Software patents pose dozens of problems, but some are fixable and some are not. If we focus too much on the fixable problems, politicians and judges will think that the whole problem is fixable.
Splitting between core and non-core problems
Take these two problems for example:
- Cost of getting patents and maintaining them - individuals and SMEs can't afford this
- Harm to standards and compatibility - I can't use Mpeg video because of the MPEG LA patent thicket
If we focus on the first problem, a politician can say "Ok, the new price for patents is €1". We would be getting a solution to our problem, but the real problem wouldn't be solved.
If we focus on the second, a politician can either solve the problem by abolishing software patents (complete win) or by creating an exception so that compatibility code is never considered to be capable of patent infringement (partial but big win)
How non-core problems are still valuable
The suggestion above (about reducing the price of patents to €1) could work in theory. There's also the problem that patent applications take too long, which could theoretically be solved by doubling the number of patent examiners. And there's the problem of silly patents, which could theoretically be solved by raising the standards. And there's the problem of...
The point is that software patents raise dozens of implementation problems (i.e. non-core problems). All might be theoretically solvable but it would mean throwing out the patent system and inventing a new one - which means either reforming the entire patent system to suit software, or developing a second patent system.
What are the core problems?