So DataTreasury fixed the way check clearing works, apparently pioneering a new digital approach instead of mailing around checks; implemented it; raised money and created 100 jobs; but couldn't get the banks to work with them. The banks then took the idea and implemented it themselves, driving DataTreasury out of business.
Now their investors are getting some money back by licensing the patents to the banks, so they'll (hopefully) be willing to invest in innovative startups that have to work with incumbents again.
Next time a small innovative company comes along, the banks are (hopefully) going to think twice before just deciding that they can easily re-implement their ideas.
This doesn't seem to me like a particularly good example of the patent system gone wrong. To me, this looks like exactly what the patent system is supposed to be: protection for those with innovative concepts, so that they can have the opportunity to commercialize their inventions without incumbents simply cloning their hard work.
So DataTreasury fixed the way check clearing works, apparently pioneering a new digital approach instead of mailing around checks; implemented it; raised money and created 100 jobs; but couldn't get the banks to work with them. The banks then took the idea and implemented it themselves, driving DataTreasury out of business.
(snip)
To me, this looks like exactly what the patent system is supposed to be: protection for those with innovative concepts, so that they can have the opportunity to commercialize their inventions without incumbents simply cloning their hard work.
So you'd agree that if things were right in the world, we should all be using the original version of WebCrawler today, because it pioneered the idea of web search, which these other horrible idea-stealing companies like Google took and re-implemented?
Do you really think that's the way business should work, that whoever first notices that a problem needs solving (and bothers to submit a patent application) should have a long-term monopoly on solutions to that problem?
Personally, I'd prefer that the means of solution, if non-trivial enough, was protected, rather than the entire set of all possible solutions to a problem.
I prefer we base our standard on what the economic benefit and damage there are to using a patent system, not on what constitute obvious and non-obvious.
If there are huge economic impact that are for the worse as a result of patenting non-obvious solution, than we should scrap the system altogether.
"Personally, I'd prefer that the means of solution, if non-trivial enough, was protected, rather than the entire set of all possible solutions to a problem."
Well, it's what patents are supposed to be. In reality, it all depends where you draw the line between "problem" and "solution", and IMO this is where a lot of patents, especially software ones, go seriously wrong.
If the problem is defined as "send images of checks instead of checks", then DataTreasury is trying to assert ownership over the entire problem rather than their particular solution - they're claiming that any solution that deals in images of checks instead of physical checks infringes, regardless of how it's implemented.
If you define the problem as "don't send physical checks", then sure, the particular solution, "send images of checks instead" is what they're asserting ownership over.
But in that case, the non-triviality of the solution is seriously questionable. There's no engineering insight at all required to come up with that, and it's one of the first things any group of people would think of when presented with the problem. The entire difficulty is all in how to actually implement such a system (and pre-1994, those implementation details made it unlikely that such a system would work).
IMO, in a situation like this, what the patent should cover is the new stuff that DataTreasury actually invented (and I agree that there is some non-trivial content there), not the entire high level problem that their invention allowed them to solve.
So many patents these days (in software, at least) are just as you describe: cases where someone realized that a problem existed, came up with a quite straightforward solution, and patented it.
We need Congress or the courts to tell the PTO that "obvious" means "the solution is straightforward given the problem". And then, of course, we need to get the PTO to hire software engineers who actually know how to build things, so they know what that means.
I think you're missing the whole point. The DataTreasury example illustrates a fundamental problem with patents which is only exaggerated when it comes to software and software "processes." The problem is, as the article pointed out, that the patents are construed to be so broad that basically anyone can be said to be infringing it. To patent the idea itself of 'scanning checks with a device and sending it to a server' is pretty absurd. A copyright is more appropriate. If you copied the code, etc then you violate the copyright. But to say that anyone that implements this idea is in violation ... that's not encouraging innovation. It just encourages running to the patent office as fast as you can.
I don't think it's fair to say that "anyone can be said to be infringing it". I think if you're a bank that's using scanned check imaging instead of mailing around checks, you're (likely) infringing. If you're building a social network, you won't be. In other words, the people who are (accused of) infringing are precisely those that were the target market for DataTreasury; there's also no accusation that the patents have been enforced against anyone other than the banks.
