Follow Slashdot stories on Twitter

 


Forgot your password?
Close
typodupeerror
×
Android Java Oracle The Courts Your Rights Online

Oracle Vs. Google and the Right To Use APIs 155

Posted by samzenpus from the my-precious-code dept.
jfruh writes "Even as an EU court rules that APIs can't be copyrighted, tech observers are waiting for the Oracle v. Google trial jury to rule on the same question under U.S. law. Blogger Brian Proffitt spoke with Groklaw's Pamela Jones on the issue, and her take is that a victory for Oracle would be bad news for developers. Essentially, Oracle is claiming that, while an individual API might not be copyrightable, the collection of APIs needed to use a language is. Such a decision would, among other things, make Java's open source nature essentially meaningless, and would have lots of implications for any programming language you can name."
This discussion has been archived. No new comments can be posted.

Oracle Vs. Google and the Right To Use APIs More

Oracle Vs. Google and the Right To Use APIs

Comments Filter:

  • Intel licensed the x86 instruction set to AMD.

    If an instruction set is licensable, then an API can be too...although it's unclear from what I've read if the licensing covered the instruction set itself or the right to manufacture something compatible with that instruction set. Either way, that should cover the API.

    • Re: (Score:2, Interesting)

      by Anonymous Coward

      Yet you do not see Intel going after the likes of dosbox, virtualbox, virtualpc, vmware, bochs....

      Me thinks the instruction set license was an easy way out for most parties instead of a long drawn out court battle...

      • Comment removed based on user account deletion

      • DOSBox and Bochs are the only ones there that actually implement any non-trivial part of the x86 ISA. Notably, they are a) open-source, either GPL or LGPL, and b) the developers had no special access to any of the "APIs". All they stuff they used was open literature - and you can't sue someone for violating a license you didn't have on the "intellectual property".

        VirtualBox, VirtualPC, and VMWare are all virtualizers - they *require* an Intel-compatible CPU to run. They do not implement any of the x86 instr

    • A lot of this stems from the agreement Intel made with AMD to have a second source for chips for IBM PC computers - they in turn licensed AMD64 to Intel. No-one ever went to court, but doubt you can copyright op-codes in the same manner.

      Maybe this is what Oracle is hoping for? Some sort of cross licensing agreement with Google? A piece of the Android pie.

    • Re: (Score:2, Informative)

      by Anonymous Coward

      x86 instruction set is licensed to AMD as part of a patent-licensing agreement -- because parts of the instruction set itself are patented. E.g.,:

      Intel believes that Global Foundries is not a subsidiary under terms of the agreement and is therefore not licensed under the 2001 patent cross-license agreement. Intel also said the structure of the deal between AMD and ATIC breaches a confidential portion of that agreement. Intel has asked AMD to make the relevant portion of the agreement public, but so far AMD has declined to do so.

      Copyright != patent protection. Oracle has never claimed that Google violated its patents (probably because it doesn't hold patents protecting the Java API). It's worth noting that patenting the API would clearly make it patent-encumbered and thus, potentially susceptible to exactly the kind of Shenanigans that Oracle is trying to pull. Of course, if it was paten

      • Patent suit still pending (Score:5, Informative)

        by tepples ( 727027 ) <.tepples. .at. .gmail.com.> on Thursday May 03, 2012 @06:31PM (#39884549) Homepage Journal

        Oracle has never claimed that Google violated its patents (probably because it doesn't hold patents protecting the Java API).

        The case was in fact about patents before it became clear that Google was very likely to prevail on Oracle's patent infringement allegations. At that point, Oracle decided to add a copyright infringement claim. The suit over U.S. Patents 5966702, 6910205, and RE38104 is still pending and will be resolved after the copyright suit concludes.

      • Re:Instruction sets can be licensable (Score:5, Informative)

        by exomondo ( 1725132 ) on Thursday May 03, 2012 @06:57PM (#39884777)

        Oracle has never claimed that Google violated its patents (probably because it doesn't hold patents protecting the Java API).

        Yes they did, in fact that was the initial basis for this court action [slashdot.org]. The copyright portion is being resolved first then it moves on to the patent portion of the lawsuit.

        It's worth noting that patenting the API would clearly make it patent-encumbered and thus, potentially susceptible to exactly the kind of Shenanigans that Oracle is trying to pull.

        Java is quite heavily patent-encumbered.

        Of course, if it was patent-encumbered, then the industry would [probably] know about it, and would steer clear, as appropriate.

        Well no, that would be silly, because it is patent-encumbered but the GPL status of Java means that GPL implementations - like OpenJDK - get an implicit patent license as well as all downstream projects.

