"Software lies in the antechamber of patentable invention."
October 3, 2016 12:35 PM   Subscribe

(1) patents constricting the essential channels of online communication run afoul of the First Amendment; and (2) claims directed to software implemented on a generic computer are categorically not eligible for patent.

In Alice v CLS Bank (previously), the Supreme Court provided a mechanism to determine if patents regarding computers are transformative, ruling that an invention that restates existing inventions in terms of computers is not eligible for patent.

The Federal Circuit US Court of Appeals ruled on Friday that several patents owned by noted patent troll Intellectual Ventures(previously) are invalid.

Judge Mayer's concurring opinion goes much farther:
Most of the First Amendment concerns associated with patent protection could be avoided if this court were willing to acknowledge that Alice sounded the death knell for software patents.
...
The central problem with affording patent protection to generically-implemented software is that standard computers have long been ceded to the public domain. See Flook, 437 U.S. at 593 n.15 ("[I]n granting patent rights, the public must not be deprived of any rights that it theretofore freely enjoyed" ...). Because generic computers are ubiquitous and indispensable, in effect the "basic tool[]," Benson, 409 U.S. at 67, of modern life, they are not subject to the patent monopoly. In the section 101 calculus, adding software (which is as abstract as language) to a conventional computer (which rightfully resides in the public domain) results in a patent eligibility score of zero.
...
Software lies in the antechamber of patentable invention. Because generically-implemented software is an "idea" insufficiently linked to any defining physical structure other than a standard computer, it is a precursor to technology rather than technology itself. See Mackay Radio & Tel. Co. v. Radio Corp., 306 U.S. 86, 94 (1939) (“While a scientific truth, or the mathematical expression of it, is not patentable invention, a novel and useful structure created with the aid of knowledge of scientific truth may be.”). It is well past time to return software to its historical dwelling place in the domain of copyright.
Fortune has provided some analysis.
posted by atbash (38 comments total) 14 users marked this as a favorite
 
Judge Mayer's concurring opinion ... is purely an expression of one judge's opinion and does not in any way represent a change in the law.
posted by jedicus at 12:49 PM on October 3, 2016 [6 favorites]


To be clear: the main quote in the FPP is from Mayer's concurrence, not Alice. The bulk of this FPP is about a single judge's expression of their own views, not a statement of what the law is. It doesn't represent a substantial change in the law. The decision in Intellectual Ventures was practically a foregone conclusion in light of Alice, which was decided over two years ago.
posted by jedicus at 12:51 PM on October 3, 2016 [9 favorites]


If you want to get into the weeds a little deeper then: PatentlyO
posted by caddis at 1:14 PM on October 3, 2016 [2 favorites]


Jedicus is right that this is just a single judge's opinion, but it is part of a gradual global movement towards the recognition that software patents are an absolutely terrible idea.

Alice and related decisions are a step forward, but there's still some way to go. What judges are slowly learning (Mayer seems to get it) is that the great power of general-purpose computers means that software patents work completely differently from patents in other areas. In other fields you are supposed to work out a new way of solving a problem, something more than simply putting together a series of conventional steps that together achieve a predictable result, but computers are powerful enough that defining a problem is often enough to define the solution as well, and previously courts have allowed patents which simply state a problem in vague terms and then claim a monopoly on completely predictable and conventional ways of solving it. This achieves nothing of the purpose of the patent system, and virtually guarantees that patented "inventions" will be independently discovered (leading to the problem of patent trolls, among other things).

It's as if pharmaceutical companies could not just get patents on new medicines, but also patent newly discovered syndromes and then claim a monopoly on all future medicines that treat the syndrome. Of course they can't do that, but software patentees have been able to. And it's been a complete mess.

If you want to get into the weeds a little deeper then: PatentlyO

Oh, I can't wait to hear what "anon" thinks about this! (actually I can)
posted by A Thousand Baited Hooks at 1:32 PM on October 3, 2016 [11 favorites]


I also think that it is helpful to read this case (and Mayer's concurrence) in light of MCRO v. Bandai, which was decided by a panel of the same court less than three weeks ago and came to a different conclusion about a different software patent.

