To the surprise of nobody who was actually familiar with the underlying law and precedent, the judge in the Oracle-vs.-Google mega-lawsuit ruled today that Oracle’s claim of copyright protection on the Java APIs is contrary to law.
This means Oracle’s claims against Google are toast. Their best case is now that they’ll get $300K in statutory damages for two technical copyright violations, almost noise compared to what Oracle spent in legal fees. The patent claims went just as thoroughly nowhere as I predicted back when the lawsuit was launched.
In a previous post I speculated that this lawsuit might signal the beginning of the end of the patent wars, as business managers wake up to the fact that IP litigation is usually a spectacularly expensive way to accomplish nothing. Today’s ruling, though it’s about copyrights, increases these odds. Here, as in the SCO lawsuit, copyright issues were pushed harder as what was initially a patent case failed.
My headline is admittedly a bit of an exaggeration; unlike SCO, Oracle still has a viable business to run. But the history and outcome of this lawsuit – huge stakes, shifting claims, apocalyptic press coverage all leading to a spectacular and humiliating debacle – certainly recalls the SCO lawsuit. It seems unlikely that many more of these will be required before corporate America gets the lesson.
I want the patent on the business practice of holding patents to sue other patent holders for infringement….
You’re probably right that this will decrease the frequency of inter-mega-corp patent suits, but (as somebody on your last post pointed out) I don’t think it will have any effect at all on a more important and far more damaging and chilling part of the patent wars, which is practice of small innovators getting trolled by big patent holders.
It looks like that phenomenon will continue. In these cases, IP litigation is not ‘a spectacularly expensive way to accomplish nothing’. It consistently accomplishes exactly what its practitioners intend it to accomplish.
Ken, unfortunately, I think there is way too much prior art for that one to hold up.
The biggest casualty in this patent troll suit I think is Apache Harmony, as the Apache Foundation has canceled it due to licensing issues.
I guess part of it lives on in Android, though.
‘It seems unlikely that many more of these will be required before corporate America gets the lesson.’
I’d like to nominate esr for some sort of optimism award.
Problem is, to the corporates, these lawsuits are the equivalent of some penniless schmuck that buys a lottery ticket … thinking that the payoff is potentially huge and then rationalizing that it didn’t really cost that much. Rinse and repeat.
When the, IMHO, two biggest trollers of all – MS and Apple – begin to stand down, then I’ll be glad to admit esr was right. I just don’t see it happening.
Actually, I’m somewhat optimistic, too. Yes, it will get worse before it gets better, but this was a fairly major skirmish in the war.
And we can help.
When the alleged “experts” like John. C Mitchell of Stanford University get up on the witness stand and lie about things like how “runtime” basically means “when the phone is turned on” and how simple instruction pattern matching is the same as simulated execution, we need to flame and shame them for their mendacity, cupidity, stupidity, and any other “ities” we can think of.
Judge Alsup has gone a long way towards helping us here; the Oracle case, especially the ruling on the JMOL motion on patents, is not a good resume builder for the aforementioned “Doctor” Mitchell:
Now all we need is for Stanford to wake up and realize what an idiot he’s been and that Google has the potential to be at least as big a donor as Oracle. Some censure would come in handy.
But we also need to make it well known that whenever some idiot thinks he will make a name for himself by lying for dollars, we will shame him mercilessly, and if we ever happen to be in a position to affect a hiring decision, he won’t be hired. I think this comment at groklaw is a particularly good start on doing this for the idiot professor John. C Mitchell.
“The patent claims went just as thoroughly nowhere as I predicted back when the lawsuit was launched.”
OK. Multiple pats on your back.
I think your use of the term “predicted” is bullshit, friend.
I was also of a mind to expect that their claims would crash and burn….yet “prediction” was very far from my mind – why? – because of the perverse state of the software patent infrastructure and the nebulous IP law that grants it oxygen.
I was actually horrified at the prospect of this particular case, dreading the consequences of a legal decision that could profoundly alter an economic landscape.
I had *no confidence* in our legal system to be able to exercise the competence necessary to understand how bogus Oracle’s claims were. Zero.
To make such a “prediction”, you would need far greater confidence than I in our legal system…..I don’t perceive you as being that kind of person (quite the opposite, in fact)……hence my application of “bullshit” to your claim of such a prediction.
Yours was not a prediction, it was a valiant principled hope.
And I’m fucking glad the dice rolled our way………for now
>To make such a “prediction”, you would need far greater confidence than I in our legal system
Your general point about the randomness in the process is valid.