The US banking system feels pretty backwards in terms of its reliance on checks where Europe has adopted electronic transfers. If the banks actually wanted to innovate here, they'd introduce a real electronic transfer system. The idea of scanning checks and sending around images (and presumably one day constructing virtual check images electronically and sending those) strikes me as remarkably backwards, but nonetheless a very clever solution to a real-world problem.
Maybe I missed part of the story, but didn't it say that banks implemented their own internal systems? If so, how does that violate a patent?
If you build a better mousetrap and try to sell it, but I decide to build my own copy of the mousetrap for my own use, wouldn't that be non-infringing?
That's the big difference between patents and copyrights. With a copyright, you can only infringe if you copy code. With patents, you can infringe if you implement the same idea. I think it makes many people uneasy with the patent system if someone built the same system without knowledge of a patent. The non-obvious clause is supposed to protect against that (and there are valid flaws in the implementation of this aspect of the patent system!)
It's not clear to me whether in this case the check scanning systems were in fact independently invented; there's certainly a hint in the article that meetings took place and the banks only implemented the idea once it has been disclosed to them ('willful infringement' results in triple damages, which could well explain the magnitude of the licensing amounts)
If I have a patent on the better mousetrap, that is absolutely infringing — you are building my patented invention without my approval. I think you're conflating patents and copyrights.
I think you should review the protections that copyright provides. I believe your concept of copyright does not match that provided under the law. http://www.copyright.gov/title17/92chap1.pdf
Most ideas are obvious in retrospect. Why was DataTreasury the first to implement this system? Seriously engage with this question. I had just started working in the industry in 1994, when this story started. Why didn't I create a B2B electronic transaction clearing system?
I'm not necessarily sticking up for this patent, and I tend to believe the patent system is completely fouled up for software companies, but we can do better at analyzing the issues than just automatically saying "that's brain dead obvious".
I agree that most ideas are obvious in retrospect. Still, In my experience most ideas are obvious before hand as well. So what is a good way to define obvious or novel as applied to patents? I have two ideas.
1. Boggle rule. In the game Boggle when two players come up with the same word neither get credit. If from a given time of filing a patent someone else files a similar patent both are rejected.
2. Virgin engineer rule. Take a group of people competent in the field of the proposed patent but without any knowledge of the patent. Give them the problem the patent purports to address. If in a single day their solutions substantially match the proposed patent, reject the patent.
I think this is a great discussion; much more productive than the normal religious arguments.
The first is reasonably sound. A potentially very interesting idea is to allow anything filed within 3 months of filing to be considered "prior art", although I'm not sure this would make a huge difference in practice.
The second would be very labor intensive if done in practice. Doing this as a thought-experiment is actually how non-obviousness is determined by patent examiners. If an engineer of "ordinary skill in the art" would think the approach obvious given the prior art, then the patent is deemed obvious. I don't know if it's ever been tried to locate 'virgin engineers' to demonstrate that a patent is obvious.
So the second idea is how the system works, although the examiner is supposed to be the group of engineers.
> Recognising the problem is often the innovation.
It can be an innovation, but it can be very hard for the PTO to tell just how much of an innovation it is. What happens a lot, I believe, is that the emergence of a new technology creates new possibilities, but the actual application of the technology to those possibilities raises certain problems that have to be solved. The existence of those problems is not obvious in advance of the new technology, but once people have it in front of them and start to play around with it, it starts to become more and more likely that someone will think of those problems. The evidence for this is the not infrequent occurrence of simultaneous or near-simultaneous invention.
So while recognizing the problem might be an innovation, it might be an almost inevitable innovation. Furthermore, much of the credit might arguably belong to the enabling technology that made a solution possible, and to the industry that developed and commercialized that technology.
To make an extreme simile, one could say the Internet is like the technology to allow us to colonize the moon, and so many of these software patents are like people booking trips to the moon, arriving, and then staking a land claim. It wasn't very much their own intrepid exploration that made it possible for them to stake that claim; there might have been a little of that, but mostly they're dependent on the massive effort, by an entire industry, that made it possible for them to get there in the first place. So the policy question is, have they really contributed so much that we should reward them with that land claim?
> So while recognizing the problem might be an innovation, it might be an almost inevitable innovation.
