[Insight-developers] Fixing itk::PathToImageFilter (again)

[Zachary Pincus]

Hi folks,

This is a ping about the status of the changes I recommended a couple  
of months ago for the itk::PathToImageFilter class. These changes are  
necessary to make it into a usable base class.

It was suggested that I put the modified class files up on the  
Insight Journal -- not as a journal submission per se, but just as a  
place to put a few files so ITK devs could look at the changes and  
see if they were useful. I did so, and have heard nothing since.

To recap, the current itk::PathToImageFilter is extremely  
problematic: it purports to be a base class, but actually implements  
filter functionality, and does so in such a way as to make it very  
difficult to subclass the filter.

My proposed fixes include refactoring this class into two, and  
deprecating the inappropriate functionality in PathToImageFilter. And  
yes, the changes scrupulously follow the deprecation policy. (Because  
of the deprecations, I was hoping to get this into ITK 2.6, so that  
the nasty stuff would be rooted out *before* people started relying  
on that functionality, but oh well.)

Here's the IJ link (I wrote up a brief PDF describing the changes in  
more detail, too):
http://www.insight-journal.org/view_publication.php?&pubid=74

Now that the push for ITK 2.6 is over, could we look at getting these  
changes into ITK? Remember, these are maintenance changes to make  
existing functionality work, not new functionality. It's only on the  
IJ because I was told to put the files there.

Zach