Chance to kill software patents opens
End Software Patents (ESP) and the Free Software Foundation have filed an amicus curiae brief in the Court of Appeals for the Federal Circuit's (CAFC) rehearing of the In re Bilski case set for May 8, 2008. The rehearing could lead to the elimination of patents on software.Link,/a>With the boundary to what can be patented effectively destroyed by previous Federal Circuit rulings, massive-scale liability has been created throughout the US economy. Over the last few months alone, ESP has tallied over fifty non-software companies being sued for infringement regarding their web sites or other course-of-business software, including the Green Bay Packers, McDonald's, Dole Foods, Kraft Foods, Caterpillar, J Crew, Burlington Coat Factory, Wal-Mart, and Tire Kingdom. The rest of this list can be found here.
Ironically, the Federal Circuit's own web site is produced using software that likely infringes some number of software patents.
ESP executive director Ben Klemens said, "This is an historic opportunity to fix the US patent system, as the Bilski rehearing will directly address the boundaries of the subject matter of patents. In our brief, the End Software Patents project supports the Supreme Court's long-held position that computer software should not be patentable, and has highlighted to the Court the real economic harm software patents cause the US economy."
the latest
latest episodes
Red Hat also filed a friend of the court brief yesterday:
http://www.press.redhat.com/2008/04/07/red-hat-asks-federal-court-to-limit-patents-on-software/
I genuinely don't understand why everyone hates software patents so much. Why do people say that the plans for a new type of engine are patentable while the plans for a new type of software aren't?
They're both intricate pieces of design that required valuable hours of research and ingenuity to create. The argument that software is easily copied and modified is invalid, as the blueprints for an engine are just as easily photocopied and modified. The patent system has always been about novel ideas, not physical artifacts.
I agree wholeheartedly that the system has been abused. However, that should lead for calls for better patent inspection not an abolition of the patent system overall.
NB: I'm not trying to start an ideological war here, I just genuinely don't understand your position. Please, educate me!
Tally-ho!
Let's keep our fingers crossed this actually goes further than boingboing.
Bugs: even if one grants that software patents might be sometimes valid, the sheer timeframe involved is a problem. A good engine design will be useful for a long time. A new drug requires so much testing and development that there's only few years in which to reap patent-protected profits. But look at the mp3 patents: even though better formats have already been developed and mp3 is basically obsolete, the patents are still good till, what, 2015? The patent protects them for longer than their entire useful life. That's clearly broken.
The answer is patent system REFORM and BALANCE, not abolition.
The system is undoubtedly being abused. Claims should be more finite and less vague, technologies should be proven/prototyped/working... not some science fiction idea.
But patents exist for a reason - they were designed to spur innovation. They were designed to say "hey small person with an idea... develop it! make it work! quit your job, spend your savings, compete against the large companies!"
The world is not idyllic or utopian. Abolishing patents isn't going to free innovation - its going to mean that small firms will lose all incentive to innovate. It's going to mean that only the Googles and the Microsofts can compete, because they have the marketing dollars and ability to clone fast. It's going to mean the end of progress, and the rise of braindead corporate lock-ins of technology.
-signed a former open source developer turned patent holder when large companies decided not to work with him, but instead clone his technology and push his company out of the market through economies of scale.
Matthew - I'm not sure I understand your position. If mp3 is basically obsolete, then the patents to mp3 technology are not relevant.
On the other hand, if the mp3 patents cover not only the original commercialized technology, but also are broad enough to cover the improvement mp3 technology, then it appears the original patented idea still has a useful life since idea is still being used.
Regarding software patents, I don't think the PTO was adequately funded (Congress would siphon its fees) or trained to examine such applications initially resulting in many invalid patents being granted. They are better equipped now. I agree broad, invalid patents are problematic, but not sure applies to all patents, even software patents.
Now, we have Congress, the Courts and the PTO all independently making changes that will make it more difficult and more expensive to obtain and commercialize a patent. I agree some reforms were necessary, but concerned we may be watering down patent rights too much.
@Jonathan V. I came out of years of lurking to respond to your comment. The problem is your underlying assumption that your ideas are a property that you should be able to own, keep others from using, earn revenue from, and that you alone could devise. I think this is an unrealistic and counterproductive way of looking at the world.
Shouldn't your coming up with novel, useful ideas be proof enough to an employer that you may come up with more in the future, thus earning you a nice salary? Or if you don't like the idea of an employer, contracts and commissions?
The idea of a lone inventor making it rich from a patented invention is a myth of American elementary schools. Even if an individual comes up with patentable idea on their own- software or otherwise- the expected pay out is through selling the idea to a corporation.
Corporate patents don't promote innovation, again and again the news shows them being used as a weapon against competitors. They are used to quash innovation.
