A student writing a thesis about software patents emailed me asking about cases "in which having a patent on an algorithm prevented some significant technological progress." HN seemed the best place to find answers. What are the clearest examples of this happening, and how much did they slow things down?
I have a couple of examples from my personal experience, neither earth-shattering, but when you consider X million engineers running into similar issues, the cumulative tax on innovation is pretty large.
Working on F1 '98, we had to scrap a whole training mode where you'd see your time compared to previous laps because lawyers were concerned it overlapped with the Atari 'ghost car' patent:
We spent a lot of time and energy trying to work around this, we weren't showing any kind of ghost car, just a time indication, but apparently the patent was broad enough to cover any kind of comparison.
The second was a patent on controlling any kind of video effect based on a sound input(!). This severely constrained what we could do on a major video-processing package, forcing us to avoid some features. They tried to extort independent developers out of a lot of cash, despite being a painfully obvious idea that had been around for decades before the patent was filed:
Xerox Parc has a patent on a type of filter viewer that acts like a looking glass to see the effect of the filter. Every year, someone reinvents it and presents it at some HCI conference, only to be told they can't use it. They are sitting on it, not building it, and also enforcing it.
The patent on SIFT is particularly frustrating because it's free for researchers to use so it ends up in every paper that concerns interest point recognition. So if you ever wanted to build an application implementing paper research, it makes it kind of tricky.
A fairly decent alternative (although less resistant to rotation) is SURF which isn't under a patent AFAIK.
The SIFT patent probably is preventing progress in computer vision products, since I think it's exclusively licensed to one or two companies. I know of folks who have asked permission to use SIFT within a product and been told that they can't have it.
Does Amazon actually enforce that? My understanding (perhaps incorrect) was that they quit after the hubbub when they tried it on Barnes and Noble. Also, isn't Bezos notoriously opposed to software patents?
It's hard to tell if a company with a large patent portfolio is only planning to use them defensively.
Does Amazon lend any support to patent reform efforts?
"Reback often tells the story of how a team of IBM patent lawyers went to Sun Microsystems Inc. in the 1980s and claimed that the then start-up was infringing on seven of its patents. After Sun engineers explained why they were not infringing, the IBM lawyers responded that with 10,000 patents, they would be sure to find some infringement somewhere." The Washington Post
"One could be tempted to consider ever stricter IP protection regimes to provide ever more stimuli for innovation. This conclusion is wrong, however. A prime example is patents on software, which might at first sight be seen as a logical expansion of the classic technology patent. But creating software differs markedly from creating machinery and the like."
Deutsche Bank Research
"If people had understood how patents would be granted when most of today's ideas were invented and had taken out patents, the industry would be at a complete standstill today."
Bill Gates (1991)
"The world's patent systems need reform so that innovation can be properly rewarded. (...) It is becoming ever more apparent that the patent system isn't working."
The Economist (11 November 2004)
"Building up a patent portfolio by engaging in defensive patenting cannot always protect against hold-up."
Federal Trade Commission of the USA
"More patents in more industries and with greater breadth are not always the best ways to maximize consumer welfare."
Federal Trade Commission of the USA
"The mild regime of IP protection in the past has led to a very innovative and competitive software industry with low entry barriers. A software patent, which serves to protect inventions of a non-technical nature, could kill the high innovation rate." PriceWaterhouseCoopers
"SMEs are crucial providers of pathbreaking innovations, but would be most adversely affected by patentability. The majority of them is deterred by the costs of patenting themselves, but would have to navigate around software patent portfolios of large corporations."
Deutsche Bank Research
"Many large companies operating on a global scale, including European ones, seem to be in favour of a software patenting regime. But most small enterprises are strongly opposed."
PriceWaterhouseCoopers
Microsoft Corp. warned Asian governments on Thursday they could face patent lawsuits for using the Linux operating system instead of its Windows software.
Reuters (18 November 2004)
"A future start-up with no patents of its own will be forced to pay whatever price the giants choose to impose. That price might be high: Established companies have an interest in excluding future competitors."
Bill Gates (1991)
"Patent law provides to inventors an exclusive right to new technology in return for publication of the technology. This is not appropriate for industries such as software development in which innovations occur rapidly, can be made without a substantial capital investment, and tend to be creative combinations of previously-known techniques."
Oracle Corporation Patent Policy
Read B again, and recognize that there is no such thing as a software patent "mainly to protect themselves". Consider the patent portfolio effect right from the first, wsj quote.
One thing that's difficult to measure but probably not trivial is the overall cooling effect. What I mean is that the tangible cases in which patents really ran into conflict don't indicate the number of cases in which technologies were never researched to begin with because of the risk of infringing a patent.