Or, it might not.
> Furthermore, much of the credit might arguably belong to the enabling technology that made a solution possible, and to the industry that developed and commercialized that technology.
So? They're free to recognize the capabilities and invent.
However, that's a red herring because you also don't want them to be able to get patents for recognizing the possibilites.
I get that you don't like "recognize the problem" inventions. Howver, can you argue against them on the basis of "it was inevitable" or "someone else should have gotten the credit"?
Which reminds me - can you name three inventions that weren't inevitable and how you know? I'm trying to figure out if your argument against "recognize the problem" patents applies more generally.
Why was DataTreasury the first to implement this system? Probably because technology was just getting to the point where their idea was feasible, and megacorp banks do not tend to be very fast-moving. I mean, look at the state of banks today — they're still 5–10 years behind the state of the art.
Was there a flood of overlapping patents for similar technology at the same time?
I brought up my entry into the industry in 1994 because '94 wasn't a slow year for tech at all. It was about a year before the bubble started blowing up, but there were lots and lots of small companies doing software.
The question is: Did the "pioneering a new digital approach instead of mailing around checks" represent a significant investment that needs to be protected, or was it just a good idea?
Patents do not exist to give you a an easy monopoly on any random idea (or at least they shouldn't). Patents exist to give companies an incentive to invest into costly research and then be able to publish the ideas, while getting some time to recoup the cost of the investment.
The litmus test is, would this idea have seen the light of the day without a lot of upfront cost? If so, it is not in the interest of society as a whole to grant a legal monopoly here.
This is a frighteningly naive article. Consider the following extract:
"There are at least three reasons to exclude software from patentability. First, software development is an individual, creative activity, more akin to writing a novel than designing a jet engine...Second, software patents are especially prone to litigation...Finally, software patents are unnecessary because software is already eligible for copyright protection...As long as programmers write their own code from scratch, they can be confident they aren't infringing others' copyrights."
The idea that software development is the sole province of the lone inventor is both wrong and irrelevant. The reason that our constitution facilitated patents was nothing to do with how many people invented something but rather with encouraging disclosure of the result.
If we stipulate that software patents are frequently violated...so what? Without further analysis of why or whether specific improvements to the system would be helpful this is a dangling factoid - no more.
Finally, the idea that copyright provides any kind of protection for software is so ludicrous the writer's naivety is laid bare. Copyright protects only the expression of an idea. In the case of a work of art such as a novel or painting the expression is critical to the work and copyright is a valuable protection. But if a programmer sits with original source code before him and rewrites it in another language thereby benefiting enormously from the original work but changing the expression completely they may indeed be programmers writing "their own code" but it certainly isn't "from scratch", and whilst he is right that "they can be confident they aren't infringing others" it is blindingly obvious that they have copied. Copyright is worthless if the expression doesn't matter and all that matters is functional equivalence. And guess what, users don't give a rat's ass how a program is expressed so long as it works.
Thanks for reading my article, and for sharing your perspective on it. However, I don't think you understand how copyright law applies to computer software. Translating a computer program line-for-line to another programming language doesn't avoid liability for copyright infringement. If it's "blindlingly obvious" that a copy was made, a court is going to say infringement has occurred.
I recommend reading about clean room reverse engineering. That's what you have to do to copy software without infringing its copyright. It's doable, but in most cases it'll be more, not less, work than writing your own software from scratch.
I am afraid that it is you who evidently does not understand my point. I am not suggesting that some imbecile may transcribe in a manner to blindingly obvious that they may run into trouble. What I am suggesting is that access to the code for anyone with an IQ over 50 is a HUGE help in building your own version and that this assistance is not inhibited by copyright.
I am extremely familiar with clean room reverse engineering. You are absolutely mistaken if you believe that 'That's what you have to do to copy software without infringing its copyright.' Reverse engineering actually has a rather different history. Given a device, say a chip, take the inputs and outputs of said device which defines the machine's behavior. Now design and build a machine which reproduces such a mapping. This practice was frequently associated with 'clean room' protocols. The key point is that the team had absolutely NO access to the source code. And because such a clean room team would never hire anyone who had previously had such access they could be confidently said to have reverse engineered the behavior from the APIs alone. This is a VERY different situation.
