Bad day for Oracle
June 1, 2012 12:07 AM   Subscribe

Following a jury finding that Google had not infiringed upon Oracles patents, a development described as a near disaster for the database company, Judge William Aslup has ruled that the Java APIs cannot be copyrighted. That leaves Oracle with only the 9 lines of rangeCheck code and a handfull of decompiled test files to show for the massivecourt case. CEO Larry Ellison remains confident, claiming that the aquisition of Java creator Sun has still paid for itself.
posted by Artw (45 comments total) 11 users marked this as a favorite

 
In depth coverage from Groklaw.
posted by Artw at 12:12 AM on June 1, 2012 [1 favorite]


Yep, that's pretty much what technical folks with a clue, instead of an axe to grind, expected. The whole idea of being able to copyright an API is just silly. An API is just a description of what code will do, it has nothing to do with implementation details, or even the language, really.

If you could copyright an API, then once someone wrote a function to produce say, a cosine, then nobody else could ever write a cosine function again... or at least not within the lifetime of anyone then living.

You can pretty clearly copyright the code you use to implement the function, but the description of the function's inputs and outputs is just a fact, not an expression.
posted by Malor at 12:12 AM on June 1, 2012 [10 favorites]


Another way of putting it: by filing this case, Oracle was making the bald assertion that it owns any code you've ever written in Java, that nobody else is ever allowed to make a competing implementation that will run your program. By their interpretation of copyright law, by simply providing a compiler, they get to be your only vendor for your working lifetime. There can be no competition to run your program faster or better; Oracle owns your code permanently.

It's fortunate that this judge did some programming of his own, so he understood just how ridiculous a claim this is.
posted by Malor at 12:16 AM on June 1, 2012 [8 favorites]


Indeed...
Oracle v Google could clear way for copyright on languages, APIs and here's why that would be a VERY BAD THING
posted by Artw at 12:20 AM on June 1, 2012 [1 favorite]


[Judge] Alsup told Boies, "I have done, and still do, a significant amount of programming in other languages. I've written blocks of code like rangeCheck a hundred times before. I could do it, you could do it. The idea that someone would copy that when they could do it themselves just as fast, it was an accident. There's no way you could say that was speeding them along to the marketplace. You're one of the best lawyers in America --how could you even make that kind of argument?"
I think the lede has been buried here a bit: a great part of the story is that a judge decided to dive in to a technical field so completely that he'd be able to ask truly staggeringly pointed, educated questions and not just buy what both sides were trying to sell him.

We need more judges like that if we're going to continue to have insane software patent (and copyright) lawsuits and troll-fests and all the horrible economy-dragging crap that comes with it. If judges gain some insight into the subject matter and become familiar enough with it, common sense, even as it pertains to the law, has a chance of prevailing.
posted by disillusioned at 12:30 AM on June 1, 2012 [50 favorites]


Meanwhile: Intellectual Ventures' Nathan Myhrvold defends patent trolling, calls tech industry immature
posted by Artw at 12:33 AM on June 1, 2012 [1 favorite]


I don't feel like the Oracle lawyers in this case really expected they had a chance in hell of getting any sort of windfall from this. I see this as having Larry Ellison's personal, maniacal, not-entirely-sane touch to it. When a judge tells you, to your face, "It will not end well for you if you pursue this case," and you pursue it anyway, that's probably because there's someone much higher up than you that you're afraid of.
posted by WaylandSmith at 12:36 AM on June 1, 2012 [1 favorite]


Linus Torvalds:
Prediction: instead of Oracle coming out and admitting they were morons about their idiotic suit against Android, they'll come out posturing and talk about how they'll be vindicated, and pay lawyers to take it to the next level of idiocy.