@#2:
Bugs,
One reason for opposing software patents which I think is a valid and logical opposition is as follows. According to the US Constitution, the raison d'etre of patents is to promote "the useful . . . [s]ciences."
A patent does something for its creator (gives them a temporarily monopoly) while doing something to society (restricting its freedom). This temporary monopoly (monopolies are something US economic policy is against, so they are abhorrent unless balanced by benefits to society) is 20 years long, which is more than a lifetime in software development.
The length of the term is too long for software (for other inventions, 20 years might be fair, but 20 years ago we were using 5 1/4" floppies, dreaming of 1MB hard drives).
We may also look at the costs/benefits to society of software patents. Patents typically benefit society as they incentivize otherwise cost-prohibitive investment on the part of businesses, with the promise that they will get protection IF they disclose to the public how they accomplish their invention (in the form of the patent application).
The costs are monumental, and I'm sure that a Boing Boing reader doesn't need to be refreshed on these. The costs are similar to the costs to society of any patent issuing, except that perhaps the costs here are more egregious simply because software moves so fast and 20 years is similar to, say, a 100-year patent in industrial machinery.
The "Boing Boing line" or ("Cory Doctorow line"?) on software patents is that there should be none. We should inquire as to how the cost/benefit analysis would change. Clearly the aforementioned costs to society if there were no patents would go away.
Would the benefits? The question is two-fold: (1) do software developers NEED a patent incentive to develop new computer technology, and (2) would the technology become publicly accessible without the incentive to publish as required in patent applications?
I think the answer to (1) is no. Because software moves so quickly, and it's relatively easy to develop new techniques compared to the material costs of industrial design and patents, developers don't need the defense of patent monopoly. By the time their competitors ripped off their products, we'd be using even better technology, and the original developer will presumably have gotten a jump on that and will have already taken advantage of his other development to a sufficient extent.
I think the answer to (2) is no as well. We constantly hear of cases in which multiple people are rediscovering things that are already patented (submarine patents, anyone?). My experience is that we don't need public disclosure of a company's software because others will be (typically) able to trivially invent the same thing.
So we come down to whether software patents actually perform their constitutional mandate of "promot[ing] the useful . . . [s]ciences." I think my brief analysis at the very least provides an argument that no, software patents do not promote software progress.
Surely your answer to (1) unfortunately contains the meat of its own fallacy?: "Because software moves so quickly", the value of "By the time their competitors ripped off their products" is severely reduced. Any competetive advantage produced by new software methods is rapidly lost. To protect R&D investment you would therefore need to pass on more cost to the consumer.
Not flaming here, just asking! I've been a FOSS advocate since forever, and firmly believe that the patent system fails to be relevant to many of the new technologies which have emerged in the last couple of decades. On the other hand I can see that investment in new, innovative ideas needs to be protected.
@9 - if an innovation is small, so that competitive advantage is lost because competitors can duplicate the technology quickly, then the investment was likewise small -- thus society should not be prevented from reinventing such an obvious thing.
But if investment in ideas is required, then clearly those ideas are not trivial or obvious, so competitors would likewise need to invest substantial sums in order to reinvent them.
I don't pretend to know the answer to the best possible balance of monopoly and free market, but I do know that the current system is not it. Patents were not meant to be bludgeons ensuring that the company with the best-paid lawyers got to keep all the money.
Software patents are bad primarily for one reason: they stifle innovation.
If software patents are allowed it may become impossible to write software unless you work with a huge corporation with a defensive patent portfolio. It will be virtually impossible to write code that doesn't violate at least somebody's patent--and it will be impossible to know if you have because you can't review all the patents--especially since patents are deliberately written to be vague to give the patent holder as much wiggle room to sue anybody as possible, thus even experts often have no idea what the patents really cover.
The purpose of patents is to encourage innovation with a trade off. In exchange for revealing your secret technology in public you get the temporary right to a monopoly of your invention. The idea is that the publication of your technology will allow others to build on it once the patent expires. With the software industry ***we already have tremendous innovation*** patents are not needed to encourage innovation as proved by all the innovation that has already occurred. Software patents would do the opposite of what they are intended to do.
Patents are not a right, they are a privilege--a privilege that can only be extended you you in the interest of *advancing* the useful arts and sciences. Software patents do not do that and you do not have an inherent right to a monopoly on ideas.
@#2:
Computer programming is a branch of mathematics, and one of the core Enlightenment principles of a free/open society is the unrestricted use of algorithms as they are discovered.
Mathematical algorithms are considered to already "exist" independently in nature until they are discovered, not so much "created" by people.
This is what software patents boil down to: Who has access to a mathematical concept, and what kind of cost and litigation people must endure to get it.