It's true (AFAIK; IANAL) that if you are reading someone's source code and writing your own based closely on theirs, you may well still be infringing their copyright.
But the parent has an valid point nonetheless. If you do a clean room reimplementation of a program, then you are off the hook as far as copyright goes. Yet the user may not be able to tell the difference. In short, the user cares about the functionality of the program, not its expression (the source code).
I think that's a valid argument that copyright protection is not entirely sufficient for software, and that patents could have a useful role to play.
Personally, I think the problem with software patents today is that the obviousness bar is far too low -- we're patenting problems, not solutions. See my comments elsewhere on this page.
But clean-room engineering is almost always slower than writing the software from scratch. Which, AFAICT, is what most companies actually do except in the rare case where they need to achieve interoperability with an existing product (like the early PC clones where bug-compatibility with the IBM PC was essential). Right?
I don't know if it's slower or not. It takes a certain amount of time to develop a functional spec; you have a head start if you have a working example in front of you.
I think the main reason you don't see more clean-room reimplementations is simply, as you say, that level of compatibility is rarely necessary -- indeed, it's often considered undesirable, even by the users.
That doesn't make any sense. It is the same as writing the software from scratch. The only difference is that the behavior of the original software which is being copied is effectively the spec.
I agree with the article that broad patents are harmful -
and if massive banks cannot get a fair outcome in the courts, what hope is there for the rest of us?
I also agree with the implication that patenting "specific techniques" is beneficial. The history of the industrial revolution is littered with such innovation. It's only correlation, but the places with patent laws also had the most vigorous innovation. At that time, the newspapers were filled with new innovations and inventions, and people seemed to be falling over each other to invent something new, and to patent it. They thought patents promoted innovation - and the drafters of the US constitution thought so too.
Personally, I would love to see new specific techniques being invented, rather than the next facebook, or the next webapp of something already done offline. The latter can be really useful and a great benefit to the world, but it doesn't stir my soul. I want new technology and entirely new ways of doing things! To make this more concrete:
- I think that Ken Thompson's specific mechanism for rapid regexp matching should be patentable (http://www.cs.princeton.edu/courses/archive/spr09/cos333/bea...) The idea of using regexp for search was new and cool (actually groundbreaking), but it's only the specific mechanism that should be patentable. Which is what Ken did.
This is nice because the Republican Supreme Court justices voted 4-0-1 in favor of the maximum possible patenting of software in the Bilski case last year. Democrats were 4-0 against software patents.
Scalia was the one deciding vote and refused to join either side's opinion on the actual subject of software patents. His decision is tentative and equivocal on the subject of software and business method patents.
This article appears in the National Review, the thinkiest of the Republican news magazines. Perhaps it can help bring a little of the craziness among Republican judges back to reason on a subject central to our industry's future on these shores.
Generally speaking, even though it probably feels unnatural and even dumb to have to think this way, if you have to frame a comment in terms of "Republican" and "Democrat", you're probably hurting the discourse on HN.
As it stands, I promise you there are many many conservative technologists who agree with your take on the patent system, but have wildly different takes on e.g. public sector unions and school choice.
(I'm not one of them; I'm a dyed-in-the-wool liberal. Ironically, I'm actually less sold on how horrible patent litigation is than most of my conservative friends.)
Speaking of judges --- and I thought your horse-race analysis was interesting, even if it made me queasy --- to the extent that there is really a such thing as a "Republican" or "Democrat" judge (as opposed to a liberal-leaning or conservative-leaning judge, or an "orthodox" or "reform-minded" judge) --- there are influential conservative judges who do not have knee-jerk pro-patent thoughts. What does Posner think about the patent situation?
Now their investors are getting some money back by licensing the patents to the banks, so they'll (hopefully) be willing to invest in innovative startups that have to work with incumbents again.
Next time a small innovative company comes along, the banks are (hopefully) going to think twice before just deciding that they can easily re-implement their ideas.
This doesn't seem to me like a particularly good example of the patent system gone wrong. To me, this looks like exactly what the patent system is supposed to be: protection for those with innovative concepts, so that they can have the opportunity to commercialize their inventions without incumbents simply cloning their hard work.