Sometimes I really wish I wasn't always right. It's a curse, I tell you.

posted by Artw at 12:37 AM on June 1, 2012 [5 favorites]


I see this as having Larry Ellison's personal, maniacal, not-entirely-sane touch to it. When a judge tells you, to your face, "It will not end well for you if you pursue this case," and you pursue it anyway, that's probably because there's someone much higher up than you that you're afraid of.

This guy reckons it was all out of loyalty to Steve Jobs: What was Larry Ellison thinking in Java Android lawsuit?

...as good as any explanation at this point, TBH.
posted by Artw at 12:44 AM on June 1, 2012 [1 favorite]


Oracle advocate Florian Mueller weighs in: Judge says Google only used uncopyrightable elements of 37 Java APIs in Android
posted by Artw at 12:52 AM on June 1, 2012


Florian Mueller summarized: Ignore what I said before, the judge is a Communist in these matters:

"Just to be very clear: Judge Alsup despises communism (he blamed Google in one of his orders last year for a "soviet-style" approach to business negotiations). But his order on this particular issue of API copyrightability happens to come down on the anti-property side of the spectrum. He could have held the asserted API SSO copyrightable and pointed to fair use and antitrust as the right ways to address any legitimate concerns."

Just to be very clear: Florian Mueller has become a self-parody.
posted by jaduncan at 1:45 AM on June 1, 2012 [7 favorites]


Yep, that's pretty much what technical folks with a clue, instead of an axe to grind, expected. The whole idea of being able to copyright an API is just silly. An API is just a description of what code will do, it has nothing to do with implementation details, or even the language, really.

First, I agree with you 100%. Wonder if Oracle was taking a (or should have taken a), "look at this entire forest!" angle - the entire collection of the various API's for Java most be insanely immense. All designed and documented, I bet it weighs a bit if printed out. Joking about Java's weaknesses aside, I bet it took quite a bit of thinking to get it to work together the way everyone involved were happy with, without even talking about the implementation of the methods, etc itself.

Maybe I'm just used to making API's that shuffle complex data structures around, instead of a single argument (or two), so the APIs of my apps are kind of, I dunno, complex. Thems still is the breaks for Oracle.

It's still weird that the only shared code found was for something trivial. It smells a little fishy, as "easy" methods/subs usually have all these weird edge cases, and if someone wrote it already, it may be a good idea to reuse their code, anyways. I mean, do people rewrite C's stdio often?
posted by alex_skazat at 2:01 AM on June 1, 2012


This guy reckons it was all out of loyalty to Steve Jobs: What was Larry Ellison thinking in Java Android lawsuit?
Yeah, I was thinking the same thing. Jobs wanted to "Destroy" android, and took action by filing patent lawsuits. Larry Ellison was actually really good friends with jobs (Ellison was distraught over Job's illness). There's a good chance this was all about avenging Jobs.

Anyway, here's something odd: The JDK is actually available under the GPL, Did Oracle sue them over code that was actually covered under the GPL? Or was it a side package that wasn't available as a part of OpenJDK? If Google had released android under the GPL, rather then a less restricted license could they have avoided this? That would have been kind of hilarious. (I guess it doesn't matter too much in the end, since they lost)
Yep, that's pretty much what technical folks with a clue, instead of an axe to grind, expected. The whole idea of being able to copyright an API is just silly. An API is just a description of what code will do, it has nothing to do with implementation details, or even the language, really.
Well, it would be a disaster. You would no longer be able to create implementations of things that were compatible with other things. But it seems like you could argue that it might be copyrightable. After all, if you're talking about an API that has tens of thousands of functions in thousands of classes you could argue that there is a significant amount of "human readable information" in that.

But like I said, it would be a very bad thing if it were the case. It would be the end of making software that worked with other software without the permission of the person who made the original. So it's certainly a good thing that the case wasn't decided that way.
If you could copyright an API, then once someone wrote a function to produce say, a cosine, then nobody else could ever write a cosine function again... or at least not within the lifetime of anyone then living.
Not exactly. You could still do it you would just have to change the name of the functions or packages. So Trig.cosine(x) and Math.cos(y) wouldn't infringe on each other's copyrights. But copyrights don't apply to a single word or either.

