3 ms·

Groklaw on the Bilksi decision

ovi256 16y ago
I'm now waiting for the HN grok of the Groklaw discussion of the Supreme Court opinion :P /me goes back to reading.
10ren 16y ago
I've been working on product development and "customer development" recently, and this has emphasized for me the differences they have in perspective. The first (engineering) is about features: how to implement something, what it needs, what would be the cleanest conceptual approach; the second (marketing) is about benefits: what those features mean to someone, is it useful, what can it do, what extra features are needed just to make it work (eg. an editor isn't just about editing: you also need to be able to save your work, even though saving isn't editing). There's a distinction in this case that strikes me as similar: software for its own sake is abstract; software that has a useful application is not abstract. The question then becomes whether the patent as a whole is new and inventive. Just applying an abstract idea to a concrete problem isn't enough; the way it is applied has to be inventive. In legal terms, I think this throws the burden of patentability from "subject matter" to "novel and nonobvious" (Bilsky majority on Flook, emphasis added): > unpatentable under §101, not because it contain[ed] a mathematical algorithm as one component, but because once that algorithm [wa]s assumed to be within the prior art, the application, considered as a whole, contain[ed] no patentable invention. tl;dr you can't patent software, only its doing of something useful EDIT btw, typo in title: s/Bilksi/Bilski/
joe_the_user 16y ago
Well I am disappointed but not surprised that SCOTUS didn't throw out software patents. The patent system has been the foundation of the business processes in many industries. It's natural that a lot of powerful stakeholders expect this to continue into the information age. Despite this, when production in general is going towards information, patents are doomed to become more pathological and more a rent-seeking activity as information processing expands. There is simply no even slightly objective measure for determining whether X algorithm and Y distinct algorithm are "the same". The existence of such a measure is essentially a mathematical impossibility. Among other challenge, a sameness measure would probably have to solve the halting problem (otherwise you substitute an algorithm with the same results for any given patented algorithm). More generally, any "sameness" criteria must really be based on a "close enough" measure and closeness itself is a characteristic of the three dimensional metric space we humans live in. Closeness doesn't exist and can't meaningfully exist in an infinite dimensional space of abstract algorithms. So "sameness" is going based entirely on subjective and arbitrary judgments of juries and judges, who naturally know little of the situation (indeed, ignorance is something of a necessary condition for having an honest opinion on the similarity of algorithms). Thus, software patents give any given patent holder of any algorithm an unlimited terrain to sue people on.
grellas 16y ago
When we stand back and assess what happened here, a few things become evident: 1. The Court did not have to hear this appeal. It did so under a discretionary grant of certiorari. So the question is, what did it intend to accomplish? Well, since it did not arrive at any new test for assessing whether business methods are patentable, and since it simply affirmed the decision of the Federal Circuit on the merits of Bilski's particular claim, the logical answer is that the Court was very uncomfortable with the Federal Circuit's adoption of the so-called machine-or-transformation test as sole test for whether processes are patentable and it ruled in a way to make it clear that this is not to be an exclusive test. 2. Thus, in spite of the hopes of many that the Court was using this as an opportunity to rein in a wildly out-of-control system of business-method and software patents, the Court was in fact aiming to prevent any form of categorical limitation from being put on what or what does constitute patentable subject matter when it comes to processes. In this sense, the Court's ruling was definitive in rejecting any categorical attempt to limit the scope of such patentable subject matter ("Congress plainly contemplated that the patent laws would be given wide scope"). 3. What this means, for good or for bad, is that the Court is not about to hold any form of broad category (e.g., software patents) as being inherently unpatentable. 4. Ergo, there is no radical answer here for those who hoped the Court would strike down software patents. As noted in this piece, however, the Court definitely rejected the other extreme as well by saying that it was not endorsing the Federal Circuit's 1998 State Street decision that had served to open the floodgates to software patents (by ruling that any process can be patented as long as it leads to a "useful, concrete and tangible result"). 5. The Court also explicitly acknowledged the issues that are troubling so many: "It is important to emphasize that the Court today is not commenting on the patentability of any particular invention, let alone holding that any of the above-mentioned technologies from the Information Age should or should not receive patent protection. This Age puts the possibility of innovation in the hands of more people and raises new difficulties for the patent law. With ever more people trying to innovate and thus seeking patent protections for their inventions, the patent law faces a great challenge in striking a balance between protecting invention and not granting monopolies over procedures that others would discover by independent, creative application of general principles. Nothing in this opinion should be read to take a position on where that balance ought to be struck." 6. I read all this as the Court's way of saying, in effect, "we are not about to wade into this mess and solve a problem that fundamentally belongs to Congress and, thus, under the current statutory scheme, we will guide the lower courts only to the extent of telling them to approach the issues case-by-case while not going to the extreme of trying to limit patentable processes through arbitrary tests that Congress did not authorize and while being sensitive that their rulings in the information age might have the unintended consequence of stifling innovation through the grant of improper monopolies - thus, to be vigilant to use all the tools available under the statutory scheme (e.g., other requirements that patents be novel, nonobvious, and fully and particularly described; that they not constitute abstract ideas, etc.) to prevent the system from getting wildly out of control." 7. Thus, one can say that the Court basically upheld the status quo but with a few twists. I know this is profoundly anticlimactic and disappointing for those who were looking for a comprehensive answer to the problems raised by process patents, but I personally don't think the Court had a lot of choice in being bound ultimately to rule on what Congress meant when it enacted a broadly-worded statute (for a technical analysis of the statute, see my prior comment: http://news.ycombinator.com/item?id=1350273 http://news.ycombinator.com/item?id=1350273). 8. Any more fundamental answer will lie with Congress. In that forum, the practical impediments will be large and that is perhaps why so many had hoped the Court could cut through all these without needing to deal with the vested interests who will seek to block reform at every turn.