However, I had specific reasons for confidence in this particular case that would not generalize to other lawsuits. That is, I have been deeply involved in a lawsuit with a very similar fact pattern in both patent and copyright claims. I am intimately familiar with the key cases (such as Altai) and how the doctrines of merger and scenes a faire are applied to software copyrights, and I’m also familiar with the case law bearing on patentable elements of software.
Accordingly, I actually knew, as a matter of experience and expertise, how weak Oracle’s case was. I wasn’t just guessing or being optimistic – I had the exact knowledge base required to predict the outcome. Could it have gone the other way? Yes…but only Google’s lawyers were incompetent, or the judge was deeply stupid, or Oracle had a handle on the judge. In this case the first two possibilities could be excluded by observing the trial.
I think IBM actually has that patent pending…
Had Oracle gone after anyone significantly smaller than Google, the defendant company would have settled out of court, and everybody in America would start thinking twice before writing something that’s API-compatible with a proprietary platform.
And, of course, had they sued in Europe it would have been over within a couple of weeks and Oracle would have been stuck with the defendant’s litigation fees.
Just because this particular lawsuit was frivolous doesn’t mean that megacorps won’t try — and succeed — to use suits of this type to menace smaller companies. It’s an intrinsic feature of the American legal system.
And the litigation isn’t even over for Android. Hell, Android still hasn’t reached the boss battle yet. That will come when the patent-troll Voltron (Trolltron?) that Apple and Microsoft formed from all that Nortel IP starts filing lawsuits against handset manufacturers, and possibly even Google itself (though that seems less likely).
….and when that storm crosses the horizon, I will be one of those that emphatically demands that their world should burn.
Eric, now that the opinion is available as text, I’d like to call your attention to the judge’s particular phrasing in the introduction. We know he’s got programming experience—let the speculation begin…
>Eric, now that the opinion is available as text
I have read it. I’m going to blog about it.
The real question here is: what does this do to RMS’s assertion that an API created specifically to sever GPL’ed code from non-GPL’ed code is copyright infringement?
Wow, I can’t imagine being this optimistic.
Look at Apple, Samsung and . And that’s before you get to Microsoft’s former CTO’s gig (when he’s not writing cookbooks), or any of the other NPEs that suddenly discover they invented “…”, “… on the internet!”.
IP is going to get far, far worse, to my eye. Larry v. Larry is just a pissing match that went too far.
The patent wars will move from very public lawsuits to backroom blackmail like Microsoft does. Everyone is already paying the Microsoft tax on Android. And now that Nokia sold most of its important patents to a patent troll, while retaining a license for itself and Microsoft this will only get worse.
In fact, the acquisition of Motorola by Google changed nothing. Microsoft is king of hill and has more profits on Android than Google.
The sad thing is that for every, scarce, wins in court against patent trolls and bullies there are thousand of losses of big players against smaller ones, of patent and copyright fundamentalists against reason, of backroom legislation to limit freedom through political corruption against humanity.
As Russel Nelson notes above, what will happen to the GPL. Are companies free to link to some GPL libs in their proprietary software products now?
I could just impement a MIT licensed API stub that does nothing, ant then claim that my product is a derived work of that, not the original GPL lib.
Presumably they couldn’t, as the library as a whole is still copyrighted. But they could re-implement the library with the same API and link to that, without being bound by the original license. If I’m understanding the decision correctly, anyway.
Isn’t that about what google did, and what this fight was about?
Never mind…I think I misunderstood what you were suggesting. I still think it wouldn’t fly, but I’m not clear enough on what actually happens during linking to be sure.
>In fact, the acquisition of Motorola by Google changed nothing. Microsoft is king of hill and has more profits on Android than Google.
As concerning as that is, and as maddening as it is that Microsoft is getting away with shaking down so many Android manufacturers, what I’ve heard is that Google isn’t really trying to profit from Android directly. They’re mostly using it as a way to make sure that Apple (or Microsoft, if Windows phone were doing better) can’t shut down their search business by simply refusing to let their phones access Google. So their acquisition of Motorola means that there’s a phone manufacturer that Microsoft can’t extract its tax from without attacking Google directly. That changes things significantly for both the consumer (who can be certain of finding a phone model that is not under the Microsoft tax if they wish, whereas previously things were less certain), and for Google (who can be certain that at least one Android manufacturer won’t have its strings pulled by Microsoft).
>As Russel Nelson notes above, what will happen to the GPL. Are companies free to link to some GPL libs in their proprietary software products now?