If I write "cosine(x)" it doesn't infringes the copyrights of countless math textbooks that have also mentioned "cosine(x)". If you had hundreds of math functions it would be easier to make them not match up with other math libraries.

But anyway, like I said -- it would obviously be terrible if APIs, rather then their implementations.
It's still weird that the only shared code found was for something trivial.
It was probably just an error. Someone at Google had probably just thrown those files into their build in order to test some code that had been half written, and then somehow they ended up getting checked in.
posted by delmoi at 2:13 AM on June 1, 2012


jaduncan: Just to be very clear: Florian Mueller has become a self-parody.

Oracle is sending Mueller checks, right? I mean, that guy has got to be prostituting himself for money. He couldn't really believe that bullshit he's spewing. If you were to stamp SHILL on his forehead, it still wouldn't be enough to describe how mendacious and misleading he is.

alex_skazat: It's still weird that the only shared code found was for something trivial. It smells a little fishy, as "easy" methods/subs usually have all these weird edge cases,

Wait, so you're asserting that Google's very, very careful clean-room reimplementation of Java is fishy because ithey really didn't copy any code but that one function? That's a pretty strange argument. The fact that they did what they set out to do is an indicator of malfeasance?

In essence, you're pointing at a lack of evidence as somehow being evidence. That's.... well, let's call it a confused argument, and leave it at that.
posted by Malor at 2:14 AM on June 1, 2012 [1 favorite]


On failure to preview:

delmoi: If Google had released android under the GPL, rather then a less restricted license could they have avoided this?

Actually, I've seen arguments that this is exactly correct, that had Google just used Sun's free version of Java directly, there would have been no possible infringement in any area. Oddly, their decision to come up with their own codebase made them more vulnerable to the lawsuit than just using Sun's, for free.

At the same time, however, that would also make them hostage to Oracle imposing new and draconian license terms, so overall, it was probably a smart decision to roll their own. Oracle is a terrible company, and you don't want technologies you depend on under their control.
posted by Malor at 2:20 AM on June 1, 2012


I don't feel like the Oracle lawyers in this case really expected they had a chance in hell of getting any sort of windfall from this. I see this as having Larry Ellison's personal, maniacal, not-entirely-sane touch to it.

When I think about Oracle the main thing that comes to mind is their hilarious, cartoonish level of evil. In the years between Episode III and Episode IV, Darth Vader was an account executive at Oracle. His boss, Satan, was a middle manager. That's how evil Oracle is. People who yammer on and on about Microsoft don't know anything about anything. If you've licensed an Oracle Enterprise cluster before you know that there are much greater evils in the world. If you haven't, here are two words for you: per core.

I've always viewed this lawsuit as simply a thing that some executives at Oracle were compelled to pursue because of their Oracle-ness. Because it's wrong, and they could. Just, you know, Oracle. Doing Oracle stuff. It doesn't have to make sense. Like Fox News. Or Cirque Du Soleil. Just a bunch of bad people doing bad things.
posted by tracert at 2:30 AM on June 1, 2012 [20 favorites]


I don't feel like the Oracle lawyers in this case really expected they had a chance in hell of getting any sort of windfall from this.

In a moment of naïveté, I admit my heart sank a little when I saw David Boies' name attached to this case. This is the dude that overturned Prop 8 in California, whom Bill Gates said was "out to destroy Microsoft", and defended Gore (albeit vainly) in Bush v. Gore.

Then again, Andy Fastow.
posted by mykescipark at 4:02 AM on June 1, 2012


Wait, Cirque Du Soleil is bad people doing bad things?
posted by snwod at 4:16 AM on June 1, 2012 [2 favorites]


Boies really isn't a very good lawyer for technical stuff. His strategies, to be blunt, suck. I don't know how he has the rep he does, because the cases of his I've seen, SCO and Oracle, were both very poorly handled. And the SCO case was particularly full of what I'd characterize as dirty lawyering.