This problem right here: but computers are powerful enough that defining a problem is often enough to define the solution as well, and previously courts have allowed patents which simply state a problem in vague terms and then claim a monopoly on completely predictable and conventional ways of solving it.

is what the slow and plodding process of Alice jurisprudence is trying to fix. I think that the results in this IV case and MCRO show that the U.S. courts are getting there. Especially given that Judge Stoll, who upheld software claims in MCRO was also in favor of invalidating bad software claims in this IV case. I think that Mayer's approach has some merit, but as articulated (paraphrasing to: "no generically-implemented software claim is patentable") throws out the baby with the bathwater.
posted by sparklemotion at 1:56 PM on October 3, 2016 [2 favorites]


Judge Mayer's concurring opinion ... is purely an expression of one judge's opinion and does not in any way represent a change in the law.

Don't you have obiter dicta in the US?
posted by Sebmojo at 2:22 PM on October 3, 2016


Question from someone who doesn't really follow this:

If software patents were canceled, or if they were made more "up-for-grabs" or what have you, what kinds of technology would we see more of? What are some of the most coveted software patents out there these days?
posted by rebent at 2:27 PM on October 3, 2016 [1 favorite]


I believe a concurrence is not even obiter dictum. It might function as an invitation to litigants to bring that type of case (and this concurrence might do that too) but by definition obiter dicta discuss matters that were not necessary to the reasoning that supported the decision. So if my "software" patent claims are rejected "because Judge Mayer said so," I'm gonna win that appeal. (Similarly, if I sue somebody on a "software" patent and the infringer says "patent invalid because Judge Mayer said so," I'll win [unless the infringer has some other argument, too.])
posted by spacewrench at 2:31 PM on October 3, 2016 [1 favorite]


rebent, it's not so much specific patents you want to use but can't afford, but rather the fact that patent trolls are armed with frequently vague, overreaching, etc. patents and form a kind of environmental threat. But that's in my neck of the woods - perhaps there's a second person writing software for a living registered on Metafilter who might have another opinion.

In addition, many folks (I'm mildly among them) aren't convinced that software patents have any utility - I honestly can't think of any societally-useful innovation in software that wouldn't have been made if it weren't patentable.
posted by The Gaffer at 2:36 PM on October 3, 2016 [4 favorites]


Re: what kind of technology, it's hard to say. Maybe people would copy and extend previously-patented software ideas, but maybe people wouldn't bother, because Google-Apple-Microsoft could just steal whatever good idea you come up with and implement.

The problem with software patents, IMO, is not the software or the patent, but the dearth of competent programmers on both sides of the patent tug-of-war (both patent attorneys and examiners). If we had better-drafted claims and better examination, none of this would be a problem, and the intended benefits of the patent system would work to encourage innovation here, as well.
posted by spacewrench at 2:36 PM on October 3, 2016 [4 favorites]


societally-useful innovation in software

The problem is, you don't know what will turn out to be societally useful in advance. The patent system encourages people to try stuff (i.e. to devote time, effort and assets to things that may pan out, or may not.) Stuff that didn't work, nobody cares if there's a patent on it. But if you say "no patents at all," then at least some people who would have "tried stuff" may say "why bother? If it _is_ societally useful (and valuable to me), somebody with a bunch of programmers on staff will just copy it." And that person may have been the one whose fiddling with stuff would have yielded the societally-useful idea and initial implementation.
posted by spacewrench at 2:41 PM on October 3, 2016 [4 favorites]


what kinds of technology would we see more of

Hard to say what new tech could/might be there. But one thing you would see is faster adoption of new tech/standards, in a world where patents and licensing issues didn't block things.

HTML5 video is one where patents had a big effect for a while. Some browsers were concerned about using h.264 video decoding because it is protected by patents and requires licensing. So there was a split in the community between some browsers supporting that, and others supporting WebM or other competing compression standards. This requires video sites/providers to provide multiple encodings of each video.

Eventually this worked itself out and h.264 is supported in every major browser now, but it was a huge hassle for a while when HTML5 video was starting out.
posted by thefoxgod at 2:44 PM on October 3, 2016 [2 favorites]


what kinds of technology would we see more of

I work for a small family owned e-commerce company in the midwest. A few years ago we got what was basically a shakedown over a patent for filtering items in a search, such as, show me all the red items that match my keywords. We ignored it and it went away, but there are all sorts of silly things out there like that.

