Scientific Libraries in Python

[Horatio Davis]

On 6 Dec 2001, Prabhu Ramachandran wrote:

> Umm, I'm not sure I understand.  Did you really mean that MayaVi would
> be able to eat the Mathematica and IDL graphics for breakfast?  I
> haven't used them much to actually comment on that but if they do
> choose to use MayaVi for their visualization then it isn't really
> possible to eat them for breakfast is it?

Hmm. Almost.

What makes MayaVi interesting to me is the combination of VTK and Python
to add value to the VTK. Mathematica and IDL may be able to use VTK, but
they won't be switching to Python to build on it. The best they could
hypothetically hope for is doing MayaVi-type stuff implemented in their
own languages, with VTK as the visualization library.  Which license
MayaVi has is irrelevant to whether this can happen.

Now posit some weird parallel universe where Mathematica tries to
incorporate Python and the VTK into itself (no, I can't think of a reason
for them to, either). In this _particular_ niche, MayaVi is far ahead of
any proprietary software, and will handily out-compete any such
hypothetical monstrosity. That's what I mean by "eaten for breakfast."

As for the question of whether MayaVi's versatility and graphical quality
would _in_general_ beat Mathematica or IDL, I haven't played with either.
much. Matlab, now, that I could give you a comparison with. Ask me again
after I finally get libVtkPython*.so to link correctly on Linux.

> Anyway thanks for the compliments.

I read this obscure paper somewhere on the Net once that reckoned ego
satisfaction was the major incentive for open-source hacking. (:

It's probably about time this thread moved off the general Python list to
some place more specialized. As soon as the scipy.org people unstick their
list server and let me successfully subscribe (hint, hint) I'll try there.

Cheers,

Horatio