Location search and software patents

Posted on June 13, 2007


According to Wired, Jason Galanis, founder of London-based Geomas, has filed a patent infringement case against Verizon for their Superpages.com location search (story here). For those of you who aren’t immediately clear on what “location search” is, it’s any search you run looking for restaurants or hospitals within a certain distance of your zip code. And, according to the Wired article, over 100 companies cite the Geomas patent as “prior artwork” in their patent applications.

Geomas produces nothing except patent applications, and as such they sound a lot like a patent troll – someone who is a parasite on legitimate businesses and exists to make money via litigation off obvious ideas that they lacked the wherewithal to make honest money off of. Patent trolls are right up there with spammers as far as I’m concerned, and I’m all for updating federal law so it’s a LOT harder to bring spurious lawsuits based on bogus patents.

This brings up a related issue that I’ve thought about some over the years – software patents. At present, software is patentable, meaning you can claim unique use of a way of doing things in software for the period of the patent (currently 20 years), and you are permitted to charge others for using that software method you invented. As an example, if you come up with a new way of performing a data sort, you can patent it and keep anyone else from doing it that way for 20 years.

The problem is that it’s very, very hard to claim that an “invention” in software is non-trivial. Developing the very first lossless compression scheme probably qualifies as a non-trivial effort if only because the idea that data could be compressed without loss is a pretty radical idea. But once that initial compression scheme was developed, the vast majority of compression schemes that come afterward are direct derivatives of the first and thus don’t really qualify as original, non-obvious work. And so, according to my understanding of the requirements for a patent, shouldn’t qualify for patenting.

The problem is that, in today’s software patenting paradigm, every single new compression scheme could win a new patent even if the changes were so minor as to be completely obvious to anyone in the field. To use a real example, Amazon.com has patented their 1-click shopping method. Basically, if you have entered all your payment and shipping data into Amazon.com, you can use the 1-click button on the site and order a book or CD. Other on-line companies may use 2-click solutions (as Barnes & Noble apparently has), not a true 1-click purchasing option. But can we really expect that, given the data that most shopping carts are abandoned before their items are purchased, someone else wouldn’t have come up with this idea? In my opinion, the process of simplifying a means to purchase stuff online is obvious, and so even if it was an Amazon.com innovation that they spent years and millions of dollars on, it still didn’t deserve patenting.

And that’s really what this is about. I couldn’t care less if Amazon.com’s 1-click patent threatens online commerce or not. I care that patents represent truly innovative thinking, and by their nature, “small inventions” (as Jeff Bezos called it in his discussion with Tim O’Reilly) don’t represent major innovations and so shouldn’t be patented. At all.

And this is the case with the vast majority of software patents, including the location search patented by Geneos. Software development is inherently an incremental process, and patenting every single obvious incremental improvement just bogs down the already overworked patent office and stifles further innovation. And given that the purpose of patents is to enhance further innovation through revealing innovations publically, any patent application that stifles innovation should be flat out rejected.

Posted in: Uncategorized