> I think the real answer is compulsory licensing of patents.

The trouble with compulsory licensing is that it doesn't solve one of the most fundamental problems with software patents, which is how do you set the price? An iPhone is probably covered by some tens of thousand of patents. How much of the value of the iPhone should be attributed to each patent? There are too many patents to spend any real resources evaluating each individual patent; the lawyers fees would surpass the royalties. On the other hand, not looking closely creates a powerful market incentive (which is what has been happening) to file for thousands of low-quality patents, and thwarts the patent incentive to make significant inventions because high-quality patents requiring significant R&D are awarded no additional royalties over the crap.

On top of that, circumstances change quickly, and how do you respond to that? If you establish royalties for specific patents one year and the next year the new model comes out which is covered by all the old patents plus thousands of new ones, do the old patent holders get less money for the same patents until the royalties from a patent fall below the cost of filing for one, or does the price of a phone keep going up until they each cost $100,000 and everyone is forced to go back to pen and paper? And what do we do about free software -- percentage of zero and no money for the patent holder or fixed fee and free software ceases to exist?

Compulsory licensing can't fix software patents. They probably can't be fixed. The patent system was never designed to apply to products with zero marginal cost and a combinatorial explosion of patent liability.

We could apply a kind of value-added tax to every good sold that might possibly be encumbered by patents, say 5%, put that money in a pool overseen by a special court, then let patent-holders file with the court to claim a portion of the money based on what products their patent applies to and how important it is relative to the other patents that apply. The other patent-holders can then file challenges if they think they're getting short shrift. The filing fees should float such that it covers the costs of the entire system.

I'm assuming you're only wanting to implement this for software. (It obviously doesn't work for e.g. drugs, the amount that patents contribute to drug prices is far in excess of 5%.)

But I don't see that it works for software either because you still have the same problem. 5% of a billion dollars is 50 million dollars. A billion dollars on the high side for a single product; the Galaxy Note doesn't infringe all the same patents as the Galaxy S3. 10,000 patent holders fighting over 50 million dollars means they'll get an average of $5000. You can't hire a lawyer to fight over your share and expect to have any left over when "your share" is $5000. Which means that a) the lawyers would get all the money, and b) world+dog would still have the incentive to file for a million low-quality patents because having a patent gets you a share and nobody can afford to look too closely at whether you deserve one.

For drugs, you could allow the product maker to opt into having a greater percentage going to patent holders. (Edit: Sorry, I see what you're saying. You couldn't keep another drug company from copying the drug and only paying 5% to the inventor. Fair point.)

I doubt that a typical smartphone actually infringes on 10,000 patents held by outside entities, or that 10,000 separate entities each hold one of those patents apiece, but taking that for granted, if filing fees are set so as to cover the entire costs of the court and system, and the system has a fairly inelastic capacity (supply), the increased demand for services would cause the filing fees to rise. Presumably an equilibrium would be reached where only those with large and realistic claims would bother to file, whereas those with low quality patents would be priced out.

Ideally the system could always afford to look closely, because it's self-funding. It wouldn't necessarily need to be adversarial, either -- it could be more like arbitration, where you file, paying the court to evaluate your claim, and the court makes a decision. No need for lawyers, unless you need help putting your claim together.

> I doubt that a typical smartphone actually infringes on 10,000 patents held by outside entities, or that 10,000 separate entities each hold one of those patents apiece

10,000 is probably in the ballpark. People have no recognition of how incredibly complicated phones are today. They're obviously not owned by that many separate entities, but that doesn't really help much. It takes about the same amount of time to determine whether a device infringes the claims of a patent regardless of whether the same patent holder also holds other patents. The "savings" come when you stop looking at the claims and start paying royalties based on the quantity of patents the patent holder claims you're infringing, but that's where the incentive to get low-quality patents comes from.

> if filing fees are set so as to cover the entire costs of the court and system, and the system has a fairly inelastic capacity (supply), the increased demand for services would cause the filing fees to rise. Presumably an equilibrium would be reached where only those with large and realistic claims would bother to file, whereas those with low quality patents would be priced out.

The counterargument you're going to get there is that you'll price small inventors out of the patent system.

> Ideally the system could always afford to look closely, because it's self-funding. It wouldn't necessarily need to be adversarial, either -- it could be more like arbitration, where you file, paying the court to evaluate your claim, and the court makes a decision. No need for lawyers, unless you need help putting your claim together.

But you're still not solving the pricing issue. What objective metric is an arbitrator supposed to use to evaluate the value of a patent?

