[Numpy-discussion] F2PY future
Fri Apr 11 07:44:55 CDT 2008
I'm afraid I'm not much help answering your questions. But one thing
I've wondered about f2py is if it could be generalized into an f2***
tool. How intertwined is the analysis of the fortran with the
synthesis of the python? There are lots of languages that could
benefit from a fortran wrapper generator tool.
On Fri, Apr 11, 2008 at 8:07 PM, Pearu Peterson <email@example.com> wrote:
> I am in a process of writing a scientific paper about F2PY that will
> provide an automatic solution to the Python and Fortran connection
> problem. While writing it, I also need to decide what will be the future
> of F2PY. In particulary, I have the following main questions to which I
> am looking for suggestions:
> 1) where the future users of F2PY should find it,
> 2) how the users can get support (documentation, mailing lists, etc).
> 3) where to continue the development of F2PY.
> Currently, F2PY has three "home pages":
> 1) http://cens.ioc.ee/projects/f2py2e/ - this has old f2py. The old f2py
> is unique in that it covers Numeric and numarray support, but is
> not being developed anymore.
> 2) http://www.scipy.org/F2py - this covers the current f2py included
> in NumPy. f2py in numpy is rather stable and is being maintained. There
> is no plans to add new functionalities (like F90 derived type support)
> to the numpy f2py.
> 3) http://projects.scipy.org/scipy/numpy/wiki/G3F2PY - this is a
> wiki page for the third generation of f2py. It aims at adding full
> Fortran 90/../2003 support to the f2py tool, including F90 derived types
> as well as POINTER arguments. It should replace numpy f2py in future.
> Obviosly, the three "home pages" for f2py is too much, even when they
> cover three different code sets. So, now I am looking for to unify these
> places to one site that will cover all three code sets with software,
> documentation, and support.
> Currently I can think of the following options:
> Use Google Code. Pros: it provides necessary infrastructure to develop
> software projects and I am used to it. Cons: in my experience Google
> Code has been too many times broken (at least three times in half a
> year), though this may improve in future. Also, Google Code provides
> only SVN, no hg.
> Since f2py will be an important tool for numpy/scipy users,
> it would be natural to continue developing f2py under these projects.
> However, there are rumours of cleaning up scipy/numpy from extension
> generation tools and so in long term, f2py may need to look for another
> home. So, I wonder if a hosting could be provided for f2py? Say, in a
> form of f2py.scipy.org or www.f2py.org? I am rather ignorant about these
> matters, so any help will be appreciated.
> Numpy-discussion mailing list
More information about the Numpy-discussion