As much of a fan as I am of the GPL, I’ve always thought its linking restrictions were probably unenforceable. (Unfortunately, those restrictions close some fairly significant loopholes that could allow people to make end-runs around the copyleft provisions).
If one takes the linking restrictions at face-value, it would seem to me that you could view a kernel as a special type of shared library linked to programs in a special way, so that writing a program that used Linux-specific system calls could be considered a copyright violation (though I doubt Linus would sue over it).
Bjorn writes: “As Russel Nelson notes above, what will happen to the GPL. Are companies free to link to some GPL libs in their proprietary software products now?
I could just impement a MIT licensed API stub that does nothing, ant then claim that my product is a derived work of that, not the original GPL lib.
I would argue that this changes nothing regarding the status of GPL’d works.
The infringement that results from using a GPL’d work linked to proprietary software is that the resulting program is a derivative work, not that the API itself is outside of copyright. It is therefore a violation of the copyright holder’s exclusive right to create derivative works and outside of the scope of the license granted by the GPL.
SPQR, of course if you link and create a single executable, that’s a single work and is covered by the GPL. But what if you link to a stub API that loads a library dynamically? RMS’s position is that if you create a stub API specifically to work around the GPL, that you are creating a derivative work even though no file has been created which mixes GPL and non-GPL software.
Russell Nelson, its my position that a symbolic link and a dynamic link are the same thing in copyright law. The result is a program that is a derivative work. I know that people think that when I say “derivative work” what is referred to is the actual linkedited program binary. But its my belief that copyright law does not distinguish between them. Just to illustrate, there is Ninth Circuit precedent that says that a running program is a “copy” for the purposes of the copyright act such that just running a program in a manner forbidden by its license is an infringement of copyright (MAI Systems v. Peak Computer).
And many years ago, he claimed that CLISP violated the GPL by merely being able to use readline…
He’s an overreaching zealot.
I thought congress implemented legislation to fix the unjust result in MAI v Peak.
In any case, though, I would agree that it’s a defensible position that when a program is dynamically linked with a library, a derivative work has been created. But it’s also a defensible position that it’s the end-user creating the derivative work. And since the end-user has a valid license to use the GPLed component in any way he desires, there is no copyright violation, so I can’t even be held responsible for indirect copyright violation.
I think the ground is really firm if I don’t distribute the GPLed component. But even if I do distribute it, I’m not distributing a single derivative work. I’m distributing two components, which can be linked together if you choose to do so…
Patrick, the end result, that running a diagnostic program outside of the scope of the license (e.g., by the third party vendor of service for the hardware ) was infringement was sort of the justification for an amendment to the Copyright Act regarding rights of an owner of a copy of software – but arguably they failed to actually fix the problem. But the holding that running a program is itself creating a copy is undisturbed.
I don’t agree with your position that dynamically linked programs do not create a derivative work in the resulting running program under the Copyright Act. I’ve got another case citation, unfortunately it is buried in some notes that I’ll have to go dig up from a CLE I taught on Open Source some years ago. It involved software that dynamically pulled up code from another company’s DB product and ran it. It was found to be infringing. You’ll have to forgive me for not having the case cite handy, I’ll see if I can find it.
I didn’t finish my thought above. To continue:
Patrick, please take this with the understanding of my respect for you. As a software guy myself, I would point out that you are looking at a legal question – what is a derivative work under the Copyright Act – and trying to fit an engineering answer. I understand this, its how we want to address that question. But its not an engineering question, its a legal question. And I think my position addresses it as the courts would.
Believe me, I understand there’s a huge difference between legal and engineering viewpoints, and I appreciate your thoughts on the matter. It would be interesting to see if the case you mention is on all fours with the case of GPL vs. non-GPL software. I can see lots of almost-but-not-quite possibilities. As you mentioned, the congressional fix didn’t quite fix the problem, partly for the reason that most software is licensed rather than sold.
But I’m explicitly concerned with the GPL. RMS and the FSF have, I think, attempted to achieve the unachievable with the license. They want maximal freedom for the end-users of the software, but restrictions on developers to force them to license their own creations under the GPL. I could be wrong, but I think this is where it falls apart — you can’t have maximal freedom without having maximal freedom.
To wit, the license for the GPL explicitly allows (1) a developer to use GPL-licensed software in-house, which should cover his testing of his own program, and (2) the developer’s end-user to use the software however he wants, which should cover linking it (dynamically or statically) with the developer’s software.