If it hadn't gone away, we almost certainly would have had to pay for a license, or back fees (or both), or whatever else they would have come up with that would have been cheaper than fighting it in court. All of a sudden the patent system becomes "this is why we can't have nice things" for your small business.

For some additional reading, New Egg has been pretty great at actually taking these things to court when they get nastygrams. Imagine if everyone who had an HTTPS enabled site had to pay a license fee, for example.
posted by Nonsteroidal Anti-Inflammatory Drug at 3:00 PM on October 3, 2016 [8 favorites]


It's very difficult to write commercial software without falling afoul of someone's intellectual property, because patents exist for many really trivial things, and also for many things that, while not entirely trivial, can still be independently discovered without a lot of effort. It's possible to violate these patents without realizing you're doing it. Many of these patents could probably be shown to be invalid due to prior art -- for instance, although it might be expired now, somebody once held a patent on the idea of a linked list. This is a really, really basic data structure and the patent should probably never have been issued.

So large corporations like Amazon and Microsoft tend to agree to licensing deals amongst themselves. They use their patent portfolios for leverage in negotiations with each other. It seems like they don't often choose to go after small companies for stuff like this, and small dev shops, in my experience, don't tend to worry that much about patent issues.

But if they wanted to, they could easily choose to use their enormous patent portfolios to extort money from small developers. This strikes me as a big theoretical problem, albeit one that doesn't seem to arise in practice all that frequently.

Then you have patent trolls, which are small companies or individuals that use patents in a predatory way to hassle people into settling out of court. This is the kind of behavior that NSAID described in the comment above.

Then there are cases where a particular piece of technology is patent-restricted and it causes legitimate problems in the industry regarding standardization. The video encoding stuff thefoxgod mentioned is an example of this, and it's also happened in the past with image and audio encoding standards.

So these are all examples of why software patents are bad. But there is one point in their favor, which is that they allow small companies engaged in research and development to realize a profit without engaging in large-scale marketing of commercial software. If a small company develops a really awesome new algorithm that does something very useful, they can acquire a patent on it and sell that patent to (say) Google. Google may choose to acquire the entire company to get access to the patent. This means that private researchers can: 1.) freely publish their results in peer-reviewed journals, which will allow the state of the art to advance faster than if these things remained trade secrets; and 2.) make money by doing research without having to turn their research into a product, which is really a totally different can of worms and imposes all kinds of additional burdens on the company in terms of development, testing, marketing, and so forth.

So I don't really know where I stand on software patents. Intellectually, they seem like a bullshit idea, and they clearly have bad effects. But I don't like the idea of good fundamental advances in computing being hidden as trade secrets because private researchers are disincentivized from publishing their work.
posted by a mirror and an encyclopedia at 3:15 PM on October 3, 2016 [9 favorites]


in a world where patents and licensing issues didn't block things. HTML5 video is one where patents had a big effect for a while. Some browsers were concerned about using h.264 video decoding because it is protected by patents and requires licensing

The first public working draft of HTML5 was published c. 2008. MPEG LA made h.264 royalty free for websites in 2010. As far as I know MPEG LA never blocked or threatened any software projects or sites for their use of h.264 video.

There is zero need to categorically exclude software-implemented inventions from patentability. The problem lies entirely with poorly specified, overly broad patents, which is a problem that can exist in any area of technology. There is nothing inherently unique or special about software in this regard. There are reasons* why patents on software-implemented inventions tend to be more problematic than most, but they aren't inherent to software.

* The USPTO dragged its heels on allowing patents on software-implemented inventions, leading to an examiner corps that was woefully underequipped to examine applications properly. The courts took too long to make it clear to the USPTO that patents that take a known invention and add "on the internet" or "on the computer" or "on a phone" are obvious. The USPTO's guidelines on software-implemented inventions favor things like flowcharts rather than code or pseudocode, making software patents difficult to search and allowing applicants to submit underspecified applications that don't actually provide a concrete solution to the problem they purport to address. I could go on, but the point is that those were all policy choices, not some inherent incompatibility between software and patents.
posted by jedicus at 3:20 PM on October 3, 2016 [4 favorites]


The first public working draft of HTML5 was published c. 2008. MPEG LA made h.264 royalty free for websites in 2010. As far as I know MPEG LA never blocked or threatened any software projects or sites for their use of h.264 video.

