W
Walter Harley
- Jan 1, 1970
- 0
Paul Hovnanian P.E. said:[...]
That's how understood them as well. But that's where software patents
will run into trouble. Software, a.k.a. source code, given sufficiently
intelligent compilers and other tools, may be nothing more than the
expression of an idea. The compiler takes it from there and produces the
executable.
Ironically, although I think a lot of software patents are utterly bogus, I
think that some software might be about the best example of why patents are
worthwhile and useful.
Google probably has patents on their search algorithms. (I don't know, I
don't compete with them.) They could easily keep them as trade secrets, and
nobody would be able to reverse-engineer them. Since people have thought
very deeply about search algorithms and published the results (read Knuth,
for instance), any new algorithm is just about guaranteed to *not* be
obvious. Whatever Google is doing is worth knowing about, and could be
helpful for many folks, even folks not directly competing. But if they
reveal it, they lose their competitive edge. Perfect case for a patent:
they disclose, academic researchers can leap-frog, competitors can license,
net gain for the world.
IF their patent actually can be implemented. Which I know is the legal
requirement, but IME there can be many man-years between knowing an
algorithm and being able to successfully implement it in a commercial
software product.
And IF their algorithm actually is novel, instead of being something that
they found in Knuth and figured out how to apply to the web. I think that's
what the problem with a lot of software patents is.