Re: Help with [] for defined types

[David Bateman]

According to Paul Kienzle <address@hidden> (on 07/28/04):
> I've considered adding sparse support to mex by calling the
> functions "spfind" to get the matrix out of the octave value and
> "sparse" to put the matrix back into a sparse value.  You could
> do something similar in load/save for the matlab binary type.
> Until octave has a sparse type of its own it can include fake
> sparse m-files for these two functions which save the result
> in full matrices.  If octave-forge is present, then it will pick up
> Andy's version of the functions.  I guess you will also need
> "issparse" to detect whether the matrix type is a sparse matrix
> before storing it.

But the matlab load/save code is all concentrated in ls-mat*.cc and so
some supporting code will be needed there to load the matrix. Loading
them as full matrices and then converting them to sparse seems to be
against the wholw spirit of using sparse matrices in the first place..

Ultimately, I believe the ideal solution is to convert all of the octave
types so that there is no distinction between internal and external types.
This means that this matlab issue needs addressing, but also what to do
with

* The type registration, how to do it partially in advance while not 
  increasing the memory footprint of octave for those that don't use 
  all of the additional types

* How to either get rid of stuff like octave_value(NDArray) or allow it
  for user types

* Getting the internal types out of the liboctinterp.so library, including
  ls-mat.cc, xpow.cc, etc..

Cheers
David

-- 
David Bateman                                address@hidden
Motorola CRM                                 +33 1 69 35 48 04 (Ph) 
Parc Les Algorithmes, Commune de St Aubin    +33 1 69 35 77 01 (Fax) 
91193 Gif-Sur-Yvette FRANCE

The information contained in this communication has been classified as: 

[x] General Business Information 
[ ] Motorola Internal Use Only 
[ ] Motorola Confidential Proprietary