I was pretty involved in the rollout of codec support on at least one major browser, and it absolutely was an issue for Chrome, Opera, and Firefox at various points. They were very concerned about licensing related issues. The effect was, however, more one of FUD than direct action --- you are certainly right that no action was taken by MPEG LA. However, the threat of it kept these browsers from quickly implementing (Firefox and Opera held out for a while, and went with a WEBM only strategy at first for this reason).

MPEG LA made it royalty free for websites but _not_ for browsers, which was the whole issue.
posted by thefoxgod at 3:30 PM on October 3, 2016 [5 favorites]


I'm having a hard time with the legal argument. For instance, a spam-catcher program is an "abstract idea" already in use that is applied to computers. So:
[... ] it was long-prevalent practice for people receiving paper mail to look at an envelope and discard certain letters, without opening them, from sources from which they did not wish to receive mail based on characteristics of the mail. The list of relevant characteristics could be kept in a person’s head. Characterizing e-mail based on a known list of identifiers is no less abstract.
It seems to me that spam filters use a variety of mechanisms -- Bayesian or not or a mix -- and I doubt the analogy to "relevant characteristics" kept in someone's head.
But IANAL...
posted by CCBC at 4:20 PM on October 3, 2016


I doubt the analogy to "relevant characteristics" kept in someone's head
It's not relevant... It's an example of people who don't know software, having power over software. But it'll fuck up the virtuous and the idiotic in equal measure. Unfortunately.
posted by spacewrench at 6:06 PM on October 3, 2016 [2 favorites]


Judge Mayer's concurring opinion ... is purely an expression of one judge's opinion and does not in any way represent a change in the law.

I believe a concurrence is not even obiter dictum. It might function as an invitation to litigants to bring that type of case (and this concurrence might do that too) but by definition obiter dicta discuss matters that were not necessary to the reasoning that supported the decision.

To back this up, the common law is made of judges opinions. Where the opinion is the main one that decides the case it's called the ratio decidendi, 'reason for the decision'. Where it's sort of off topic musing it's called obiter dicta, 'other stuff they said'.

In this case if the judge decided that software patents are bunk and this is a core part of his reasoning, then it's a ratio decidendi and every lower court is bound to follow that decision where the facts are the same, or very similar. Superior courts don't have to care, but will need to be convinced the lower court was wrong.

If only one of the judges held that view then that's less convincing than if all of them did, but nonetheless a judge's opinion in deciding a case very much does change the common law (as distinct from statute law, which is made by Congress or Parliament or w/e).
posted by Sebmojo at 6:33 PM on October 3, 2016


Sure you're not misstating things? Concurring opinions are not binding precedent, at all. They don't have any impact on the common law. (From a purely technical standpoint, anyway)
posted by naju at 7:24 PM on October 3, 2016 [1 favorite]


It sounds like we're going to end up with a world where software is like fashion and has very little IP protection. You come up with an innovative new bit of code and someone can just reimplement it -- knocking off your design to get around you copyright and you can't patent it. I'm ambivilant in that on the one hand I do a lot of open source stuff and have seen a lot of ugly patents. On the other hand I've seen how frustrating it is to build an App, get some traction and then see clones and look alike apps flood the App Stores killing your profits.
posted by humanfont at 7:38 PM on October 3, 2016


a world where software is like fashion and has very little IP protection
I hope not, and it won't happen on the basis of this opinion (though there are plenty of software know-nothing judges who could jump on the bandwagon.)

The open-source thing is a concern, but the fact that people are willing to copy a great idea "for free" should not affect what the originator of the great idea can do with her idea. If open-source people are really being innovative, then their open-sourced code should be the prior art that knocks out patent attempts by dirty trolls. (And if examiners and attorneys could read source code, that might happen.) But if the "dirty troll" thought of (and implemented?) the idea first, then the open source copyists are no better than the movie pirates.