So it would be really interesting to see a ruling that would be applicable to the GPL. The developer is explicitly allowed to run the GPLed software as long as he is not distributing modified versions, and (especially with the recent Alsup ruling) writing code that is capable of invoking GPLed software through an API shouldn’t infringe. If the end-user was infringing, then the developer would certainly be liable for contributory infringement, but the GPL explicitly allows the end-user to run the GPLed software as well. Restrictions are on “modifying” and “conveying”.
I must confess that GPL v 3 is much more confusing than GPL v2 on some of these points, and the last time I studied this deeply was with v2. There might be something in v3 that covers this, but it seems like a difficult, if not impossible, problem for them to try to solve.
For me this is mostly academic. I do contribute to the rst2pdf project, which is under a permissive license, but which can, in a few cases, use GPLed libraries if the user has them installed. Nothing bad is going to happen because the permissive license is compatible with the GPL, but I can forsee cases where someone writes a proprietary program that rst2pdf and thinks they are “safe” but invokes it in such a way that it uses a GPLed licensed library…
Patrick, I do not think that Alsup’s opinion stands for the proposition that one can invoke another’s software “through” an API. Alsup’s opinion stands solely for the proposition that the API itself, and those lines that define the API interfaces, are protected by copyright.
The case I was thinking was closest (the last time I did the research on the topic was five or six years ago – working on a CLE presentation on Open Source licensing issues) to holding that a program using a dynamically linked library created a derivative work of the library is Dun & Bradstreet vs. Grace Consulting, 307 F.3d 197 (3rd Cir. 2002). Around page 213, the opinion discusses where the defendant company’s software calls functions out of the plaintiff’s software to access customer data from the plaintiff’s software’s database. It finds that it is infringing and appears to base its holding on the exclusive right to create a derivative work.
In the Ninth Circuit, Microstar vs. Formgen, 154 F.3d 1107 (1998), in a different analogous situation, a company was distributing its own work – new levels of the Duke Nukem game – without distributing the Duke Nukem game itself. This was viewed as creating a derivative work. In this analogy, the court is not looking at the software itself as object code. It is really looking at the resulting “work” that is the game in operation – viewing software more as an audiovisual work than the plain literary work analogue of source and object code. The new level created a derivative work when ran with the copyrighted software.
So I think that the FSF’s FAQ on the GPL, that is its sections on what is and what is not a derivative work misleads people by supplying engineering definitions to a legal question. (Ignoring that an estoppel argument naturally arises from FSF’s explanations of the meaning of the GPL).
Here is a note I wrote back then about Rosen’s treatment of the topic in his book:
Crap, three posts in a row, that makes me the loser.
Correction to my 3:10pm: “Alsup’s opinion stands solely for the proposition that the API itself, and those lines that define the API interfaces, are not protected by copyright.”
Thanks for the snippet. I read your well-written paper, and in that it seems that (like any careful attorney) you can see both sides of the problem. Obviously, a decision could go either way, based on the exact fact pattern.
Nonetheless, I think the GPL provides much stronger protections for people writing programs that communicate with GPLed programs (the license itself, not just estoppel from the FAQ) than either the Formgen or D&B licenses. Yes, the D&B decision came down hard on the developers, but the developers (and their customers) violated so many of the contract provisions before they even wrote the code, it would have put the judge in a bad mood. And finally, of course, they did, in fact, modify the D&B code in order to make it possible to use it from their program.
Obviously, someone could attempt a similar thing with GPLed code, which is why the FSF had all the hand-wringing over allowing plugins for GCC. I think their current solution is rather an elegant hack, in that it at least lets the code developers stop worrying about legal issues and concentrate on making a good architecture.
But if someone just wrote a program that could use unmodified (else they might have to distribute it) GPLed code, it’s hard to see, in most cases, that there would be a problem, because the user is doing the linking, and the license allows that. (It’s interesting to note that the Formgen decision anticipated the libedit clone of readline…)
Finally, of course, if I am giving away open source code that interacts with GPLed code (even under a non-GPL-compatible license), fair use and other issues come into play. In your paper, you mentioned a section Eben Moglen’s slashdot interview. That’s actually some verbiage that I personally found offensively misleading, as I wrote in 2003
My paper? I think you may be confusing me with someone else, Patrick. What did you find?
Oh,then it wasn’t your paper. I googled for some of the words you used and found somebody else’s paper. Now that I think about it, the note you refer to wasn’t in there. I think I may have known your identity in the past , but forgot it — getting old.
Anyway, the paper’s here:
Yep, not me. Hadn’t seen that paper either now that I think of it. Thanks for the link. As for my real identity, even the Commissioner does not know. Now why he’s never traced the phone wires to my lair is a mystery …