Re: Nd cell array

[Philipp.Batchelor]

David,
great, thanks a lot for the answers! it's exactly what I needed.
I didn't know that 3D arrays were available in the newest version, I'll 
download and install asap then.



Maybe I can give a bit of background on what I'm trying to do.
vtk (Visualisation Toolkit) http://www.vtk.org is a quite popular 
computer graphics C++ library, traditionally used with Tcl, Python, or 
recently Java, wrapping. I came across instructions for using the Java 
wrapping to load vtk classes into Matlab, (cf 
http://www.cns.bu.edu/~speech/VTK.php) which sounded like the 'best of 
both wolrds' for me. So the issue became, how does one load vtk into 
Matlab, and as I just bought a laptop and put Redhat 9 on it, which came 
with Octave, I the same question arises for Octave.

There are different possibilities:
1) Matlab can use a java virtual machine, hence can read any java class, 
hence without too much effort (actually it wa a lot, but wel...), one 
can load any vtk class into Matlab. As far as I can see, this solution 
doesn't extend yet to Octave. There are problems with memory management.
2) The same person who created the java solution was unhappy with it, so 
created a 'vtkmexifier' which I had adapted to Octave, but this is 
limited to vtk pipelines I think.

3) Writing directly C++ code which links to vtk, and loads into Octave, 
(this would be much harder in Matlab I suppose). I will try to write a 
volume renderer as a test.

Thanks again
Ph


David Bateman wrote:
> According to David Bateman <address@hidden> (on 03/04/04):
>
> >   for (int i=0; i < dv(0); i++)
> >     for (int j=0; j < dv(1); j++)
> >       for (int k=0; k < dv(2); k++)
> >         m (i, j, k) = ...;
>
>
> of course that should read
>
>     for (int i=0; i < dv.numel (); i++)
>       m(i) = ...;