[vtk-developers] vtkUniformGrid

[Berk Geveci]

I didn't implement vtkUniformGrid as a subclass of vtkImageData
because I wanted to prevent filters that cannot support blanking (for
example all image filters) to work with vtkUniformGrid. When I get
around to it (soon I hope), I will revisit this issue. One way of
removing the code duplication is to move the common code to another
class and subclass both vtkImageData and vtkUniformGrid from that.
That still leaves us with the fact that only dataset filters work on
vtkUniformGrid and that there are no "fast paths" for it. Also, people
will eventually want to apply image filters to AMR datasets. At that
point, it might become necessary to break up that dataset so that
there are no overlaps (ChomboVis does this for everything). I am not a
fan of doing this by default because 1. it changes the structure of
the AMR dataset 2. it requires keeping two copies (and uses twice the
memory) if access to the original dataset is necessary.
I am open to suggestions, btw :-)
-Berk


On Fri, 10 Dec 2004 14:21:44 +0100, John Biddiscombe (CSCS)
<biddisco at cscs.ch> wrote:
> Hello,
> 
> I'm implementing a reader for vtkUniformGrid data and looking at the
> code I wonder...
> 
> Is there any reason why vtkUniformGrid wasn't implemented as a subclass
> of vtkImageData - it seems like a great deal of code has been duplicated
> (speed impact of virtual calls maybe?)
> 
> thanks
> 
> JB
> 
> --
> John Biddiscombe,                            email:biddisco @ cscs.ch
> http://www.cscs.ch/about/BJohn.php
> CSCS, Swiss National Supercomputing Centre  | Tel:  +41 (91) 610.82.07
> Via Cantonale, 6928 Manno, Switzerland      | Fax:  +41 (91) 610.82.82
> 
> _______________________________________________
> vtk-developers mailing list
> vtk-developers at vtk.org
> http://www.vtk.org/mailman/listinfo/vtk-developers
>