I agree. And we don't always agree - do we.
I'll add that I do think that the process should be twofold
When a patent gets filed - it should be thoroughly examined. Once examined it can be approved or declined. The it should be sent to an "auditor" who verifies/validates the approval or denial.
The first person should be paid based on the number of UNCONTESTED approvals. This way - they just don't approve willy-nilly. And this way - denying every claim doesn't get them a bonus either.
Heh. Based on memory, I think this might actually be the first time.
Not sure I agree with the details of your idea, mostly because it would end up spreading the patent load across fewer people in the initial approval phase (further swamping the examiners), but it's almost certainly better than what we've got now.
We should also bring back the working model requirement. It was dropped because the USPTO couldn't afford to store the models, but that can be solved by requiring that the *applicant* pay for the storage. Failure to keep the storage fees paid could simply result in abandoning the patent early.
Regardless of the particulars, there needs to be a better solution than 'let the courts sort it out'. That's supposed to be the last-ditch solution to a bad patent, not the normal one.