The cooling effect probably isn't significant for the mythical basement hacker who does it for love, but decisions about what research to pursue are made all the time in larger companies, and one of the factors is potential for legal trouble.
This is an excellent point. Basically the patent minefield concept. The insanity of software patents has created an overall sense of fear. There is no way to quantify this aspect of patents' negative effect on progress.
I'd say that HTML5 is the tip of that particular iceberg.
I mean, what was HTML 5 supposed to change? The fact that video (and audio) is broken on the web, leaving to flash as the closed de-facto standard. The whole situation that HTML5 was supposed to fix is, itself, a result of codec licensing/patent issues.
I'd venture that without those issues, the <object> and <embed> tags would have had a bigger impact.
June 2009 is not "current". That issue (of making one or another codec mandatory in the HTML5 spec) is dead with decision you linked, for the reason that "if a browser
refuses to implement something, then we can't require it".
I'm being pedantic here, but I think it's very misleading to connect the spec issue with your conclusion that <video> could have been nice but somehow isn't. Had one or the other codec been made mandatory in the spec, the reality today would be exactly the same.
Singular Value Decomposition was held back for 20 years by patents owned by BellCore. I had to scrap large interesting portions of Dowser to avoid potential problems.
As a counterexample, Autonomy has a patent on Bayesian text classification dating from the late 90's. I've always wondered why they didn't say anything after Paul popularized it for spam filtering.
Elliptic curve cryptography and LDPC codes are two that come to mind.
Though afaik the latter hasn't really been improved since ~1960 and were only patented after rediscovering them in 1993, and I have no clue whether progress in elliptic curve cryptography stalled due to patents or because they don't offer any practical benefits over RSA.
They ended up losing a lot of those claims recently if I recall correctly, but that being said I agree for a long time it did and still does hurt the LMS environment.
It could have been a backwards compatible extension to GIF instead.
More formats = fragmentation = consumer confusion and problems.
Progress is not a single variable either, there are plenty of ways that you can define progress that are more or less orthogonal to each other.
Forward, sideways and 'up' are all movements that you can interpret as 'progress' depending on where you are and where you want to be.
From that point of view PNG was sideways and 'around' rather than forward.
Any kind of dataformat, any kind of protocol, anything that is not a completely closed box that performs a function (an 'appliance') should be as open as possible.
And that should be enshrined in law, interfaces and formats are automatically in the public domain, and have to be documented.
PNG as a GIF replacement is superior, for most applications. But if you think in terms of browser support then you could interpret it as a step back during the introduction phase as well.
For example, the PNG standard still isn't fully supported by Internet Explorer. Who knows what could have been different if the early standard were open and extensible (without patent difficulties)..
I still don't know what the issue with supporting multi-touch devices in North America is but I strongly suspect it's because of software patents. Anytime a corporate rep is asked why their device doesn't support it they hem and haw and say that it's complicated but never give a real answer.
There is a ton of really interesting and cool stuff that was done in wavelet research years ago which got patented and hasn't really seen the light of day in the way it should.
If the patents hadn't existed I'm sure we'd now have much better widely available voice recognition, image recognition, and related data processing software than we do now.
During development of Doom 3, id Software fell into a bit of a patent conflict with Creative. Relevant snippet from http://en.wikipedia.org/wiki/Doom_3:
"A week before the game's release, it became known that an agreement to include EAX audio technology in Doom 3 reached by id Software and Creative Labs was heavily influenced by a software patent owned by the latter company. The patent dealt with a technique for rendering shadows called Carmack's Reverse, which was developed independently by both John Carmack and programmers at Creative Labs. id Software would have placed themselves under legal liability for using the technique in the finished game, so to defuse the issue, id Software agreed to license Creative Labs sound technologies in exchange for indemnification against lawsuits."
I remember something about Carmack coming up with a workaround that paid a performance penalty so others licensing their engine could bypass the Creative mess.
Creative has also taken quite a bit of heat over the years for sitting on Aureal's IP and stifling innovation in sound technology, but I don't have specific examples to cite.
I worked at a company called Pica9, generating PDFs on the fly. Another company has the patent on that, so we had to spend a considerable amount of effort working around it, making sub-optimal choices at easy decision point in order to make it legal.
The potential "technological progress" doesn't exist because it was thwarted by software patents, so there really is no way of knowing what doesn't exist because it couldn't be created to begin with.
No; the question asks what things would have existed. This is a concept people use all the time in everyday life without provoking ontological crises. E.g. if you say "he would have scored a touchdown if that last defender hadn't tackled him," everyone knows exactly what you mean. You can never be 100% certain about things that would have happened, so you qualify your conclusions, but it's not the case that you can't reasonably talk about such things.
