[Insight-developers] Rigid3DTransform

[Bill Lorensen]

Luis,

As part of our Slicer4/ITK4 experiment, there are a number of compile
errors because Slicer4(3) instantiate Rigid3dTransform.

Recall that we removes the itkNewMacro for the class in response to
bug 8320: http://public.kitware.com/Bug/view.php?id=8320

The corresponding gerrit topic is here:
http://review.source.kitware.com/#change,504

I believe that Slicer4(3)'s use of Rigid3DTransform is valid.
Slicer4(3) have saved transforms that are part of the test suite. Note
that in your comments on the bug tracker you state:
"It doesn't hurt to have a valid implementation of Get/SetParamters at
the level of the Rigid3DTransform. However, this class simply
shouldn't be used as a full fledged transform. (e.g. for example as
part of an image registration framework). "

Note the word "shouldn't". Well, Slicer4(3) did.

I suggest that we restore the itkNewMacro and resolve bug 8320 another
way. I think it is hard to justify such an API change in light of a
valid use in Slicer4(3).

Bill