3 ms·
As a meta-item on this discussion, it appears that we have a disagreement of value systems. Leaving aside the nuances of measuring positive and negative aspect
by JoshTriplett 11y ago
As a meta-item on this discussion, it appears that we have a disagreement of value systems. Leaving aside the nuances of measuring positive and negative aspects, it's quite possible for two people to look at the same set of positives and negatives and value them differently, or even to classify them differently, counting some of the stated positives as negatives and vice versa. It seems fairly clear that that's happening here, as I'm assigning positive value to several things you either don't care about at all or assign negative value to, and you're assigning positive value to several things I either don't care about at all or assign negative value to. I suspect we'll have a hard time agreeing on any high-level conclusion if we don't even value the same things.
>> It's quite easy to show the ill effects thereof.
> I haven't found a study that does so convincingly.
I'd recommend "Sequential Innovation, Patents, and Imitation" (http://www.researchoninnovation.org/patent.pdf http://www.researchoninnovation.org/patent.pdf), among others.
There also exist sources of information other than studies. Pointers to specific cases suffice to demonstrate that harm exists and to show the nature of that harm.
My original point was not that the sum total of either the positive or negative effects could easily be quantified, but rather that the negative effects can more directly be observed in any given case, while the positive effects cannot. Evaluating the negative effects simply requires showing an example of an invention that would have been published or product that would have been (or was) created, but was blocked by a patent; plenty of direct evidence exists for that. Evaluating the positive effects requires not only identifying an invention/product, but determining whether the invention/product would still have been published/created in the hypothetical scenario in which patents did not exist. An invention/product that made money from patents does not count in favor of patents; it only counts if it would not have been published/created at all without patents.
> > I read the meta-study you linked to, and found it quite unmoving.
> Yes, that's why I pointed to the meta-study. I find it fairly balanced on the whole, identifying benefits as well as costs. The point was not to convince you either way, just to highlight that the situation is pretty nuanced. For instance, while classical assumptions (like disclosure) may not hold, there are a number of alternative ways patents encourage innovation (e.g incentivizing R&D investment, signalling, facilitating VC funding and entry, etc.)
I meant the description "unmoving" differently than you took it. I did not mean "neutral/inconclusive", though in many ways it was that. I mean that I saw absolutely nothing in it that gave me the slightest inclination to support a patent system (and plenty not to). The meta-study covered three points: 1) efficacy at promoting innovation (one genuine purpose of the patent system), for which it failed to provide any direct evidence (things that would not have happened otherwise) and could only fall back on economic claims (rewarding things that were done; see (3)); 2) efficacy at promoting publication (another genuine purpose of the patent system), which it rather conclusively refuted; and 3) economic incentives and monetary evaluation of patents, which as a non-goal of the patent system (only a means), is irrelevant when evaluating whether a patent system should exist.
I care about code published, ideas created, ideas made practical as products, and other things that create value. I don't care at all about things that just move value around, and neither does the stated rationale for the patent system. The economic arguments made or mentioned in the meta-study only really talk about moving money around, but don't seem to provide evidence to support any claim that any value is created as a result. This seems, to me, like a circular argument, presupposing without evidence that "money flowing towards patent holders" promotes innovation.
Patent lawsuits are a zero-sum game with zero value: anyone's gain is someone else's loss, and no value is created when such money changes hands. When I see money just moving around based on patent lawsuits, no matter who the plaintiff or defendant are, small or large, practicing or not, individual or company, I just see wastefulness. Not quite as bad as destroying something entirely, but nothing of value either.
Inventions and products themselves, by contrast, can actually advance the progress of science.
>> Both, and other cases as well...
> Can you point to some examples of a novel technology that would have been available in the market if not for patents? Note that from-scratch re-implementations are just "new", not "novel".
I'm quite aware of the distinction drawn by patent law there, though I don't consider the patent system's definition of "novel" the only measure of a technology's inventiveness or usefulness. More critically, though it's possible to use the idea covered by a patent (or to be sued claiming such) while also making novel advancements beyond that idea; that includes both novel applications of an existing idea, as well as novel inventions threatened by general patents on entire classes of ideas. Thus, a program that is itself novel can be suppressed by a patent that applies (or threatens to apply) to part of it. See also the "Sequential Innovation" concept and paper I pointed to earlier in this comment.
In any case, some specific examples that qualify as "novel":
- Public-key (asymmetric) crypto was patented from 1977-1997. Multiple people during that period attempted to develop programs for novel applications of asymmetric crypto, but could not due to the patent. PGP was developed towards the end of the patent's life, in spite of the patent, and the patent holder subsequently forced PGP to limit it to "non-commercial use" (and AFAICT the general belief is that they'd have killed it entirely if not for potential bad publicity). In the absence of patents, encrypted email would have been available in the market far earlier. (Note that the patent wasn't for encrypted email; it was for asymmetric crypto in general; PGP was novel, as was asymmetric crypto before it.)
- WebRTC (and, in general, the availability of video in web standards) was blocked for years over what codecs to use, due both to patents on h.264 and threats of patents against VP8. Video on the web would have been available years earlier if not for those patents.
- Some of the software implementing MP3 (and undistributable in the US or anywhere that enforced the patents on MP3) included multiple novel algorithms to produce better MP3 files (e.g. in bitrate allocation, models, performance), and produced significantly better and smaller MP3 files (and did so faster) than the "official" implementation.