I can't believe I just had to write that. This is why forums are such a time sink. If you were watching a football game with your friends and someone said "he would have scored if that guy hadn't tackled him" and you replied "how can you talk about touchdowns that don't exist?" everyone else would just roll their eyes and not invite you over anymore. But here in the world of text these subtly graduated social cues don't exist; all you can do is reply with more text. I wonder if there is some kind of solution to that.
> This is a concept people use all the time in everyday life without provoking ontological crises.
+1 for common sense; however:
> This is why forums are such a time sink.
I doubt that you can realize the benefits of an open forum without paying what we might call the free speech tax - the inevitable disruptions of pedantry and/or trolling, which under the right circumstances are insightful and even revelatory.
Sadly, you're probably right. What's depressing is that HN wasn't like that for a very long time (2 or 3 years, maybe?), but it seems to be descending into a Reddit-like cesspool of crazy very rapidly.
Having said that, this is still by far the best non-trivial online community that I know of.
EDIT: Wow... downvoted in 4 minutes, for a mostly harmless comment. I guess we've become closer to Reddit than I thought. I normally hate elitist groups, but there's a part of me that would like to see every account newer than, say, 200 days just get nuked. <sigh>... I really loved this site.
Please note that I'm not saying HN is turning into reddit - I don't believe my score is high enough for me to be allowed to make that claim. (I do believe HN would benefit from a measure of reddit-style levity: right now it's all Serious Business, all the time.)
What I'm saying is that real constructive critical debate requires strong-willed, pedantic, even curmudgeonly debaters. The marketplace of ideas doesn't work without the push and pull of challenges, refutations and even meta-refutations ('You're asking the wrong question').
If Paul Graham wants good answers to his question, the price he has to pay is an allotment of bad answers - answers that are revealed as such on the marketplace of ideas. You can't get the good stuff without also allowing the bad stuff (and having reasonable filters to distinguish them).
P.S. Not sure why I was downvoted for my comment above.
The tone might have been better, but it's a valid argument that I have heard in conversations in person. Unlike football physics, society is a much more complex system and it's really hard to play what-if scenarios with any confidence about it.
In society usually there are many forces acting in opposing directions and the outcome depends on how strong they are, which is a quantitative difference we currently have no way of estimating.
First things first, Brett Favre should not have thrown that pass last night. Vikings would have won.
Second, I'm taking this from the perspective that we have a college professor (most likely) being asked by a student to cite examples of cases where software patents have stifled innovation.
So, does this professor just want his student to start making up random ideas that infringe on patents and argue that those random ideas don't exist because of patents?
Seems unlikely.
I'm arguing the question itself because I am trying to help this poor student who may be faced with more of a paradox than a fair assignment.
I think it is a perfectly valid question. If someone got sued by the holder of a patent and withdrew a piece of software or an application then that's proof positve.
Would be a nice example, in http://brej.org/yellow_star/ there is a bit that says "There are instructions which are patented by MIPS and have been removed.".
That's a clear impediment to progress, once the MIPS instruction set was documented and people started to write software for it anybody ought to be free to re-implement these instructions for the express purpose of interpreting MIPS object code.
By analogy - name a great business impeded by corruption in the government. Unm. We know there are great businesses in America and not that many in (more) corrupt countries, but we can't point at killed businesses themselves.
I can name you a great business from one country that was impeded by corruption (or at least a connection to industry that should not have existed) in the government of another though:
But the magnitude can only be determined by having lots of examples. You are saying that there is no point in collecting any examples at all but when you don't have any then the magnitude could still be '0'.
By showing examples, preferably lots of them the magnitude will get at least a lower bound.
I'm not sure it's possible for people to imagine a different world that easily. The differences caused by software patents not existing are pervasive and cumulative. Like the surprising effects of compound interest, I think a world without the 20 years of software patents would be in many ways unpredictable or even inconceivable for those of us who lived through this reality.
The original question may make sense in a limited fashion, but it frames things poorly as it's asking for specific incidents to illustrate a systemic effect.
By progress do you mean creation of new technology or making technology available? A case I followed (involving infringement on a fingerprint recognition patent) shows how even weak patents slow the development process.
Summary - Small company receives patent, sues large companies for infringement, two years and more than a million dollars later patent is invalidated, technology moves on. All the gory details here:
In general, anywhere you find Huffman coding being done, it could achieve superior compression with Arithmetic Coding but is not because of IBM's patents.
This is essentially every JPEG image on the planet. The standard provides for arithmetic coding, but no one implements it because of the patent. Wikipedia asserts that arithmetic coding saves about 5% of the files size.
So that is 5% of every flash card sold for a camera wasted and 5% of the bandwidth costs for images on the internet wasted.
If GATT hadn't tacked on three more years in 1995 these would have expired by now, but it looks like 2012 or 2013 is the date now.