(My pet idea for fixing software patents is to require the submission of a working implementation. If your implementation doesn't satisfy some claim, then that claim is invalid. One of the problems with the patent system is that technical ideas must be translated into English for examination, but we already have a better languages than English for describing software; we should use it them!)
posted by spacewrench at 7:51 PM on October 3, 2016


Concurring opinions are not binding precedent, at all

Wait, doubting my own take on this now. It was a three-judge panel, with one opinion for the court, one concurring opinion, and one dissenting opinion. So 1-1-1 splinter. As such there is no "majority" precedential opinion and any court looking at this case would have to muck through and examine the opinions of the individual judges, without any clear-cut takeaway. Is that right...
posted by naju at 7:55 PM on October 3, 2016


@naju: One opinion was clearly entered as "the opinion of the court." Without reading through the whole thing I'd assume Dyk (who wrote that opinion) used a narrower rationale than Mayer (who wrote the excerpt quoted here). So Mayer chose to concur, and everything Dyk said the force of binding law. Where Mayer went further than Dyk, that has no precedential value.

IANAL though.
posted by mark k at 8:00 PM on October 3, 2016 [1 favorite]


In chemistry researchers read patents because it helps them innovate. They learn about their subject. The published secrets have competitive value but they are disclosed because they are protected.

Everything I've heard is that in software practitioners are told not to read patents to avoid the charge that they "knowingly" violated them. The expectation is that if you are working on something you might come up with it on your own and infringe uninentionally, or draw off something public (like "rounded corners") that you wouldn't think patentable, and a lot of patents are ridiculous, so it's best to leave patents to lawyers. The main impact is to restrict research, not encourage it.

This doesn't quite prove the whole ideas of software patents are ridiculous--you could make a fairness argument--but I do think they are pretty far removed from the point of the patent system. Especially as they are used in practice. Copyrights seem a better choice.
posted by mark k at 8:06 PM on October 3, 2016 [3 favorites]


The problem with software is that the standard for non-obviousness is roughly speaking: "this isn't in a textbook". This encourages the filing of trivial patents, which do not encourage progress in the useful arts. No one would ever look through software patents for ideas on how to solve a problem. This is not true for fields like chemistry or drug development, where existing patents are a useful record of what other people have done.
posted by monotreme at 9:17 PM on October 3, 2016 [3 favorites]


I don't know that software patents had to have been a bad thing, if the legal system had understood software better. Because of the history in practice and the culture of software as it exists though, I think the majority of software people I know would gladly give up the option of ever getting a patent in exchange for being able to just write code without being concerned with other people's patents. (I'm sure agreement on that point is not universal but I think it's not an uncommon sentiment.)
posted by atoxyl at 10:42 PM on October 3, 2016 [1 favorite]


The ineptitude of the USPTO is the source, ultimately, of much wailing and gnashing of teeth with respect to software patents. The application reviewers often lack the technical skills to properly evaluate the applications, and are swamped, so tend to rubber-stamp approve them and let litigation sort it out. This amounts to just passing the buck, but I understand why they do it.
posted by axiom at 11:26 PM on October 3, 2016


Wait, doubting my own take on this now. It was a three-judge panel, with one opinion for the court, one concurring opinion, and one dissenting opinion. So 1-1-1 splinter. As such there is no "majority" precedential opinion and any court looking at this case would have to muck through and examine the opinions of the individual judges, without any clear-cut takeaway. Is that right...

I'd think it would be persuasive but not have binding authority, but while IAAL, IANAUSL so imma just link wikipedia and shut up now.
posted by Sebmojo at 11:29 PM on October 3, 2016


The thing about software is that it tends to consist of many ideas assembled in a specific, very complex way to some end, with the value of a particular piece of software lying in what it does and the quality of the overall implementation.

To my mind patenting algorithms and data structures comes dangerously close to patenting mathematical facts - and anyway in practice due to the number of such pieces involved in any complex program and their abstract nature it can be very difficult to keep track of which are safe to use if some are not. Software development shouldn't be a minefield - especially because, again, the value really comes from the particular assemblage. The particular assemblage - the exact implementation - is protected by copyright, as it should be. That leaves the actual function of the software. But of course you can't patent "[every solution] to [specific problem]." You have to have "[specific solution] to [specific problem]" - but that takes us back to what part of a specific solution, other than the exact solution, constitutes a patentable solution?

