Current Situation
Today's patent law also punishes an inventor, when (s)he re-invents something that has been patented before (accidentally or knowingly, that's not important). This easily hurts us programmers because there is no chance that I know all those crazy patents out there. So by just using 10% of my brain cells, I easily come up with a solution such as the "double click" and bang – I broke about 100 patents at once without knowing it. (I am over-exaggerating here but the basic point is valid).
(Side-note: Oh, and did I mention that we just went snow shoe hiking and our snow shoes had about 8 patent numbers printed on them?!? Assuming an average of 15 pages per patent, that makes 120 pages description of my snow shoe! What are those guys writing about? )
My Proposal
My proposal aims at avoiding those (arguably) ridiculous patents that just protect brainless "inventions" but really hurt everybody else and kill innovation. And it works like this:
The invention effort of an idea is the effort (e.g., in dollars or man-hours) that were used to make a given invention. The complexity is the effort that is needed to implement the solution, given the patent document (and any other available information). Then the novelty is essentially the worth divided by the complexity:
novelty = invention effort / complexity
(If the resulting novelty isn't high enough, a patent wouldn't be accepted in the first place.)
This means, that an invention is the more novel (and therefore the more valuable) the higher the effort was that it took to invent it and the smaller the complexity to simply reimplement it. So putting billions into a new drug that once it has been discovered is almost trivial to reproduce has a very high novelty. Amazon's one-click isn't novel at all because it's totally stupid in the first place. And something that was super expensive to invent but the whole thing is so super complex anyway, that's a huge engineering effort but not such a great patent. (arguably)
Patent Disputes
Now, if there is a dispute about a patent the court checks the effort that both parties put into the invention (the complexity of re-creation should be the same, obviously, since it's about the same thing). If both efforts are comparable, then that's OK and the patent hasn't been infringed. Note, that here it doesn't really matter if the second party "stole" the invention or not. It spent the same money and therefore it's its own "fault" (see paragraph below for licensing issues)
If the potentially infringing party spent considerably less on the invention, then experts have to decide. They will get the invention documentation of both parties (see Invention Documentation below) and from those they have to decide if one party really "stole" the invention or if the other party just claimed a much higher invention effort, either because they are slower or because they wanted the invention to be more valuable than it really is. If the experts find an infringement, it's clear. If they don't then the first patent's novelty will be downgraded (with possible effects on the license prices).
Licensing
Licenses play a slightly different role in this system. If a party sees an idea, it can decide to re-invent it or to license it. Since re-invention is likely to require the same invention effort than the original (unless the invention isn't all that great and it can be done faster), licensing the invention makes a lot of sense if the license price compared to the invention effort is reasonable. But that's just a free market mechanism. So the license helps you to save money, actually.
If the invention isn't really all that great (such as Amazon's "one click" and the Google "map-reduce") but the other company claims a huge invention effort, then just re-invent it and document the effort. The other invention will then just be downgraded. The same mechanism holds when somebody accidentally re-invents something (s)he didn't know of.
Documentation Requirements
In this scheme, both parties must show "documentation" about their invention. Otherwise it's hard to judge what happened. (Note, that we assume that no party fakes documentation. That's an entirely different issue handled by different laws.)
But I claim that that's OK, because if I am a little garage-based inventor, I am in the "inventor business" and know about the documentation requirements. If I am a programmer who sells stuff, I am also a business and have many other requirements anyway. (Also, version-control system logs are already a very good documentation in that case, which comes for free). And big corporations document every little thingie anyway. Private people, on the other hand, are unlikely to invent something that requires a huge effort, then become rich with that, but don't have any documentation such as notes or other witnesses.
Patent Expiration
The time before a patent expires should be also proportional to its novelty. So the higher the effort was that you had to invest (and the more trivial it is for others to steal your idea instead of reinventing it) the longer you can keep your patent. So Amazon's one-click would have expired about 2.3 nanoseconds after it was filed.
Revocation of Unused Patents
It could also make sense to restrict the time one can own a patent to the fact that the business really uses it. So I cannot simply collect patents to threat others if I am not "in the business" (a.k.a. Patent Trolls). However, this whole issue is much more difficult because what if I just don't use the patent right now, but I might use it next year? Also, since I already spent the money it might be fair that I can license it to others even though I don't use it myself? So my proposal doesn't say anything about revoking unused patents even though that might be a good field for further ideas.