        Basically, Oracle is trying to have its cake ("open") and eat it too ("license").

        Just because something is open doesn't mean it can't be patent-encumbered, if you understand the patent-specific elements of the GPL this should be quite clear.

    • Re:Instruction sets can be licensable (Score:4, Informative)

      by Anonymous Coward on Thursday May 03, 2012 @06:25PM (#39884489)

      As someone who worked in x86 Architecture, I'm pretty sure (though I'm not a lawyer, so get their opinion definitively first) that the issue here is not the instruction set Per Se, but various physical implementation details that would occur building real circuits to produce a processor chip.

      So, for example, that's why various software-based x86 emulator/simulator writers have not been sued.

      • Shame you posted that as AC: it's exactly correct. The issue with instruction sets is that there are often very few sensible ways of implementing a specific instruction in hardware. The canonical example of this is the MIPS unaligned load / store patent. If you can figure out a way of implementing an unaligned load that doesn't involve the mechanisms covered in the MIPS patent, then you are free to do so. If you can't, then you need to license the patent from MIPS to be able to implement the instruction

    • Re: (Score:1)

      by Anonymous Coward

      Intel holds patents on various necessary mechanisms for implementing the x86 instruction set. Their licensing regime is mainly based on strong patents, not on mere copyrights or the spelling of assembler opcodes.

      In the 1970's, there were companies that cloned Data General's NOVA instruction set and ran Data General's RTOS operating system. RTOS was offered below cost, with DG expecting to get compensated by hardware sales. Data General defeated those clone companies by lawsuit. I think in that case, it

    • Well, anything is licensable. The question is, can you use it without a license?

      Consider the "cloud music" stuff. Apple signed licenses with all the music companies so that their customers could link their music libraries with iTunes Music Store. Turns out, they didn't need to do that.

      Which is cheaper? Fighting it in court or just paying a license?

  • hope we luck out (Score:5, Insightful)

    by Eponymous Hero ( 2090636 ) on Thursday May 03, 2012 @06:19PM (#39884421)
    the rest of us who know what we're talking about have to sit around and wait for outsider judges and juries to decide the context of things far outside their grasp. it takes years for an engineer to become competent in these technologies, and now we have bus drivers and secretaries deciding what applies to us and our trade in the span of mere weeks. can we get specialized jury selection for cases involving specialized knowledge?

    • Re:hope we luck out (Score:5, Insightful)

      by sribe ( 304414 ) on Thursday May 03, 2012 @06:59PM (#39884793)

      ...and now we have bus drivers and secretaries deciding what applies to us and our trade in the span of mere weeks...

      If you'd been following the case you'd know that this judge has been working very hard, for months, to understand the issues and make sure the jury is presented with well-formed questions and good background. How good a job he has been able to do, I don't know because that would require wading through hundreds, perhaps thousands, of pages :-)

      • "If you'd been following the case you'd know that this judge has been working very hard, for months, to understand the issues "

        I think the parent poster did understand that. He was merely pointing out that a judge even after months of study is unlikely to fully understand the ramifications as well as someone who has been studying and developing API's and languages for decades.

        Never mind the jury; no matter what the judge tells them, their judgement will essentially be random.

        • Re:hope we luck out (Score:5, Funny)

          by gman003 ( 1693318 ) on Thursday May 03, 2012 @08:51PM (#39885769)

          What we need is a professional, full-time expert in car metaphors, who can go before the court and say "This case here is like a carburetor..."

          • "This case here is like a carburetor..."

            Are you trying to imply that Java is old and busted?

            • I.... I don't actually know what a carburetor is.

              I had to try three times just to get spell-check to correct it properly.

        • Re: (Score:2)

          by rtb61 ( 674572 )

          What is interesting here is if Oracle wins they will lose. The reason they will lose is because a win will affect every single other developer using Java as they also come under risk, end result, 'Class Action Lawsuit' pretty much guaranteed. A professional legal team experienced in computer litigation will take up the challenge to sue Oracle for misleading developers who have made substantive investments in Java with regard to their rights to use the language and risk that implies to the ownership of the

        • Re: (Score:2)

          by curunir ( 98273 ) *

          He was merely pointing out that a judge even after months of study is unlikely to fully understand the ramifications as well as someone who has been studying and developing API's and languages for decades.

          Similarly, the judge is much less likely to have any personal stake in a legal decision on the issue than someone who has been studying and developing APIs and languages for decades.