I'm now waiting for the HN grok of the Groklaw discussion of the Supreme Court opinion :P /me goes back to reading.

10ren 16y ago

I've been working on product development and "customer development" recently, and this has emphasized for me the differences they have in perspective. The first (engineering) is about features: how to implement something, what it needs, what would be the cleanest conceptual approach; the second (marketing) is about benefits: what those features mean to someone, is it useful, what can it do, what extra features are needed just to make it work (eg. an editor isn't just about editing: you also need to be able to save your work, even though saving isn't editing). There's a distinction in this case that strikes me as similar: software for its own sake is abstract; software that has a useful application is not abstract. The question then becomes whether the patent as a whole is new and inventive. Just applying an abstract idea to a concrete problem isn't enough; the way it is applied has to be inventive. In legal terms, I think this throws the burden of patentability from "subject matter" to "novel and nonobvious" (Bilsky majority on Flook, emphasis added): > unpatentable under §101, not because it contain[ed] a mathematical algorithm as one component, but because once that algorithm [wa]s assumed to be within the prior art, the application, considered as a whole, contain[ed] no patentable invention. tl;dr you can't patent software, only its doing of something useful EDIT btw, typo in title: s/Bilksi/Bilski/

joe_the_user 16y ago

Well I am disappointed but not surprised that SCOTUS didn't throw out software patents. The patent system has been the foundation of the business processes in many industries. It's natural that a lot of powerful stakeholders expect this to continue into the information age. Despite this, when production in general is going towards information, patents are doomed to become more pathological and more a rent-seeking activity as information processing expands. There is simply no even slightly objective measure for determining whether X algorithm and Y distinct algorithm are "the same". The existence of such a measure is essentially a mathematical impossibility. Among other challenge, a sameness measure would probably have to solve the halting problem (otherwise you substitute an algorithm with the same results for any given patented algorithm). More generally, any "sameness" criteria must really be based on a "close enough" measure and closeness itself is a characteristic of the three dimensional metric space we humans live in. Closeness doesn't exist and can't meaningfully exist in an infinite dimensional space of abstract algorithms. So "sameness" is going based entirely on subjective and arbitrary judgments of juries and judges, who naturally know little of the situation (indeed, ignorance is something of a necessary condition for having an honest opinion on the similarity of algorithms). Thus, software patents give any given patent holder of any algorithm an unlimited terrain to sue people on.