- HyperProf, a piece of Java profiling visualization software (back in the JDK 1 days), was blocked due to a patent by Xerox on displaying trees (unrelated to profiling).
- GCC would have had runtime memory bounds checking (to prevent array overruns, etc) in 1996, years before the availability of programs like valgrind and similar, if not for a patent by a company named Pure Software. The GCC implementation included several novel additions beyond the patent, and the patent was not known until after the work was done (and subsequently shelved).
> Since this is a Microsoft thread, appropriate examples would be Stax electronics and i4i. See also the Chien paper above which may have more numbers.
Both of which I'd hold up as examples for "why we shouldn't have software patents". In particular, if you see i4i as a good example, then I'd deeply question your evaluation of value.
>> I don't believe technological progress will grind to a halt, or even slow down, with patents (or even just software patents) abolished. I won't necessarily argue with as much confidence that it will massively accelerate, either, though it certainly will in areas I care about.
> Depends on what you count as "technological progress". I don't necessarily count re-implementations of existing technologies as progress.
See the above cited paper on "sequential innovation", and several of the above examples of novelty above and beyond a threatening patent.
(Apart from that: even a reimplementation of an algorithm without algorithmic improvements or other associated novelties may solve practical problems or otherwise improve on execution. But even if you don't consider that progress, see "sequential innovation" and the above examples.)
> I guess if you think, say, the Wright brothers didn't really make modern flight possible or that Apple didn't really revolutionize the smartphone industry, not much I can say.
That strawman sure was flimsy. First, I neither said nor implied either of those things. I'd quibble over the term "revolutionize" in the latter case, but hyperbole aside, Apple certainly made notable advancements. But in both cases, so what? Neither of those contradicts my statement that "I've never once seen a patent lawsuit in which I could muster the slightest sympathy for the plaintiff or the slightest belief that they deserve anything". If I see Apple suing someone over smartphone technology, or for that matter someone suing Apple over smartphone technology, in neither case would I think the plaintiff deserved anything.
You seem to presuppose a pile of additional reasoning that does not follow.
- throwawaykf05 11y ago> As a meta-item on this discussion, it appears that we have a disagreement of value systems. I agree :-) > I'd recommend "Sequential Innovation, Patents, and Imitation" (http://www.researchoninnovation.org/patent.pdf http://www.researchoninnovation.org/patent.pdf), among others. Personally, I would not recommend anything by Bessen without significant cross-checking. Many of the author's works have been shown to have poor methodology. As I recall (it's been a while since I read it), this particular paper presents a theoretical model with little empirical evidence to back it up. Additionally, it posits that R&D intensity does not increase with more patenting by firms, something that later empirical studies refute. > Evaluating the positive effects requires not only identifying an invention/product, but determining whether the invention/product would still have been published/created in the hypothetical scenario in which patents did not exist. That is one way to do so, but that is a narrow view. At the macro level, this can be shown by the amount of investment into R&D activity that is evidently incentivized by patents. More money and resources funneled towards innovative activity enables more innovation. While it may not serve as direct proof of specific innovations, it is obvious that without resources being invested into R&D, many innovations would not have occurred sooner or at all. Would Microsoft pour billions into MSR if it did not hope to profit from it at some point in the future? > The meta-study covered three points: 1) efficacy at promoting innovation (one genuine purpose of the patent system), for which it failed to provide any direct evidence (things that would not have happened otherwise) and could only fall back on economic claims (rewarding things that were done; see (3)); 2) efficacy at promoting publication (another genuine purpose of the patent system), which it rather conclusively refuted; and 3) economic incentives and monetary evaluation of patents, which as a non-goal of the patent system (only a means), is irrelevant when evaluating whether a patent system should exist. The goal of a patent system is "To promote the Progress of Science and useful Arts". Promoting disclosure, as is capturing rewards, encouraging financing and increasing innovation through forced work-arounds, are all post-hoc rationalizations of that clause rather than explicit goals. It is incorrect to say economic incentives are a non-goal. Insofar as it diverts more resources towards R&D, patent systems evidently do well. > ... I don't consider the patent system's definition of "novel" the only measure of a technology's inventiveness or usefulness. Agreed. I have a number of issues with the current implementation of patent systems, but not with the fundamentals. > PGP, GCC, WebRTC, etc. Thanks for the examples. I'll note that often what was prevented was re-implementation of existing technologies available on the market, and nothing prevented the research and development of follow-on innovation, only its distribution to the market, which ostensibly would also have been allowed if a license was taken. Sure that would have increased costs of downstream applications, but why should products not pay their fair dues to the upstream innovations that made them possible, even if the products themselves are given away for free? In the case of MP3s, patents were directly responsible for spurring the then-languishing development of Vorbis, which many consider a better technology. > Both of which I'd hold up as examples for "why we shouldn't have software patents". In particular, if you see i4i as a good example, then I'd deeply question your evaluation of value. Both of them were operating companies whose technology was ripped off by Microsoft. I don't see what's so bad about i4i - they were a practicing entity that filled a business niche that MS ignored for years and then decided to just take over once i4i proved it was valuable. > But in both cases, so what? Neither of those contradicts my statement that "I've never once seen a patent lawsuit in which I could muster the slightest sympathy for the plaintiff or the slightest belief that they deserve anything". So this was not a strawman after all, and I guess therein lies our disagreement. I believe these inventors took enormous risks and caused an undeniable leap forward in their respective fields, and should have reaped proportional rewards. They certainly did well enough, but not nearly as well as they should have. Different value systems, I suppose.