(sorry if this doesn't make sense I'm kind of trying to unpack an intuitive set of concerns which I ended up having a hard time articulating)
posted by atoxyl at 11:47 PM on October 3, 2016


You could probably argue a lot of that isn't actually unique to software and hasn't presented an insurmountable obstacle outside of software. Even if there is a difference of degree somebody could perhaps draw a reasonable line somewhere. I think more than anything just to me as a software developer educated in the culture of software it doesn't seem to me like we have done poorly with non-patent incentives to innovate methods or features.
posted by atoxyl at 11:59 PM on October 3, 2016


There is zero need to categorically exclude software-implemented inventions from patentability. The problem lies entirely with poorly specified, overly broad patents, which is a problem that can exist in any area of technology.

I couldn't possibly disagree with this more strongly. There are a lot of problems here, and that is but one of them. Among them:
  • It's completely impossible to actually look up if your software would violate any patents. If you try to do this, your lawyers come back with a list of patents too large to search through in any meaningful way, and you spend large amounts of time answering abstract questions about your software. Then they ask a lot of questions where they read a line from some patent (that you're told not to read) and ask "does that clause read on this implementation?" in really oblique ways. It's not the best way to spend weeks of your engineering time, especially since there's no reason to believe you've found everything even if you do all this. It's expensive and completely ineffective. Because of this, most small companies literally do not even try.
  • Instead of worrying about that, large companies often treat their portfolios as a nuclear arsenal—the largest purpose for them is deterrant in a system of Mutually Assured Destruction. This is why you see things like the Open Invention Network, which pools companies' patents in order to build a bigger arsenal than they'd have independently, as well as enforcing a non-aggression pact for licensees of the patents. (full disclosure: some of my software patents are probably assigned to OIN.)
  • Some markets are, as a result, impossible to actually enter unless you've got billions of dollars to spend on either patent defense or building up your MAD portfolio. This is why Google bought Motorola for $12.5B and sold it without the patents 3 years later for $2.91B. (Many, but not all, of these were software patents.)
  • The RSA cryptosystem was patented in 1977. Until 2000, you couldn't ship web browsers or servers that had https without paying RSA for a license, and RSA would only sell licenses for its closed-source implementation. This is a completely well specified, specific patent on a software algorithm, and it actually wound up holding back web security for years. Even if there's a legitimate need for software patents (which I doubt), the term for patents in a highly competitive, quickly developing technology market is much too long.
  • The FAT filesystem was introduced in 1977. For various reasons its implementation changed over time to a thing called FAT32, to support newer hardware, and (on top of that as sort of an optional part) VFAT, to support longer file names. These patents are essentially required for compatibility with things like SD Cards, and so if you're making a device like a GPS with SD Card map updates or a digital camera, you really don't have a choice but to implement it. Microsoft has been litigious about FAT, and following their success in that suit has taken approaching device vendors with threats citing that case. In several cases these vendors have waited for all the previously known patents to expire before they implement this, and Microsoft shows up with a new (old) patent to threaten suit with. Since these cases usually end in sealed settlements, which patents actually get brought to court is not widely known, nor is whether or not an implementation will violate them. In this case it's also true that the patents should be offered on RAND terms, but that doesn't help if there's no way to look up what it is you need to license. It's completely nonsensical—and this is for 1970s technology!
  • In some cases (FAT again) where the patent is known, software has been changed so that the patent doesn't read on it, without actually breaking compatibility. This is possible because the patents are narrowly defined. But the result is just that future suits are over a different claim or another patent covering a different aspect of the same thing. And in these cases it's not even really an algorithm: they're phrased as algorithms, but they're really about the structure of the resulting data.
You could argue that this all spells out the need for reform, and that software patents are fundamentally sound. I don't think that's it. Companies pay good bonus money for filing patents on "inventions" they never expect to implement, whether they would be valid under some better criteria or not. There's no end of lawyers fees, and that wouldn't change by making the patents less broad. There is not any method to look up what patents a piece of software might be violating, and rigorous writing and vetting isn't going to help that. In the general case, it's literally an intractable problem—ironically it's an (admittedly strange) form of one of the oldest problems in computing. For a very small number of patents, you might be able to approximate an answer, but that's not what we've got. What we've got is something like 100k software patents per year being granted. If even 10% of these are what you'd think are legitimate, that's still enough to make this task impossibly expensive for most software authors. The ones who can afford it would ultimately still not try - it's easier to do MAD.

