Let's go troll hunting
The third thing that's grossly wrong refers mostly to the US, to the non-practising entities, or "trolls". Of course there's nothing wrong with the idea of Patent Assertion Entities at all. This is just Adam Smith's division and specialisation of labour all over again. It's just fine for the U. of Cambridge to set up a specialist unit staffed with patent lawyers to gain maximum value from the patents the university owns. Better, certainly, than each individual department or college trying to do the same. Yes, this PAE won't be producing anything, all it will ever do is negotiate with and sue people but that's still a worthwhile activity.
But some of the mroe litigious PAEs are a rather different beast. These are the people who spray writs around and rely on the fact that it costs more to defend against such a suit than the amount being demanded in fees.
In US civil litigation, barring extreme circumstances, every party to a lawsuit pays their own legal costs. So even if you won against a PAE, you'd still have to pay your own millions in lawyers' ransoms. The PAE can just sit there watching you do so and fold before it has to spend any more than that initial $500. Asymmetric costs, that's what we've got going on here. The solution is to make the PAE liable for the costs of fighting for licensing fees for users. Switch the US system to loser pays.
You're the loser, you can pay...
People who are really asserting a real patent won't be put off here: they will be confident of winning their case and thus happy to risk the money knowing they will get it back during the victory lap. People who are confident of seeing off a "troll" will happily rack up the defence costs knowing that they'll get it back when they win (and for PAEs that are shell companies it is possible to insist they lodge likely fees with the court before the hearing).
The people who won't go around gaily issuing writs are the trolls: for they're the people who know they have a weak case and whom with every writ will be opening themselves up to millions in potential fees as and when they lose.
We've even evidence that this system works. Commercial law in England and Wales works on loser pays. We do not have patent trolls in this American sense in England and Wales. We have PAEs asserting real patents, yes, but not the $500 trolls.
To conclude, I wish I could offer more and better solutions. Make patent issuance better, use more copyright and less patent law and make sure that losers pay all the legal fees. I'm afraid that's about all I've got.
For, as mentioned right at the beginning of this two-part discussion, we have a real economic problem here with public goods and as the first great law of economics states: there are no solutions, there are only ever trade-offs.
We want to maximise original invention but also derivative invention: which means that we cannot have perfect protection for the former for fear of choking off the latter. We also cannot allow free reign to derivative for fear of reducing original. That means that we'll never have a perfect patent system but we can, perhaps, have one rather better than the current one. ®