[vtk-developers] Correct handling of information propagation

Ken Martin ken.martin at kitware.com
Tue Jan 18 11:09:12 EST 2005


In your request information you should look to see if there is a bounds key
in the input's information, if so then use it. Currently I don't think
anything actually provides bounds during request information but that is the
correct place for it and it will start showing up more as
filters/source/reader start providing it. If bounds are not provided then
you must go with some default extent.

> In the old pipeline, we had vtkStructuredPointsSource, which was
> interesting in that it didn't necessarily compute the correct
> information (regarding extents).
> I would like to implement a filter which takes points as input, but
> produces an imagedata as output similar to ShepardMethod, except that
> vtkShepardMethod uses a fixed extent, which is set by the user (so the
> information output can simply set its whole extent without issue)
> 
> Suppose my Image that is produced as output, depends upon the extents of
> the polygons passed in (the bigger the polys, then perhaps the larger
> the wholeextent of the output).
> 
> In the UpdateInformation method, I used to do an Update on the input (or
> inputs), get the relevent info from them, then compute a whole extent,
> then make that visible to the output...
> 
> Can anyone tell me the correct way to handle this in the new pipeline
> (inside RequestInformation). Is there an example of this I can look at
> anywhere?

I do not think anything does what you describe yet but clearly the other
point to ImageData classes are a rough starting point.

Ken





More information about the vtk-developers mailing list