This week’s software patent comes courtesy of my friend Rob LeGrand, a grad student at Washington University who previously worked for IBM. As he recounts on his blog,
I and two other guys filed for four software patents while working for IBM in Austin about five years ago but I never got around to checking whether any of them had made it all the way through the process. Well, we checked, and two of the four are now patents! You can find them at the Patent and Trademark Office’s database; the patent numbers are 6,778,837 and 6,898,628. (Checking the patent application database shows that the other two filings never made it even that far; presumably, IBM’s lawyers didn’t think they were worth the expense. Anyway, IBM gave us inventors little bonuses for all four filings.)
I suppose they’ll look good on my r©sum©, and I guess I’m proud to have my name on something that was deemed important and original enough to be patented, but I’m also a little ashamed. I mean, what is a patent, really? It’s an artificial restriction on the commercial use of an idea. Not an actual piece of property, but an idea. I have no desire to restrict anyone from implementing our ideas. And while most patents cover a specific implementation of a new idea, complete with detailed diagrams of its inner workings, software patents usually just describe what the invention does, not how it does it. Read about our patents and you’ll find prose more conceptual than concrete, written in dense lawyer-speak designed to cover as many potential products as possible.
I think Rob’s take on patents is a little more radical than mine. I don’t have a problem with patenting an idea in principle, as long as it’s novel and non-obvious as required by patent law. But Rob seems to feel that software patents are particularly problematic, given that, as he says, they tend to describe what the patents do, rather than how they do it. (since the “how” is described by source code, which is already protected by copyright law).
As Rob says, his patent doesn’t appear to be an exception. Basically, the patent covers the concept of using one’s GPS location as an input for an authentication process. Although implementing such an idea could be somewhat challenging, the patent doesn’t go into a lot of detail about how such an implementation might work. And while using GPS as an authentication method is a clever idea, it seems likely that it would be a fairly obvious to someone who had an actual application for it, rather than simply writing about it in the abstract.