On Sun, Nov 15, 2009 at 5:33 AM, Gabor Szabo <[email protected]> wrote:

> I had an interesting chat with someone who pointed out that moving people
> and companies away from Matlab will be extremely difficult due to the huge
> amount of legacy code.
>
> So I wonder if anyone has ever experienced with parsing Matlab code and
> then running it?
>

The open-source alternative to Matlab is Octave.  Most Matlab scripts should
be able to run unchanged under Octave, and vice-versa.  I don't believe that
PDL has a direct interface to Octave, but there is an Inline::Octave
module.  If somebody is looking for an open-source alternative to Matlab,
they should probably first look into Octave, and then slowly migrate
functionality over to PDL.

I have never tried Inline::Octave.  It looks like it hasn't been developed
for a few years and it talks to Octave by opening a process and
communicating with it over pipes.  The author states that he tried to
interface with liboctave but it got too complicated.  Although the
compuations on Octave's side should be as fast as native Octave code, there
will probably be a bottleneck when sending or receiving large quantities of
data, as well as a large memory overhead when that data is turned into a
string and sent over the pipe.  (At least, that's the problem I've run into
with my Asymptote pipe.)

This points to an excellent opportunity with somebody on this list with
spare time to kill.  :)  If we could create some sort of Matlab quasi-parser
using liboctave and/or IDL quasi-parser using
GDL<http://gnudatalanguage.sourceforge.net/>,
migration to PDL would become much simpler for lots of potential PDL users.
Probably the most Perlish way to implement this would be to create some sort
of Inline::PDL::Octave and/or Inline::PDL::GDL modules.

David
_______________________________________________
Perldl mailing list
[email protected]
http://mailman.jach.hawaii.edu/mailman/listinfo/perldl

Reply via email to