[Insight-developers] Re: ICP (Iterative Closest Point) Pate nt

[Peter Cech]

On Tue, Sep 28, 2004 at 19:16:18 -0400, Lorensen, William E (Research) wrote:
> I don't think we can distribute patented code without the permission of the
> patenjt holder. For vtk, GE gave us that permission.

That's the question for a lawyer. I do not know the law in US, so I have
no idea how much control has patent holder over implementation of the
patented algorithm. They can demand royalties, so I would at least ask
them about royalties model (there is a good chance that they won't ask
royalties as long as you are not making a profit of it).

Alternative solution is to just risk it. I encountered a few arguments
why open source projects are unlikely being sued by patent holders:

- As long as you don't make any money of it, you are not interesting
  enough to be sued.

- Suing open source developers or educational/research institution is
  likely to damage image of any company pretty badly.

As you are already distributing at implementation of least one patented
algorithm (marching cubes), one more will probably make little
difference. I'm all for clearly marking algorithms as patented (and even
having and option to compile/install ITK without patented algorithms;
even better for each algorithm separately) and mark all depending
algorithm implementations as depending on patented stuff.

Regards,

Peter Cech