I'd say actively use or pro-actively license to someone else. If a company can't show that their patents are generating revenue somehow other than lawsuits, the holder should be forced to give the clarinet to a kid who is gonna use it.
So there are a lot of problems that arise either way. Fist, the way it works now, you file for a patent and get approved, you own it for the generic term that applies to everyone, special considerations depending on the type of patent not withstanding. This leaves us with patent trolls, yes.
The alternative: grant a patent but if it not used within some scope of time in some way it becomes public domain or otherwise no longer considered a person's intellectual property. So you go to some company and say "I've solved your problem and I want to license my solution to you" and they tell you either you're going to sell it to them cheap or they'll just wait you out. You either take it or lose the patent in short order and now they can use it free. Now the only people that even own intellectual property are the big guys, and if you don't sell they let you lay fallow until your ownership becomes forfeit. You either sell to them or lose the patent.
I hate patent trolls too, but this alternative is worse. I personally don't like intellectual property in general, I think the best solution is to just not have it. Trade secrets are one thing, making it illegal for someone to reverse engineer something they bought or draw pictures they saw elsewhere in public is a bit absurd.
That is no the whole point, and certainly hasn't been going back to the 18th century at least. It is routine for patent holders to license their patents to manufacturers for production. That would not be possible if inventors had to make the product for commercial sale by themselves. It would render novel inventions by independent inventors untenable. Only big corporations with manufacturing resources would ever be able to patent anything.
Take James Watt for example, Matthew Boulton could just have told Watt to take a hike and started manufacturing improved steam engines by himself. Watt couldn't have afforded to make steam engines independently, and why would any investor bother supporting him with capital if they could just set up production without him too?
The problem is with trivial patents, not patents in general.
Your facts are completely false, and in fact it's easier to bring things to market than it's ever been.
Watts held back progress significantly and steam engine development was hampered by his and others patents until the expired. The steam engine would have been built to pump water out of mines whether patents existed or not.
Split a company's research from its production and sales departments into two organisations, and one licenses its patents to the other.
You have transformed a structure where one entities invents and produces to a market structure with one where those are separate roles.
You like the former and abhor the later. But from the outside, this change is absolutely meaningless. It's no different than any other make-or-buy decision.
(Note that "sells a license" is exactly the same as "patent trolling".)
All these feelings you people have are based on the idea of "superficial" patents that aren't "real innovation". Which is, indeed, a problem. But it has nothing to do with the above. Assume some actually useful patent, and these objections fall apart.
It is not meaningless. The split structure is actually better (if it is genuinely spun off, rather than created as a tax evasion mechanism) because it can licence the technology to multiple entities, which spurs competition among them.
A research program that only produces patents for patent trolls seems like a useless research program.