A disturbing piece has popped up, via Slashdot: a guy named Greg Aharonian is suing to remove copyright protection from computer programs. He maintains that instead, IP protection should be offered by patents.

This makes sense in principle, but is a terrible idea in practice. Getting a copyright is easy -- if you've written something, you have at least some copyright protection. Getting a patent is difficult -- it's expensive, it takes a long time, and you have to demonstrate that your invention is new and non-obvious. A lot of redundant and obvious software techniques have managed to get patents, but this is due to the USPTO's almost complete inability to deal with software patents competently, not because patents are well-suited to software. If the patent system -- properly applied -- was the sole source of IP protection for software, not only would the system be shifted to favor large companies with the resources to file lots of patents, but the vast majority of software authors would be producing work that could be legally appropriated by whoever cared to copy it. I'm not someone to equate piracy with the end of civilization, but if businesses weren't compelled to buy software licenses it does seem likely that the industry and our economy would suffer.

Aharonian has a point, though. Despite some valiant efforts by computer scientists to pawn it off as speech, computer code really doesn't express ideas in the way that copyright law seems intended to protect. Still, just because our IP law is ill-equipped to handle the information economy doesn't mean we should dismantle the jerry-rigged protections it does afford.

The problem with software patent is that there are very few software innovations that could not be developed given a hundred geeks at a hundred PCs and adequate supplies of cola. Those innovations that do rise to this level tend to be highly abstract algorithms that are largely mathematical in nature -- a faster way to sort data, for example. But these discoveries approach the level of fundamental mathematical truths, and allowing them to be patented would be as problematic as requiring engineers to buy licenses for the quadratic equation.

And in any case, the development of such algorithms is rare. Software patents are more commonly awared for inventions like Amazon's infamous "one-click" patent --which actually covers basically all business conducted over the internet (Amazon has thus far declined to enforce it). Patents liek this are for trivial ideas that clearly could easily have been thought up with by someone else first (and frequently have -- the USPTO is very bad at finding prior art for software patents). They're obvious -- Amazon has not done the public a big favor by developing this knowledge, and consequently they don't deserve special protection. Yet they have it -- and unlike patents most other fields, software patents are a potentially huge pain in the ass: if someone patents a better mousetrap, that doesn't interfere with my existing mousetrap business. But if Amazon starts enforcing their patent for performing transactions over the internet, I have to be on my toes to avoid infringing upon it as I write any number of glaringly obvious applications.

There's not much genius involved in software development, but there is frequently a huge amount of work. That work deserves to be protected. Just because computer code is philosophically more similar to an invention than it is to speech doesn't mean that the patent system is the most appropriate way to protect it -- in practice, protecting code as an invention would not serve the industry or the public's best interests. Mr. Aharonian is trying to make a principled point while ignoring the damage that doing so would incur. It's a good thing that he seems likely to fail.