[pp.int.general] software patents

[Richard Stallman]

    - It is not difficult to design a mechanical invention that only depends on 
    some geometry.

That's no counterexample to my argument.  I think you have focused
on a part of the argument, and not on the overall point.

    - Software design for control enginering and signal processing applications 
    is often non-trivial.

Signal processing interacts with physics; it is typically part of a
specific physical system.  So it does not really contradict what I
said.

But even if there are some specific areas of software which are
exceptions to the general conclusion, that is of small importance
compared with the rest of software, so the conclusion is still valid.
We're talking about public policy, not mathematics.
A small exception does not make a policy argument invalid.

I'm not going to defend non-software patents.
If society is ready to get rid of them, I won't stand in the way.
But it is very useful to be able to show why software patents pose
a special problem.