          As a software developer, it's occurred to me how cool it would be to have a chance to serve on the jury in this case. I think it would be an amazing opportunity to learn about the legal issues involved in my chosen profession and a chance to get a deep-dive into one of the technologies that I have to use on a daily basis. I'm no

      • Re:hope we luck out (Score:4, Interesting)

        by Simon Brooke ( 45012 ) <stillyet@googlemail.com> on Friday May 04, 2012 @04:01AM (#39887789) Homepage Journal

        ...and now we have bus drivers and secretaries deciding what applies to us and our trade in the span of mere weeks...

        If you'd been following the case you'd know that this judge has been working very hard, for months, to understand the issues and make sure the jury is presented with well-formed questions and good background. How good a job he has been able to do, I don't know because that would require wading through hundreds, perhaps thousands, of pages :-)

        The key point about juries is that they are 'of your peers', people equal to and in a similar position to the people being tried. It does not seem to me that in this case bus drivers and secretaries are in any useful sense 'peers' of software engineers - they lack the knowledge, concepts and language required to form a considered opinion of this case. It's like having a jury of Indonesian rice farmers adjudicating on the Enron [wikipedia.org] case: it doesn't work because they cannot reasonably be expected to understand the arguments. It really doesn't matter how well intentioned or well read the judge may be: without a common language, communication isn't possible (which is, actually, precisely what the case is about).

    • Re:hope we luck out (Score:5, Insightful)

      by Anonymous Coward on Thursday May 03, 2012 @07:24PM (#39885115)

      The expert knowledge in the case is supposed to come from the expert witnesses.

      You don't really want the jurors to be too deep in the field; in particular, you don't want the jurors to be privately disagreeing with the explanations of the expert witness.
      Why? Because the jurors are not examined. An "expert juror" could decide the case based on a prejudice that is never heard in court, and neither of the sides get the opportunity to challenge. A jury with few preconceptions is good for transparency.

      Besides, how likely is it that such a jury would be unbiased? Would they rule against Google despite the implications for IT if Oracle wins?

      • My kingdom for mod points...

        This is exactly right. Given the somewhat representative sample of experts on this site, consider how much bias there is about trivial things like Google playing with augmented reality or Oracle running Java on a Raspberry Pi... There's a group who thinks that Google's always out to steal information, and absolutely must have a nefarious agenda behind its glasses. There's an independent group of people (myself included) who believe that Oracle is trying to piggyback on any good P

        • Re:hope we luck out (Score:4, Interesting)

          by zbobet2012 ( 1025836 ) on Friday May 04, 2012 @01:29AM (#39887203)

          Why? Because the jurors are not examined. An "expert juror" could decide the case based on a prejudice that is never heard in court, and neither of the sides get the opportunity to challenge. A jury with few preconceptions is good for transparency.

          but just what an API is. That can be taught adequately in about two hours.

          This is wrong, and provably so. 30% to 60% of CS students never make it pass the first class [mdx.ac.uk]. Literally, 30-60% of the population can't program. To quote the linked paper "It is as if there are two populations: those who can [program], and those who cannot [program], each with its own independent bell curve."

          Statistically speaking your try to teach someone in two weeks which some where around 60% of people literally can't be taught. The authors posit as to why this is, but the important fact is your asking a population to judge something 30 to 60 percent of them will never understand.

          • Even if that paper were accurate, you're assuming that the knowledge required to pass an introductory programming course is exactly the same as (or even significantly overlapping with) the knowledge required to be an informed juror. That's simply not the case. Even the most basic concepts laid out in that paper (assignment and sequence) are more in-depth than what's really necessary for this case. For what I've seen of this case so far, a juror only needs to know a few details. Programs come from source, wh

          • (A separate post, because it's rather tangentially off-topic)

            Thank you for showing me that paper. I like a good laugh in the morning. More seriously, what the hell is this elitist ivory-tower bullshit, and who approved this?

            The premise is that a given percentage fail the programming class, and therefore must be unable to learn to program. In fact, there's a significant percentage (I've heard as high as 25% back when I was teaching, but don't have sources at the moment) that fail every introductory course, i

      • That's silly. The jurors are _supposed_ to decide from their own conscience. If one of them is an expert already and has a bias, there's nothing wrong with that and in fact it's great, because it means that his point of view is going to be represented among the twelve.

        The jurors are supposed to be representative (and chosen randomly) from the population at large, so if in the population 1% of people are experts with a certain bias, then they should be represented with approximately that frequency in jurie

  • If Oracle wins... (Score:3, Informative)

    by Anonymous Coward on Thursday May 03, 2012 @06:25PM (#39884483)

    ... I move to the EU.

    At least there is still one region in the world where the hairless monkeys haven't gone completely insane.

    • While the leader monkeys in EU aren't completely insane on their own, the same old "monkey sees, monkey does" still applies to them, though.

