Usually, I use this spot to rant about something, or someone that's riled me up in some way. My lack of discussion on software patents doesn't mean I agree with them, it's just that everyone else has been doing it. I couldn't see why I should do so and be seen as just another blogger with nothing better to do with my time.
Someone that has plenty of things to do with their time is Simon Phipps. He was brought into Sun to work up their Open Source strategy, and was instrumental in getting Java released under the GPL. And he still has enough energy left to be a great speaker. I had the pleasure of meeting and hearing him talk last night, where he introduced his ideas for software patent form. Let's face it - software patents are going to happen, so we might as well be constructive about it and guide it in the right direction, so it can be implemented in a manner with which we are agreeable.
Now, I agree with the stance and the ideas he presents, but still have some niggling doubts. Firstly, the method to minmize and/or eliminate patent trolls is wonderful. If they had to pay back the licensing costs, they'd be less quick to slap injunctions on products using "their" technology. However, if the company has a lifespan similar to those in the dotcom boom, they'd be bankrupt before you got your money back. Ugh! This problem exists because a company is not run by the company; it's run by people. And if the people running the company can use that money to finance their (private) house, car, and lifestyle, then they will. Most people would sell out their companies and careers for a few million in the bank.
Making patents invalid once they're part of an ISO standard is also a good idea. Nuff said.
Simon also suggests making software patents last for a maximum of 5 years, due to the pace of technology. I think he's wrong, and it should be limited to 3. Alas, both are arbitrary numbers, and I've seen software patentable ideas get superceded and become irrelvant in less time (e.g. computer graphics) and some which I expect would never die, given the option.
Requiring source code to be filed with the patent is also a good idea as it stops the wooliness and arbitrary language usually used in patents. Anyone that's filed a patent will usually tell you they don't understand the wording of their own patent, or to what specifically it refers.
Finally, Simon suggests perjury for patents filers that don't discover prior art. This is a little harsh in my opinion, especially for the small developer as they will need to start filing patents for the sole purpose of having a portfolio that they can use to defend themselves against the big boys. Or perhaps the idea is to highlight the fact that the entire Microsoft Windows development team are guilty in their failure to check prior art, and should be locked up.
Well, it's a means to an ends...