(prompted by Greg’s post about OSDL Conference Key Thoughts)

I believe in patenting combinations of software and hardware or human interaction which are genuinely innovative – eg: token ring networking, Lego Mindstorms, the wheelie mouse, that new thing which allows people to see by putting sensors on their tongue, performance-related algorithms such as compression, or security-related encryption algorithms.

Pure software patents, on the other hand… why? Most software is developed as a means to solve a problem. Given the same problem and two decent teams of people to solve it, it seems likely that the two teams will come up with similar designs. Patenting those designs is basically patenting the problem, not the solution. It makes no sense to me. If the problem is already in the public domain, then perhaps the solution shouldn’t be patentable.

So what’s the solution to the ‘problem patent’ problem?

There’s not a lot which can be done with regard to existing patents without some major changes to patent laws. However, for future patents… how about a site where software designs, no matter how strange or silly, can be published? I’m thinking of some kind of library in which any thoughts, architectures, bright sparks, plans, dreams, ideas etc. can be collated against future patents, so that anyone trying to patent things will find themselves losing the court battles on the easily provable grounds that the idea is already in the public domain. This doesn’t have to be restricted to software, necessarily.

For instance:

“How to communicate with Generation Ships and other deep-space bodies, in the absence of faster-than-light travel:
Use quantum entanglement for instant information transfer.”

Now anyone trying to patent the idea of using quantum entanglement to overcome the speed-of-light restriction in deep space communication will find me sitting in the witness box in court. I have now published this idea, and unless someone has already applied for the patent (which wouldn’t surprise me) you are welcome to use it free-of-charge. I’m pretty sure someone else has already thought of Generation Ships.

“How to transfer documents between two different formats, eg: Word 2003 and LateX:
Use a learning AI with pattern-matching to work out the conversion algorithms from other document fragments in both formats, then apply it.”

And so on.

Ideas could be voted on for viability, with the best ideas published to the open-source community for implementation, and the “silly” ideas given negative points… but kept around, just in case.

Carlos reckons that the c2 wiki might be a good place to start messin’.