  • Judge Alsup is on it... (Score:5, Informative)

    by sconeu ( 64226 ) on Thursday May 03, 2012 @06:25PM (#39884485) Homepage Journal

    He has asked the parties to brief him in light of the EU decision [groklaw.net].

  • Matters of fact vs. matters of law (Score:3, Informative)

    by Citizen of Earth ( 569446 ) on Thursday May 03, 2012 @06:27PM (#39884507)

    waiting for the Oracle v. Google trial jury to rule on the same question under U.S. law.

    This idea seems rather broken. Juries are supposed to decide matters of fact and courts are supposed to decide matters of law. Whether an API is copyrightable is purely a matter of law, not fact. What the hell's going on here? This decision is the judge's responsibility.

    • The question submitted to the jury was one of fact: "Let's pretend that the API is copyrightable. Was it copied?" After the answer comes back, the judge will decide the question of law as to whether it was copyrightable in the first place.

      • Re: (Score:2)

        by mark-t ( 151149 )

        The problem with that line of reasoning is that it takes one down an avenue that is wholly irrelevant (and a waste of time) if the hypothetical scenario is false. The biggest danger with this line of thinking is that if you start by assuming that it was copyrightable, then deciding that they did copy it, you've set up a framework for possibly biasing your own decision in favor of concluding that it *WAS* copyrightable, instead of making an unbiased decision.

        For what it's worth, I'd be inclined to say t

        • Re: (Score:2)

          by Xtifr ( 1323 )

          Actually, it potentially avoids a waste of time. Whoever wins, the decision is certain to be appealed. With a jury verdict available, the appeals court can focus on matters of law without forcing a new trial of fact.

          For what it's worth, I'd be inclined to say that yes... Google copied the Java API.

          That question is not in dispute. Google, like Apache and the FSF and countless others, openly and publically copied the Java API, and has never pretended otherwise.

          • Re: (Score:2)

            by mark-t ( 151149 )
            If that question is not in dispute, then why is a judge asking a jury to decide it?

            • Pleading the Seventh (Score:4, Interesting)

              by tepples ( 727027 ) <.tepples. .at. .gmail.com.> on Thursday May 03, 2012 @07:42PM (#39885279) Homepage Journal
              Judge Alsup is pleading the Seventh [wikipedia.org]: "no fact tried by a jury, shall be otherwise re-examined in any Court of the United States, than according to the rules of the common law." This way, the jury's verdict on the facts is on the books no matter what questions of law the Court of Appeals remands back to Alsup, and there's no need for an expensive retrial.

            • Re: (Score:2)

              by Xtifr ( 1323 )

              It's not. The jury is going to decide whether the copying of the API alone could be defended as fair use or de minimus. The judge is going to decide whether an API can be copyrighted in the first place. If it's fair use or de minimus, Google wins. If the API isn't copyrightable, Google wins. If judge decides the API is copyrightable (which goes against strong precedent), we all lose.

  • It's not that in the EU that APIs can't be copyrighted, it's that APIs can be clean-room reimplemented. Whether or not they can be copyrighted is another story all together.

    Google didn't do a clean room implementation of Java with Dalvik.

    • Google didn't do a clean room implementation of Java with Dalvik

      The VM itself? No. But they did base their class libraries on Apache Harmony, a clean room implementation of the Java APIs. [for which Sun refused to license the Java Compatibility Kit anyway - the shenanigans started long before Larry got involved]

  • Turn-about (Score:4, Interesting)

    by evil_aaronm ( 671521 ) on Thursday May 03, 2012 @06:40PM (#39884605)
    Wouldn't Oracle be guilty of this, as well, to some extent? Oracle couldn't possibly be using a 100% "clean" environment for their product or development systems. If they use C and standard libraries, they're using, in effect, copyright-able APIs.

    • Re: (Score:3)

      by nurb432 ( 527695 )

      While that is true, is the API for C in the public domain or controlled by a company? If its a company, does this company want to milk people or just say 'go ahead and use it'.

      Being copyrighted by its self does not mean good or bad things, it just 'is'. Its how that copyright is used is what matters.

      Just think about how the GPL, BSD, etc copyrights work.

    • Oracle released Oracle V2 back in 1979. SQL was an IBM design/language.

      IBM themselves released SQL a few weeks later!

      Of interest is that MIT (RDBMS) and U/Cal Berkley (Ingres) had those out (but Ingres used QUELL, not SQL, but Ingres became Postgres, so, it could be argued that Postgres is the real SQL leader).

      Naturally, because SQL came from academia (although Codd worked for IBM, thereby allowing IBM credit).