Not that it helped, but they were really scumbags.
posted by Malor at 4:25 AM on June 1, 2012 [2 favorites]


Anyway, here's something odd: The JDK is actually available under the GPL, Did Oracle sue them over code that was actually covered under the GPL? Or was it a side package that wasn't available as a part of OpenJDK? If Google had released android under the GPL, rather then a less restricted license could they have avoided this? That would have been kind of hilarious. (I guess it doesn't matter too much in the end, since they lost)

The JDK wasn't available under the GPL at the time Google was developing much of the Android class libraries. In fact, Android's implementations are largely based on chunks of Apache Harmony, to which IBM contributed a fair amount of code, a project which set out to build its own Apache-licensed Java implementation. Harmony eventually folded after OpenJDK was released by Sun, though pieces live on within Android.

Google couldn't use the GPL for Android (except the Linux kernel) since the handset manufacturers were and are incredibly terrified of anything resembling a strong copyleft license. Imagine Motorola setting out to develop Motoblur (yuck) and having teams of lawyers arguing with engineers over what parts of the skin were derivative works and had to be open sourced. Since a crappy and slow UI skin would obviously represent a major competitive advantage for Motorola, the idea that they might have to distribute their source would be horrific.

Finally, it's not entirely clear to me that Google would have been in the clear if they had used Java under the GPL. It's important to note here that Google has never claimed that Android implements Java with a capital J and a trademarked coffee cup. In fact, they've explicitly said quite the opposite, and it has never been Android's goal to be offer a complete implementation of Java. Several Android design features, the use of the Dalvik VM chief among them, preclude this kind of compatibility. Especially given the rather weak implicit patent grant in GPLv2, my uneducated gut feeling is that nothing about a GPL'd Java would have stopped Oracle from suing Google for patent infringement related to its implementation. Of course, Oracle lost on their patent claims anyway, but that's another story.
posted by zachlipton at 4:39 AM on June 1, 2012 [2 favorites]


Wait, Cirque Du Soleil is bad people doing bad things?

Yes. Clowns.
posted by tracert at 5:03 AM on June 1, 2012 [3 favorites]


Florian Mueller has become a self-parody.

Florian Mueller was a shill and a joke already ten years ago when he led the charge against the European Computer-Implemented Inventions Directive. That it has taken his then-followers all this time to become aware of that is a source of endless if somewhat bittersweet mirth to me. He really pwned them.

Unfortunately, much of the FUD that he spread back then about patents has remained vividly present in the minds of the FOSS community, thoroughly poisoning the debate.
posted by Skeptic at 5:39 AM on June 1, 2012 [1 favorite]


Malor: Oracle is sending Mueller checks, right?

yep:

"I would like to inform you that Oracle has very recently become a consulting client of mine."
http://www.fosspatents.com/2012/04/oracle-v-google-trial-evidence-of.html
posted by nielm at 5:43 AM on June 1, 2012 [1 favorite]


Florian Mueller has become a self-parody.

The trouble with this kind of guy is that it's very hard for a non-expert to tell a biased expert from a neutral one.
posted by smackfu at 6:01 AM on June 1, 2012


I find it interesting that the judge in this case actually has some technical knowledge in the subject domain. You have to wonder how many patent/copyright cases that were decided by judges and juries who were not knowledgeable would have ended up had they some pertinent knowledge.

