Is your software patentable?

I’ve always been curious about how “impressive” or “radically new” does a software creation have to be in order to be patentable.
Once you’ve had success implementing (licensing) a software application you’ve developed for a handful of customers the “obvious” next legal step is to obtain a patent for it, although some of us wouldn’t even think about publishing a software piece to a customer without first obtaining a patent from the PTO, the procedure is usually obscure to software developers.

Well, in this search, I found a very worrying blog post by professor John F. Duffy called “The Death of Google patents” that summarizes the current PTO position in regards to “patentable software”.

To my huge surprise I found that issues like: transforming data from one source to another in a fast or automated process, combining technologies on an innovative way to solve a problem or quantifying intangibles to generate other intangibles would have to be tied to a “particular machine” to be patentable. And by particular the PTO means a “special” computer not any kind of computer like yours or mine.

So unless your creation is not required to run on a special “machine” your software could easily be left unpatented under the current PTO’s Board reasoning in the Wasynczuk case.

The link to professor Duffy’s blog:
http://www.patentlyo.com/patent/2008/07/the-death-of-go.html

Some other useful links in this matter:
http://www.patentlyo.com/lawjournal/files/cole.pdf
http://www.law.cornell.edu/uscode/html/uscode17/usc_sec_17_00000101----000-.html

Comments

[Karl T. Muth]

From what I understand, software is protected in a variety of ways, including trade secrets (taking reasonable care to not make public how the software is developed, or particular intracacies of it that would be difficult to observe or reverse-engineer) and copyright (actually copyrighting the program itself, or crucial bits of code). Patent is part, but not all, of the IP protection puzzle in this space, from what I've read...