[Numpy-discussion] matlab -> python translation (original) (raw)
Stefan van der Walt stefan at sun.ac.za
Wed Jun 28 12:24:02 EDT 2006
- Previous message (by thread): [Numpy-discussion] fread codes versus numpy types
- Next message (by thread): [Numpy-discussion] indexing bug in numpy r2694
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Hi all,
I recently saw discussions on the list regarding Matlab/Octave to Python translation. I brought this under John Eaton's attention (he is the original author of Octave) -- below is his response.
Regards Stéfan
----- Forwarded message from "John W. Eaton" <jwe at bevo.che.wisc.edu> -----
From: "John W. Eaton" <jwe at bevo.che.wisc.edu>
On 21-Jun-2006, Stefan van der Walt wrote:
| I'd like to bring this thread under your attention, in case you want | to comment: || http://aspn.activestate.com/ASPN/Mail/Message/numpy-discussion/3174978
Would you please pass along the following comments?
Translating the syntax might not be too hard, but to have a really effective tool, you have to get all the details of the Matlab/Octave function calls the same as well. So would you do that by linking to Octave's run-time libraries as well? That could pobably be made to work, but it would probably drag in a lot more code that some poeple would expect when they just want to translate and run a relatively small number of lines of Matlab code.
Another semantic detail that would likely cause trouble is the (apparent) pass-by-value semantics of Matlab. How would you reconcile this with the mutable types of Python?
Finally, I would encourage anyone who wants to work on a Matlab/Octave to Python translator using Octave's parser and run-time libraries to work on this in a way that can be integrated with Octave. Please consider discuss your ideas about this project on the maintainers at octave.org mailing list.
Thanks,
jwe
----- End forwarded message -----
- Previous message (by thread): [Numpy-discussion] fread codes versus numpy types
- Next message (by thread): [Numpy-discussion] indexing bug in numpy r2694
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]