We talk about "a jury of your peers" but who would be the peers of Google and Oracle?
posted by tommasz at 7:09 AM on June 1, 2012


Yeah, having followed this a bit on Groklaw, this case has been full of Judge Aslup being a total badass. Here's another choice quote from his decision:
For example, Java-based code using the replicated parts of the
37 API packages will run on Android but will not if a 38th package is needed. Such imperfect
interoperability leads to a “fragmentation” — a Balkanization — of platforms, a circumstance
which Sun and Oracle have tried to curb via their licensing programs. In this litigation, Oracle
has made much of this problem, at times almost leaving the impression that if only Google had
replicated all 166 Java API packages, Oracle would not have sued.
posted by gilrain at 7:25 AM on June 1, 2012 [1 favorite]


tommasz: We talk about "a jury of your peers" but who would be the peers of Google and Oracle?

Actually, over the course of this case my impression of this jury, in particular, and juries in general has improved a lot. In this case, they constantly asked intelligent clarifying questions about the technical details being discussed. Yes, they were working on an abstract level, conceptual rather than technical, but they made themselves effectively understand the necessary issues.

The judge specially praised the jury at the end of the case, though, so perhaps they made an exceptional effort.
posted by gilrain at 7:35 AM on June 1, 2012 [1 favorite]


The trouble with this kind of guy is that it's very hard for a non-expert to tell a biased expert from a neutral one.

If somebody claims to be a "patent expert", yet has no qualification whasoever in law in general, never mind patent law in particular, it is quite easy to know that he actually isn't an expert.

If moreover, when reporting on litigation, he appears to have first-hand knowledge of the submissions and arguments of one of the parties, then it's quite safe to presume that he is quite close to that party.

Anyway, Mueller has always been a PR gun-for-hire, not an IP expert. Furthermore, he's never really tried to hide it. Indeed, he's generally been quite straightforward as to who buttered his bread. This is why I find it so amusing that just those who once took his words as holy gospel now would like to see him drawn and quartered as a traitor to the cause (while simultaneously falling for equally slippery fellows who are bound to disappoint them again sooner rather than later).
posted by Skeptic at 7:44 AM on June 1, 2012 [2 favorites]


I think the lede has been buried here a bit: a great part of the story is that a judge decided to dive in to a technical field so completely that he'd be able to ask truly staggeringly pointed, educated questions and not just buy what both sides were trying to sell him.

My impression of the writeups about that little (awesome) altercation was that the judge was indicating that he's already been doing a lot of programming? Otherwise, why mention the "other languages"—why not just use Java, since it's the language at issue? And he points out specifically that he has done a fair amount.

In which case, still an amazing story, but more luck-of-the-draw.
posted by kenko at 7:47 AM on June 1, 2012 [1 favorite]


Oh, I always thought Florian was full of it. But he's recently become almost impossible even to parody.
posted by jaduncan at 7:51 AM on June 1, 2012


You have to wonder how many patent/copyright cases that were decided by judges and juries who were not knowledgeable would have ended up had they some pertinent knowledge.

Getting technically knowledgeable judges on the bench is difficult. First, that really narrows the pool of applicants, since only about 7% of lawyers have a science or engineering degree. Second, federal judges (i.e. the ones hearing all patent and copyright cases) are appointed for life and must be confirmed by the Senate, so it's a slow process to appoint and to replace them.

But if you could put a lot of technically-qualified judges on the bench, you'd have to have a dozen or more in every district in order to cover all the various fields of technology.

In any case, it's probably unnecessary. Judges are free to appoint special masters to help them with some tasks where an understanding of the technology is important. Judges in patent cases use special masters more often than in other types of litigation (source [pdf]). For example, a judge might appoint a special master to help with claim construction (i.e. deciding what the patent claims actually cover), and the special master will then wade through the competing constructions offered by the parties and supported by their own experts.

There has been talk of creating specialized patent courts for decades, though specialization need not be the result of a top-down organizational change. The courts are fairly free to decide how to allocate judges to cases, and I believe some have experimented with having particular judges take patent cases more often than others (as opposed to a pure round robin for all cases or some other effectively random allocation). And of course judges in some disproportionately patent-heavy dockets (e.g. Texas Eastern, Virginia Eastern, Wisconsin Western) end up with a fair amount of experience.