> (It obviously doesn't work for e.g. drugs, the amount that patents contribute to drug prices is far in excess of 5%.)

For one thing, it's arguable that the amount that patents contribute to drug prices, and to drugmakers' revenues, is too damn high. Secondly, the basic scheme is obviously compatible with setting a higher percentage of sale price in the case of drugs. Alternatively, about 10 or 15 years ago some lawyers proposed a kind of mandatory Dutch-auction scheme for drug patents, which seemed interesting.

> world+dog would still have the incentive to file for a million low-quality patents because having a patent gets you a share and nobody can afford to look too closely at whether you deserve one.

The nice thing about this is that it would be somewhat of a self-resolving problem, since it would give the holders of non-trivial patents reason to lobby for high standards for patentability.

> The nice thing about this is that it would be somewhat of a self-resolving problem, since it would give the holders of non-trivial patents reason to lobby for high standards for patentability.

It would also give the holders of trivial patents (and patent lawyers who want more business) reason to lobby for low standards for patentability.

No, the trivial-patent holders already have that incentive. In any case, in a direct lobbying/political conflict between IBM, GSK, Dow Chemical and whoever versus pure-play trolls, even big ones like IV, my money is on the former. The way in which a fixed levy would or could facilitate trivial-patent holders is by putting in place a revenue-collecting process (making claims on the 5% pool) which might (or might not) be smoother and more reliable than the process by which these gentlemen make their money nowadays.

One nice thing about such a system is that would give the holders of non-trivial patents a direct interest in demanding a high bar for patentability, because other people's trivial or obvious patents would be directly taking a slice out of their revenues.

The first step is to stop using the word "compulsory licensing". Compulsory licensing sounds like some kind of crazily authoritarian evil government regulation; if I didn't know what it actually is, and there were no triggers that convinced me the idea is worth taking an hour of my time to research, I would imagine I would instinctively oppose it. The language implicitly accepts that patents are a natural right, and that what it is proposing is an infringement on that right. In reality, of course, it's a limitation on a government-granted monopoly. Nobody is being "forced to give a license"; it's actually the derivative creator who is no longer forced to face unbounded cost when using something which was patented.

Maybe it could be called something like "licensing by right" or "right-to-use", so casting it as a defence of something positive, rather than the imposition of something negative.

"Right-to-use" sounds like something people could get behind.

Congress needs to pass "Right-to-use" patent reform to prevent Saudi oil companies from buying up patents that increase gas mileage and preventing US companies from developing them into products.

I don't think this has ever actually happened, but current patent law would allow it.

Isn't the word "statutory" used in similar contexts? So there would be a "statutory" maximum for, say, individual and total royalty costs in producing some product.

You are hereby compelled to furnish the audience with an alternative less-authoritarian sounding substitution. I compel you to not "compel against" with merely strict negatives, and insist that you provide tenable positive alternates.

I would say rename the whole thing to something like "intellectual use tax".

I think a better way that might treat copyright and patents equally well is to make it a continuous act of will and treasure to keep things from the public. Right now, the default state of a copyrighted work like you mention is essentially automatic eternal restriction. Patents, once issued (for little relative cost) remain in force for more than a decade with no effort required for upkeep.

Restricting "intellectual property" from public use should require exponentially increasing effort/cost as time goes on. I think this would allow the system to automatically discover the value of the restriction, and help redistribute the cost of maintaining that restriction back to the public which provide it with the force of law to begin with.

tl;dr : Each year you hold on to a copyright or patent, you should be required to pay a great deal more to the governing body that grants you that monopoly to keep it in force.

I don't think this would be very good for small creators. Existing copyright giants could use their current wealth to overwhelm the market, making it impossible for an independent to get any exposure. Then they just wait for the independent to run out of copyright support money, create a derivative work, and never have to spend anything on original writing.

It would depend on the protection being nearly free in the short term but becoming very expensive in the longer term. This would allow small upstarts to find out if a market existed and then, if true, capitalize on it and use the profits to fund continued protection as far (and only as far) as the market would bear.

If they failed, the patents could not then be acquired by trolls for pennies where they would then pollute the ecosystem for decades.

As far as copyrights go, this would neatly solve the orphaned work problem with the added benefit of Disney having a serious think about paying a billion dollar tax to the public in order to keep "Steamboat Willie" in the "Disney Vault".

I still think such a system would need a way to prevent large players from waiting out smaller creators and getting free source material that they currently have to pay big bucks for (e.g. consider J.K. Rowling).