All this translates to not only a high expense, but no possible way to determine appropriate bounds for the expense, and no way to evaluate the risk of picking an arbitrary limit. The result is that companies spend a disproportionate amount of money on it, instead of on actually creating things. So it's a pretty big pile of sand in the wheels of economic growth.
posted by atbash at 8:26 AM on October 4, 2016 [12 favorites]


Yup. Burn it all down. It does no good and a hell of a lot of harm.
posted by Artw at 10:34 AM on October 4, 2016 [1 favorite]


We should abolish patents entirely though. It's plausible they provided a social good 100 years ago, but the record seems spotty there. It all just moves too fast today. And the genetic and medical patents create worse problems than the software patents.

Investors will sill flock to high tech industries even without patents because those high tech industries are opportunities to (a) make real money and power, and (b) feel like you actually do something useful.

We should also add some sort of "significant technological advancement award" with say a few thousand annual recipients that guarantees the recipients a salary for themselves and a few assistants for say 20 years, maybe with some requirement to publish an annual statement of what they do with the money. In essence, anyone who invents anything significant gets a cash flow sufficient to float a very small company to develop it.

Ain't likely such government investment can replace the crazy pants in private investment, but it'll add safety net for some interesting ideas that take way longer than investors can stomach.

In any case, software should not be covered by patents obviously. In fact though, compiler output should not directly be covered by copyright. We'd all fare far better if only source code were covered by copyright and a derived copyright applied to compiler output only if the source code were distributed too.
posted by jeffburdges at 6:53 PM on October 4, 2016


Could I solicit a bit more detail to the devil's advocate view? I mean in the vein of comments saying baby/bathwater and that software patents could have been done smartly and not led to the mess. I intuitively agree with that, but can't make the argument, so what I'd like to solicit is a nice collection of good, unarguably valid, society-positive existing patents.
posted by spbmp at 4:07 PM on October 5, 2016


Just a few reasons why patents have failed : Technology advanced beyond where where investors can foresee the payoffs via products, while information technology gave investors quicker and easier investment opportunities that do not involve taking risks on new technologies. And this longer path to market ensures that (a) more patents on smaller parts inhibit technological progress elsewhere, while (b) few patents on the larger parts can more easily be circumvented. Patents were a legal tool for the 1800s, not the 2000s. And parasites have now firmly attached themselves, like many legal tools.

We might consider a scheme of "national patents" where only nations could claim ownership of patents and only on work produced by their own nationally funded research institutions. These nations would attempt to do cross licensing on behalf of their domestic industries, but licensing agreement failures would result at worst in extracting national licenses through tariffs, not forcing companies to pay. The goal would simply be to provide an advantage in negotiating trade treaties to nations who provide more funding for research and development.
posted by jeffburdges at 9:54 AM on October 6, 2016 [1 favorite]


Patents were a legal tool for the 1800s, not the 2000s.

This is a post that seems mostly informed by observing software patents extrapolating across industries. Patents have different effects in different industries, and the relationship between what can be sold and what can be patented is quite different as well.
posted by mark k at 7:53 AM on October 7, 2016


"defining a problem is often enough to define the solution as well"

This is beautifully said and goes right to the heart of the problem.

The thing is this: there is a reason software is written in what is called a language. It is because IT IS!

Write the shortest sentence to inform someone the sun is setting.

The answers are limited. And unpatentable. Hell, even uncopywitable!

Now write the shortest 'sentence'/block of code in a langauge to transpose a two digit number.

Same difference.

Software is written in a language. Language is un-patentable, but does fall under copyright.

Code is the same. Anyone who does not understand that just has not coded enough and should be disqualified from the discussion.
posted by MacD at 8:15 PM on October 11, 2016 [1 favorite]


« Older Notes from the Lambda Complex   |   Rosa Parks' former home kidnapped/saved... Newer »


This thread has been archived and is closed to new comments