[vtk-developers] Scope of VTK and it's potential as a common research language

[David Doria]

On Mon, Jan 25, 2010 at 9:33 PM, pat marion <pat.marion at kitware.com> wrote:

> Adopting a math library, one thing to consider is vtk's wrapped api.  If
> new data classes are used as public method arguments or return values they
> may or may not translate easily to vtk's wrapped languages, java, python,
> tcl.
>
> Devil's advocate- If something like Eigen is so easy to incorporate into a
> project, then why does vtk need to include it, why not leave it up to the
> application developer to choose their own math library?
>
> Pat
>

First, I don't think anyone said it is "so easy". It wouldn't really help
anything if the user (developer) used their own library because the point is
that the math library has to interface nicely with VTK. Using your own math
library calls for a conversion before and after every vtk call, which gets
annoying in a big hurry.

My thought is that you should be able to do things like:
ChosenMathLibraryMatrix a = vtkFoo();

rather than the messy arrays that happen currently.

I also really think it would let us clean up a lot of quite messy internal
mathematics code.

David
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://public.kitware.com/pipermail/vtk-developers/attachments/20100125/cfe1d907/attachment.html>