This is about the triple-damages rule, which says that if you're found to infringe on a patent willingly then the damages are increased, while if you didn't know about it, they aren't. This, combined with a large number of very-low-quality ambiguous patents and the fact that the patent office's screening only creates a legal presumption of validity but not a practical one, means that everyone is strongly incentivized to avoid ever looking at anything the patent office publishes. Which makes things even worse, because the problem is hidden and the patent office can't draw on public knowledge to shoot down bad patents.
"the triple-damages rule, which says that if you're found to infringe on a patent willingly then the damages are increased, while if you didn't know about it, they aren't."
That's the triple damages rule as invented from whole cloth by the Circuit Court of Appeals for the Federal Circuit (CAFC), the patent court.
The actual law is that "the court may increase the damages up to three times the amount found or assessed" (35 USC §284). There is no special rule about knowing infringement or willful infringement.
The CAFC has been limiting triple damages because having no standards in a free for all of triple damages based only on judges' feelings would be terrible. So the CAFC picked out one egregious behavior by rich infringers -- willful infringement -- and limited the triple damages to that offense. It was a reasonable compromise but has no support in the text of the law.
The Supreme Court looks askance at such compromises outside the statute and may decide to overturn this one. That will lead to more shopping for district judges inclined to cheat and abuse the defendants in patent lawsuits now that there will be no limits on tripling. And that points to -- you guessed it -- the Eastern District of Texas.
There is another case before the CAFC -- not the Supremes -- TC Heartland scheduled for oral argument on 11 March, where the CAFC could simply abolish Eastern Texas patent litigation with the stroke of a pen. The law (28 USC §1400) says that you can be sued over patents only in your home jurisdiction and not in East Texas (since tech companies don't start up in Texarkana). The CAFC abolished that law by simply choosing to contradict it a couple decades ago and created the troll jurisdictions. But now the CAFC is considering bringing it back and obeying the law as written.
It's interesting to see the CAFC in both cases being the sane ones. For the most part historically the CAFC has been loony and corrupt in promoting maximum damage to innovation for the benefit of patent lawyers with no respect to the law or reason. Some of the new Obama appointees and the resignation of the Chief Judge Rader under federal investigation for corruption last year may have reduced the lunacy temporarily.
I appreciate your explanation, though I can also appreciate why the rule was introduced. It would seem the consequences of it were entirely unforeseen.
A devil's advocate argument: should the rule be tossed, might there not suddenly be a bottleneck imposed on new projects where legal teams would now have to research whether a particular feature duplicates an existing patent? I'd say it's an unwinnable game since doing away with software patents might also harm any commercial motivation to produce a service with even a hint of transparency into how that service does what it does.
I almost feel like the Treble damages rule is the least-bad of all of them, but that's just a gut feeling. Any armchair-lawyers (or actual ones!) want to weigh in? I'm not anywhere near qualified to actually express an opinion on this.
> doing away with software patents might also harm any commercial motivation to produce a service with even a hint of transparency into how that service does what it does
Do you have examples of FOSS or other projects benefiting from companies revealing information through patent filing? To rephrase, what exactly do we lose out if patents didn't exist?
Even if there is some marginal value in the knowledge sharing that happens in patent filings, not allowing software patents would be a much bigger benefit. To support this point, we can look at the prevalence and utility of Open Source software projects; which work without patent protection.
> To rephrase, what exactly do we lose out if patents didn't exist?
There is a lot of empirical evidence showing patents incentivize investment of resources into innovative efforts, such as R&D expenses in large corporations and VC investment in startups. Of course, it's hard to quantify how much innovation this really drives because it's very hard to measure "innovation" itself. The drawbacks of patents include, amongst other things, the potential to reduce follow-on innovation. Again, this is very hard to measure at any meaningful scale.
> To support this point, we can look at the prevalence and utility of Open Source software projects; which work without patent protection.
Open source projects are driven by people willing to donate their efforts for free. (Well, it's not really free, is it? They are, to a great extent, sponsored directly or indirectly by organizations that typically extract value from something other than the software itself, such as, maybe data.) Do you expect that business model to work for everybody?