I would argue that software patents are far less damaging than permitting copyright to extend to object code alone. The price that the patent system demands for the issuance of patent is that the inventor publish lengthy discussion of how the invention operates and the best modes to operate it. Once the patent application is published, everybody is free to learn from it, improve on it, and, if it never issues as a patent, use it however they want. In this way, the universe of globally available knowledge expands.
By contrast, the copyright system makes no sense with regard to object code alone. Software object code behaves like a machine. Its main goal is to perform functions for the user. Sure, there is some tangential expression of ideas as to how the software looks and feels, but no more so than in any other machine -- and people don't copyright machines. Copyright for object code functions like a patent, but with these huge disadvantages to the general public:
(1) A much longer lifespan -- the writers of the code will almost certainly have an enforceable copyright for their entire lives;
(2) No requirement that any part of the source code be made available for public review;
(3) No requirement that the general public be taught how the copyrighted code does what it does
Before people attack (relatively) short-lived software patents, they should first ask themselves "Is our time better spent making sure that the public gets some benefit from the monopoly it grants over use of object code?" I think the answer is yes.
In response to the idea that "we already have enormous innovation" in the software field, of course we do. But this is not proof that patents do not drive innovation. Rather, this innovation is largely because software object code gets copyright protection that is far more powerful (see DMCA notices for example) and far longer lasting than patent protection.
I would propose reforming software copyright law so that a copyright to object code only be available where the copyright holder publishes the source code. Since object code copyright functions in all practical ways like a super-patent, it should at least come with the same "share the knowledge" requirements that a patent has.
I'm a software patent holder from back in the days before I became enlightened. I recently have been struggling with the temptation to get involved in the software patent process again, because a lot of the work my company does involves data analysis in genomics, and I recently came up with a novel algorithm that has yielded excellent results from previously recalcitrant datasets.
The pros are nominal: with a patent, I have the chance to sell or license the technology, or get investment in my company from any VC's foolish enough to still go down that particular path. The odds of being able to realize any of those gains are long. I've worked for a number of startup companies that had patented software, and they are all gone now, in part because they could not survive in an environment where patent costs are so high, the process is so long, and the rewards are so problematic.
The cons are unequivocal: patents are not cheap. The costs can easily run to $25k, and the process can take three to five years. For a small company, that may as well be forever.
The benefits are dubious: a patent is nothing more than a license to sue. To enforce a patent you have to have pockets deep enough to make a plausible threat. Patent actions that go to court typically cost $3 to $5 million. It's true that most companies are relatively scrupulous about playing fair with other people's patents, but it only takes one to ruin your whole day.
Furthermore, most software patents are easily circumventable. It's a very rare case where you can define an algorithm so narrowly as to be useful, but broadly enough to cover all the minor tweaks and twiddles that could be used to achieve similar results.
On the other hand, the very idea of "equivalence" between two algorithms is problematic, which leads to what amount to multiple patents covering very nearly identical ground. The field that my one software patent is in (image guided surgery) is well-known for having a patent landscape that's an absolute mess. This has been solved by larger companies cross-licensing everything from each other, which effectively excludes smaller players. So much for innovation.
Patenting is a game played by inventors and won by lawyers.
As such, I'd be delighted if SCotUS strikes down or otherwise radically limits the scope of software patents. It would simplify my life enormously, reduce the risk that anything I do violates anyone else's patents, and remove the temptation to waste my own time on patent pursuit rather than actually making money analysing data and inventing new algorithms.
Skep - How do patents "stifle innovation"? I've heard this assertion many times but not sure I understand the basis. A patent excludes others from making, using or selling what is claimed. Unless the patent covers a tool or process required to innovation (e.g. a microscope) or you need to built it to improve (rare, but even here not likely to be found liable or sued if solely to improve), its unclear how any patent can prevent anyone from further innovating. You are free, and in fact encouraged, to look at a patent and think of ways to improve what's claimed in the patent ("I think that invention is interesting, but if also do x, y and z it will be even better"). In fact, patentees are required to describe their invention to enable others to make it and also disclose the "best mode" of the invention (sometimes something to consider when deciding between patent and trade secret protection). In fact, under our current system, your application will be published even if you never get a patent.
In short, I understand how patents stifle commercialization of technology and how invalid patents can damage any economy, but not sure how they stifle innovation since you are always free to further invent.
I believe software patents are why we've been stuck with Windows for 23 years, and will continue to be for eternity.
"Skep - How do patents "stifle innovation"?"
First off, patents are only for the purpose of advancing the useful arts and sciences. They are not for the purpose of maximizing corporate, or even individual, profits. They are supposed to advance the arts and sciences in two ways. First, by getting inventors to publish their ideas instead of keeping them secret and in exchange by giving the inventor a limited monopoly. The publication allows future inventors to build off the idea and the monopoly means that inventors have temporary exclusive rights to their invention--this encourages people to spend money and time creating inventions because they can exploit them exclusively.