The MPEG series of standards probably could also have benefited similarly from Arithmetic Coding.
> The standard provides for arithmetic coding, but no one implements it because of the patent.
That is, there is an implementation in the libjpeg source tree, but it's disabled. I thought the patent had actually expired now, though.
Also, H.264's arithmetic coding mode saves ~20% bitrate; Vorbis or AAC could presumably benefit too, but neither of them use an arithmetic coder, even though AAC should have no patent problems. Maybe there's some technical problem I don't know about.
There's the FAT32/vfat issue that caused Linus to take out the write support for long filenames. Not helpful, but I wouldn't call it impeding innovation.
Not an example, but there are many cases of people "working around" a patent. In this case, it is a spur to technological progress. That is, preventing progress can encourage progress.
There's a "prospect" theory for patenting (which with this student is probably familiar), that a patent is like staking a claim to minerals, with a samilar effect: that it encourages others to stake claims nearby, and so explore that region. If others could use the same area as the patent, they wouldn't be encouraged to explore, and you'd get a technological monoculture (in another sense, this would be greater progress, in that it is more widely adopted).
I've been trying to think of an example, but I can't. I think the main way a patent can block progress is if the technology is not exploited well by the owner (like smalltalk being too expensive), and in such cases, we don't hear of it. When a patented technology is exploited successfully, money and time is reinvested in developing it, so that progress on it is accelerated.
However --- imagine that the claims on minerals would be hard to see [1], and accidental violation would put you at great legal risk. Then people might get discouraged from staking claims in that general region.
[1] It's not easy to look through all the patents before building anything --- and may even be harmful to your legal position.
Apple's patents on sub-pixel text anti-aliasing. The text in both Windows and Linux looks hideous by comparison. You can enable these features in certain situations on Linux with extra packages, though.
Correctional update: Microsoft holds patents in this area but Apple has a cross-licensing agreement on them.
Apple has patents on truetype hinting, not subpixel anti-aliasing. Ironically, when working around this patent, it is easier to get results similar the OSX, rather than similar to Windows.
it's hard to find any examples where patents have stiffled innovation. Of all the hundreds or millions of patents it's only possible to find very few if any historical examples where patents stifle innovation.
What patents do achieve is exactly what they are there for. which is to protect the inventor of something usefull for a period of time.
If you want to use innovation that has patent protection, you should reward the person who invented it, OR invent something better youreself.
That applices to things like Ogg Thorbis, where new technologies were developed in spite of patents. If there were not patents on MPEG Ogg would not have been innovated.
Same with file systems, if there were no patentson NTFS or FAT, then why would ext3 and so on mabey would not be where they are now.
Fact is patents are there for a reason, (apart from it being law). They appear to work, and they dont appear to be affecting innovation to any degree.
Sure, there are groups who would be happy to ride the coat tails of others innovations and inventions. But that is only so they are not able to innovate themselves. They like something that works, and have a 'me too' attitude.
(why should someone else get rich, when I can copy that and get rich as well).
I would direct the student to the working Paper by James Besser and Eric Maskin, Sequential Innovation, Patents, and Imitation, which builds an economic argument against software patents. It and other relevant papers on software patents are available at
http://www.researchoninnovation.org/.
That was one I was going to mention, along with the Wright Brothers, before I reread the question and realised it was only software patents they were interested in. Still fascinating how people that I'd always heard venerated as inventors could have spent so much time and energy holding other people back.
There certainly have been cases where alleged software patent violations result in lawsuits. There the lawyers get rich while very senior engineers cool their heels in court or give depositions in places unusual, uncomfortable, and distant.
I don't know how measurable this is but it is patently obvious it is preventing progress.
Sorry, no time to do the legwork to see if this would work, but I wonder if this "wastage" could be crudely measured by taking some aggregate number of patent related lawsuits per year, multiplying by some factor to include those settled out of court, multipy again by a conservative mean cost of a patent action (attack, defense, commonwealth) == $wasted/year.
Working on F1 '98, we had to scrap a whole training mode where you'd see your time compared to previous laps because lawyers were concerned it overlapped with the Atari 'ghost car' patent:
http://kotaku.com/270035/patents-are-interesting-ghost-mode
We spent a lot of time and energy trying to work around this, we weren't showing any kind of ghost car, just a time indication, but apparently the patent was broad enough to cover any kind of comparison.
The second was a patent on controlling any kind of video effect based on a sound input(!). This severely constrained what we could do on a major video-processing package, forcing us to avoid some features. They tried to extort independent developers out of a lot of cash, despite being a painfully obvious idea that had been around for decades before the patent was filed:
http://www.trapcode.com/US_SK_advisory.html