> 1) IP companies must notify any infringing company within 3 months of the infringing product's release. IP companies can't wait until something like the iPhone blows up to sue.
So no stealth startups then?
> 2) If a company wants to maintain a patent, it must either be a commercial product or be willing to license it for a reasonable fee. Some IP companies refuse to license at first, in order to drive up the price.
Who defines reasonable? I understand that the market can determine this, but the market can't then go to the patent office and say "Yeah, this patent should not be allowed."
> 3) Every 3 years that you have the patent, you must show progress towards commerciality (commerciality remains to be defined in this thought experiment).
Show progress to whom? The patent office? How do you define progress? Then we are back at square one: the patent office holds all the keys.
> 4) The patents for ideas, strategies, etc. should either be banned outright, or the lengths made much, much shorter. There's no reason that Amazon should still have a patent over "one-click" in 2010.
Agreed. Patents on software == patents on mathematical concepts. By the same token patents on strategies.
Here is a solution that can be implemented as an algorithm:
1. The inventor registers a patent and is required to set a licencing fee.
2. If within a year nobody licenses, the idea, the licensing fee is dropped to at least half the original or lower if the inventor chooses to do so.
3. This repeats until the fee is less than $1. Then the patent expires and everything is public domain.
4. If a patent is licensed from day one at the asker's price, it still has the usual time limit on expiration.
5. As long as a new licensee was signed on in the past year, the inventor may raise the licensing fee once for that year for any new licensees.
This is not a perfect system and obviously can still be gamed in various ways. But a system like this one could use markets to determine what an idea is really worth and for how long.