Not to mention that the vast majority of open source projects actually don't do anything innovative other than just re-implementing known technologies in a different context. There can be a lot of minor, follow-on innovation in open-source projects, but in my experience, the big advances come from other sources, such as academia, startups or corporate R&D labs.
>There is a lot of empirical evidence showing patents incentivize investment of resources into innovative efforts
Do you have any sources of this empirical data? In many cases especially with software the opposite is the case now. Legacy patent holders are able to obtain such broadly vague patents and then sit on them and do nothing but try to sue anyone who tries to actually innovate.
To my knowledge, empirical evidence that incumbent patent holders in the software industry are suppressing innovators is mixed. This paper from 2006 claims no ill-effects have been felt: http://papers.ssrn.com/sol3/papers.cfm?abstract_id=926204
On the other hand, this 2006 study finds that product areas with patent thickets see fewer new entrants: http://www.nber.org/papers/w12563 -- Interestingly, this paper also finds that startups with patents of their own are much more likely to enter a patent-infested product area.
> To rephrase, what exactly do we lose out if patents didn't exist?
If you're talking about mundane things like a slider on a phone lock-screen - than obviously nearly nothing.
On the otherhand if you are talking about a result of research that might have taken 5-10 years - than patent protection is a big deal and very few private companies would be incentivized to engage in such research without a patent system.
Yes, that is, indeed, the point if patents. That's what the OP asked: what 5-10y software research, which we currently benefit from, would not have happened were it not for patents?
In medicine, this is extremely easy to answer. Hence why the vast majority of software patents opponents are fine with those.
> That's what the OP asked: what 5-10y software research, which we currently benefit from, would not have happened were it not for patents?
First of all that is explicitly not what OP asked - he asked:
>> To rephrase, what exactly do we lose out if patents didn't exist?
> In medicine, this is extremely easy to answer. Hence why the vast majority of software patents opponents are fine with those.
Then it seems you've answered your own question? Obviously the patent being in software as opposed to being in medicine doesn't make any distinction -- since there is no logical way to qualify that criteria into a rule that can be unequivocally applied; that the patent system is greatly beneficial in other major areas of research (and supposing you admit the net benefit to be greater than the net harm) makes it inevitable in software - whether you like it or not.
> A patent that can be infringed using only computer software and a prior art computer.
When is that the case? When is it not? Are medical devices that run software excluded? Are all kinds (read most) imaginable consumer electronics also excluded (they also have some software)?
You are arguing for a policy as if in complete obliviousness that such a policy must be applied in the real world (and said companies will have teams of lawyers to find loopholes, not a random HN poster).
Can you name some specific thing where you think this would be unclear?
> Are medical devices that run software excluded? Are all kinds (read most) imaginable consumer electronics also excluded (they also have some software)?
Do they consist of only computer software and a prior art computer?
It doesn't say the device has to contain no software. If you come up with some new light sensor for a camera which is only useful in combination with a particular algorithm to interpret the output then you would be able to patent the sensor (hardware) but not the algorithm (software). If it's just a new algorithm to interpret data from existing sensors then it's a software patent.
I don't see what part of this you find so perplexing. It is a software patent if running software is all that is required to infringe it.
It is difficult, as that article points out. But to suggest the non-patentability of software is "just hard to take seriously" is, well, just hard to take seriously, when the largest economy in the world is doing it.
The current state of affairs as described in the brief (accurately, in my experience) is catastrophically bad. Even if you completely abolished software patents I don't think it could be worse than this.
I would like to see a study that compare the commercial incentive vs the deincentives cause by the chilling effects from patents. It should be a standard cost-benefit analyzes with no emotional/political bias that could give a final scientific answer if patents are beneficial for software or not.
There are tons of studies around patents and the economics thereof, but you will not find such a convenient answer to your question. The reason is that it's very hard to measure the costs and benefits to quantify these incentives and dis-incentives. Furthermore the benefits and costs of patent systems are often orthogonal. If you want to get an idea of the depth of the field, look at all the studies cited by this paper, and look at all the various methods, metrics and datasets that have been considered so far: http://papers.ssrn.com/sol3/papers.cfm?abstract_id=2139260