Software patents are unnecessary. Innovation in software occurs without patents and there is no evidence that people will stop innovating software if there are no software patents.
Next, software patents are written not to be open so that other can exploit them when they expire--part of the deal in a patent--instead they are written to be as vague and inscrutable as possible so that the patent holder can sue for all sorts of people. That means that you as a software developer have no idea what is or isn't already patented--unlike copyright, independent invention doesn't get you off the hook in patents. You can't spend your entire life trying to cross check each line of code and innovation against the patent list. And doing so could even hurt you since you could be accused of knowingly violating a patent.
Big companies can afford to create Mutually Assured Destruction portfolios of patents, so that if someone sues them they can sue back on a different patent. You, on the other-hand, could be sued at any time for anything in your software. This means that you may just get out of the business entirely--I'd say that would count as stifling innovation.
Whether or not software patents stifle or promote innovation seems to me to be an empirical question. If innovation is the actual concern (and not simply another case of corporate greed), then we should test it by implementing a trial period of no patents. Say, a year or two or five.
If at the end of the trial, things are significantly worse (or if a crisis erupts in the software industry before it finishes), then we should implement a reformed patent program. Otherwise, we are better off in not trying to treat ideas as property.
(One of the advantages of federalism seemingly forgotten in our age of increasing centralization is the way that individual states can serve as laboratories for economic and social policies. Instead of letting ideology/rhetoric and wealth/influence decide these important matters, we should use experimental methods and scientific reasoning. Unfortunately, due to the nature of intellectual property, in this case I don't think an experiment would be possible except on a national level.)
@ #17:
The software industry already has such experience: Nearly all of the software innovations up to the early 1990s were patent-free. That happens to cover most every significant development from spreadsheets to databases to kernels, various UIs, networking stacks, public key cryptography and the web browser.
What has changed most about PC software since software patents came into vogue is that minicomputer-class algorithms and interfaces have migrated down to the micro level, producing PCs that are more stable and feature-rich (ultimately thanks to better hardware). The other big change has been in communications links themselves - the arrival of robust and fast Internet hardware.
So I'll venture a guess that most of what software patents "protect" are the mundane and obvious algorithms, creatively phrased to appear unique. Where the "advancement" of the art is in that state of affairs may barely be worth mentioning.
To play devil's advocate, you're only going to become a target for patent lawsuits if you have a pocketbook worth defending. Your average Joe writing code and starting a small business isn't going to be sued by patent holders, for the price tag reasons mentioned by #14. It's only when you're making enough money to make it worthwhile (or when you sue someone else for violating your own patent) that you should expect (counter)suits, and by then, you should, as a cost of doing business, have a legal team and a policy about how to handle (cross)licensing.
What I think it boils down to is that the patent system has ingrained issues and limitations that are especially exposed by software patents. Unlike the engine from #2's comments (rather than the plans), it is relatively easy and inexpensive to reproduce computer systems, which puts the tools used to "fiddle" out a new process into the hands of a great many more people. It's also easier to compose software technologies, so the quality of "obvious to try" increases with software. Furthermore, the descriptions in patents are far from easy to determine whether it applies to your in-parallel invention -- the proliferation of software patents has made the amount of work required to review so large as to be intractable, not simply because there are more patents to review but because the governing body is not capable of policing submissions sufficiently for identity, overlap, and ambiguity, and thus the quality of patents has decreased in the software realm.
Patent "griefers," who keep their patents for the sole purpose of preventing others from taking advantage of the use of the underlying art or science, while limiting the boon that the patents can give in the short term, still improve the public domain upon termination of the patent, at least insofar as the patent's description is worthwhile.
There is still one bastion for those fans of monopoly of ideas: the trade secret. Unfortunately, in the software realm, those secrets are nigh useless -- either they remain in the heads of the developer, or, if applied, are instantly exposed to reverse-engineering, and directly proportional to their perceived value will they be researched and duplicated. Eliminating software patents will probably leave very little protection to the one-person developer, and will stifle their idea production, save where they were willing to give it into the public domain anyway. (Let's hope they have a day job or patron.)
Whatever overhaul of patents happen, it must give incentive for people to publicize (ideally useful) ideas, while not hampering the efforts of the remaining populace to create in-parallel. #15's freedom to further invent can get stuck when your livelihood is jeapardy due to patent lawsuit; your family's next meal might be more important. Greater requirements of specificity, greater bar of inobviousness, especially around composition of previously-existing work, probably can help the system.
One thought experiment: If a patent application was held for a term (perhaps a year) before being released, and several other applicants came up with the same idea in parallel, then a co-patent (optionally with the first applicant getting some special rights) might be introduced -- a shorter lived patent-in-common. Not sure how licensing issues would be arranged.