On the Federal Circuit I think Judge Newman is the only Circuit Judge with a technical background, namely chemistry. Given that the Federal Circuit is a pretty specialized court, some have argued that more Federal Circuit judges should have technical backgrounds.
posted by jedicus at 8:12 AM on June 1, 2012 [1 favorite]


In other news, Oracle has a merger in the works to become Sco.racle Unix.
posted by roboton666 at 8:24 AM on June 1, 2012 [2 favorites]


Artw: "Meanwhile: Intellectual Ventures' Nathan Myhrvold defends patent trolling, calls tech industry immature"

Meanwhile: Metafilter's Own Symbioid attacks patent trolling, calls Nathan Myhrvold immature.
posted by symbioid at 8:33 AM on June 1, 2012 [4 favorites]


Larry, Larry, Larry...

No more walled gardens...

THIS is what happens when you kill off Sun's SPARC workstation line...
posted by PROD_TPSL at 8:36 AM on June 1, 2012


If you could copyright an API, then once someone wrote a function to produce say, a cosine, then nobody else could ever write a cosine function again... or at least not within the lifetime of anyone then living.

As I understand it, "API" in this legal argument refers to a collection of function signatures, not an individual function signature. An analogy would be that you can't copyright a fact but you can copyright a collection of facts.

(I'm not claiming that I support this argument, nor am I a lawyer.)
posted by Slothrup at 9:58 AM on June 1, 2012


Anyway, here's something odd: The JDK is actually available under the GPL, Did Oracle sue them over code that was actually covered under the GPL?

*If* Google had copied the code and then distributed it under a different license than the GPL, it would be a violation of the terms under which they acquired the code.
posted by Slothrup at 10:02 AM on June 1, 2012


I don't understand why this is being portrayed as a near disaster for Oracle? They still have tons of other product lines pulling in the enterprise dollars; this is just one less revenue stream they took a shot at and it didn't pan out. It's more like not winning the lottery. Oracle being Oracle, they didn't buy the lottery ticket, they stole it off a short woman as she left the 7-11.
posted by benito.strauss at 10:17 AM on June 1, 2012 [2 favorites]


Well, they were hoping for a ridiculously huge pay-off, and they'd sunk 7 billion dollars into buying Sun plus millions in court expenses into getting it.

I dunno, maybe just trashing Open Office makes the expense worth it somehow in their minds.
posted by Artw at 1:31 PM on June 1, 2012


I dunno, maybe just trashing Open Office makes the expense worth it somehow in their minds.

No, not that either: http://www.libreoffice.org/download/
posted by jaduncan at 1:39 PM on June 1, 2012 [1 favorite]


They bought Sun just to get Java? I'd believe it, but I'd never heard it said before.
posted by benito.strauss at 2:51 PM on June 1, 2012


No doubt a bunch of other reasons too, like MySQL - certainly since aquiring Sun they've been trying to transform every aspect of it they can into a quick win meal ticker. Screwing Google over using Java would have probably have been the jewel in the crown though, even in 2009.
posted by Artw at 8:58 PM on June 1, 2012


The constellations have shifted again. Oracle has just sued Lodsys, seeking to invalidate four of its patents. The complaint actually claims noninfringement and invalidity.

I know. Oracle is now the good guy. Major good guy.

posted by Artw at 2:55 PM on June 5, 2012


BREAKING
Apple v. Motorola: Judge Posner dismisses entire patent case, no injunctions to be issued. 

posted by Artw at 5:10 AM on June 23, 2012


(Ignore the BREAKING, cut and paste error)
posted by Artw at 5:11 AM on June 23, 2012


Groklaw on Oracle's motion for appeal.
posted by Artw at 10:56 AM on June 25, 2012


« Older ITU Power-grab, or is something else brewing behin...  |  Last week, I wrote about how u... Newer »


This thread has been archived and is closed to new comments