire someone to the do the job, you have to give him/her money...

--

Sincerely yours,

Michel DEVEL