Re: [PyOpenGL-Users] PyOpenGL, Numpy and 3.0

[Mike C. F. <mcf...@vr...>]

Rodney Stephenson wrote:
> This is more of a general question: where/how does Numpy fit into 
> PyOpenGl. There is the numpymodule and a few other references to numpy 
> around the place. Given that there is a lot to do to get Numpy to 2.6, 
> let alone 3.0, what is the long term strategy here?
>   
One of the big advantages of the PyOpenGL 3.x architecture is that it 
has the ability to use run-time pluggable data formats.  It currently 
can use:

    * Numpy arrays
    * ctypes pointers
    * ctypes arrays
    * strings
    * lists
    * tuples
    * VBO objects (written in Python in the module OpenGL/arrays/vbo.py )

So that it should be able to use whatever is available on the platform 
where it's running.  There's even a legacy Numeric format handler, 
though honestly I haven't got any setup for testing that it remains valid.

That said, a 2.6 port is likely to happen relatively soon (though I'm 
personally not really motivated to do it before numpy ports, as all of 
my own code is numpy-using).  The 2.6 port should be relatively trivial 
and we could likely be waiting for numpy on the other side without much 
effort.

3.x porting is going to take a while unless someone else wants to handle 
(and most importantly, maintain) it.  That's again just an interest 
thing.  I have close to 0 interest in Python 3.0 as a development 
platform, and the amount of work involved in porting to it easily swamps 
that level of interest.  I'm more likely to spend any extra time I find 
trying to get PyPy to work nicely with PyOpenGL code to produce (fast) 
executables, but realistically I don't often get that kind of extra 
time.  Maybe 3.0 will become more compelling at some point and I'll find 
my interest-level de-swamped, but I don't see a lot of movement that way.

HTH,
Mike

-- 
________________________________________________
  Mike C. Fletcher
  Designer, VR Plumber, Coder
  http://www.vrplumber.com
  http://blog.vrplumber.com