grellas 16y ago

When we stand back and assess what happened here, a few things become evident: 1. The Court did not have to hear this appeal. It did so under a discretionary grant of certiorari. So the question is, what did it intend to accomplish? Well, since it did not arrive at any new test for assessing whether business methods are patentable, and since it simply affirmed the decision of the Federal Circuit on the merits of Bilski's particular claim, the logical answer is that the Court was very uncomfortable with the Federal Circuit's adoption of the so-called machine-or-transformation test as sole test for whether processes are patentable and it ruled in a way to make it clear that this is not to be an exclusive test. 2. Thus, in spite of the hopes of many that the Court was using this as an opportunity to rein in a wildly out-of-control system of business-method and software patents, the Court was in fact aiming to prevent any form of categorical limitation from being put on what or what does constitute patentable subject matter when it comes to processes. In this sense, the Court's ruling was definitive in rejecting any categorical attempt to limit the scope of such patentable subject matter ("Congress plainly contemplated that the patent laws would be given wide scope"). 3. What this means, for good or for bad, is that the Court is not about to hold any form of broad category (e.g., software patents) as being inherently unpatentable. 4. Ergo, there is no radical answer here for those who hoped the Court would strike down software patents. As noted in this piece, however, the Court definitely rejected the other extreme as well by saying that it was not endorsing the Federal Circuit's 1998 State Street decision that had served to open the floodgates to software patents (by ruling that any process can be patented as long as it leads to a "useful, concrete and tangible result"). 5. The Court also explicitly acknowledged the issues that are troubling so many: "It is important to emphasize that the Court today is not commenting on the patentability of any particular invention, let alone holding that any of the above-mentioned technologies from the Information Age should or should not receive patent protection. This Age puts the possibility of innovation in the hands of more people and raises new difficulties for the patent law. With ever more people trying to innovate and thus seeking patent protections for their inventions, the patent law faces a great challenge in striking a balance between protecting invention and not granting monopolies over procedures that others would discover by independent, creative application of general principles. Nothing in this opinion should be read to take a position on where that balance ought to be struck." 6. I read all this as the Court's way of saying, in effect, "we are not about to wade into this mess and solve a problem that fundamentally belongs to Congress and, thus, under the current statutory scheme, we will guide the lower courts only to the extent of telling them to approach the issues case-by-case while not going to the extreme of trying to limit patentable processes through arbitrary tests that Congress did not authorize and while being sensitive that their rulings in the information age might have the unintended consequence of stifling innovation through the grant of improper monopolies - thus, to be vigilant to use all the tools available under the statutory scheme (e.g., other requirements that patents be novel, nonobvious, and fully and particularly described; that they not constitute abstract ideas, etc.) to prevent the system from getting wildly out of control." 7. Thus, one can say that the Court basically upheld the status quo but with a few twists. I know this is profoundly anticlimactic and disappointing for those who were looking for a comprehensive answer to the problems raised by process patents, but I personally don't think the Court had a lot of choice in being bound ultimately to rule on what Congress meant when it enacted a broadly-worded statute (for a technical analysis of the statute, see my prior comment: http://news.ycombinator.com/item?id=1350273 http://news.ycombinator.com/item?id=1350273). 8. Any more fundamental answer will lie with Congress. In that forum, the practical impediments will be large and that is perhaps why so many had hoped the Court could cut through all these without needing to deal with the vested interests who will seek to block reform at every turn.