      SQL wasn't standardized until 1986, so if "APIs and Languages" are ruled Copyrightable, Oracle spe

      • Re: (Score:2)

        by Lisias ( 447563 )

        Sue all them. The Oracle company was built on exactly what Oracle wishes to kill off.

        here, I fixed that for you...

    • Wouldn't Oracle be guilty of this, as well, to some extent? Oracle couldn't possibly be using a 100% "clean" environment for their product or development systems. If they use C and standard libraries, they're using, in effect, copyright-able APIs.

      +1, wish I had mod points.

      And C was heavily based on B, which was in turn heavily based on BCPL (and BCPL's CINT-code interpreter was a clear fore-runner of the Java Virtual Machine), which in turn was heavily based on the Combined Programming Language [wikipedia.org]. So if Oracle were to win this case, they owe shed-loads of money to the Universities of Cambridge and London, whose code they have so maliciously stolen. Meantime, SQL, the language which is at the core of Oracle's core database product, is clearly the prope

  • Far beyond programing languages (Score:4, Insightful)

    by nurb432 ( 527695 ) on Thursday May 03, 2012 @06:40PM (#39884609) Homepage Journal

    If API's can be restricted like this, it would go far beyond simple languages. Think about the concept of what an API actually is..

    Even something as apparently benign as replacement parts for your washer could be restricted and eliminate cheaper 3rd party parts. Even if its not electronic in nature, just claim that nothing can be compatible enough to hook up to your 'widget connector version b'.

  • Will be huge for ColdFusion (Score:5, Interesting)

    by devleopard ( 317515 ) on Thursday May 03, 2012 @06:47PM (#39884663) Homepage

    ColdFusion [adobe.com] was about the only language (in my experience - sure there's others) that you needed to pay for the runtime for your code (in a production environment; development version was free; The "Express" version went away around 2001 or so). Then along come Railo [getrailo.org] and Open BlueDragon [openbd.org], and there were open source alternatives. The "language" itself is pretty basic (most developers get by using just 5 tags), but the power of it comes when you use the various feature tags that are more akin to APIs (cfchart, cfpdf, cfsearch, etc). Railo and OpenBD of course implement all these tags. Whereas Oracle doesn't "sell" Java, Adobe sells ColdFusion - if Oracle wins, Adobe has 100% motivation to eliminate their competition. (Should also point out that OpenBD's lineage comes from New Atlanta [newatlanta.com], which sells commercial version of Blue Dragon - MySpace was built on this.)

  • It's not the what that should be eligible for copy-right protection, it's the how. APIs are the what. The implementation is the how.

  • So writing a piece of code that basically generates "Hello World" requires licensing?

  • Sadly, I absolutely do not trust the US to make the correct decision in this case. IP law in America needs to be nuked from orbit and rebuilt from scratch, and the prevailing attitude there right now doesn't seem to be amenable to something like that.

  • It seems that if Oracle successfully argues this point, everybody that provides Oracle with a library is going to demand a license fee. Imagine if Brian Kernighan decided that the C standard library ought to be generating more (well, some) income...

    • I wouldn't worry about Kernighan asking for anything.... I'd be worried that AT&T will got hold of this idea (as I believe they hold the copyrights to the c programming language)

  • I think that something is only an API if we call it an API. However, I think once an entity labels something as an API, that should be the same as saying, "this subset of our source code is freely implementable and modifiable by anybody." It should be the same as saying, "we as a company are opting out of claiming this part of our source code is our IP."

    How about this for an API:

    size_t count_chickens(const egg_vector_t& eggs);

    Ok, got a copyright on that. Don't even *think* about implementing that withou

  • Oddly (Score:3)

    by Greyfox ( 87712 ) on Thursday May 03, 2012 @11:48PM (#39886737) Homepage Journal
    When I search Google for "Oracle", I'm quite surprised to find the first hit is not for the goatse.cx guy. Perhaps they're waiting to do that until after the case is over?
  • Think about that for a moment. The x86 instruction set used in your computer today is an api. If copyright of an API is upheld, AMD disappears. Every ISA ever is now copyrighted. Competition is dead in a single fell swoop.
  • "bad news for developers"

    In the U.S.A., that is.

    On a sidenote, groklaw says the judge has asked the parties to reflect upon the EU court's decision on APIs being non-copyrightable. Should be a nice exercise :)

  • If this whole mess turns south, and Google can't use the current Android anymore, couldn't it just switch the codebase to be a derivative of OpenJDK?
    Sun/Oracle relaesed OpenJDK under the GPL. That means as long as you use OpenJDK, you can use their copyrighted works. All they would have to do is port the bits from current Android (like the bytecode interpreter and extra classes) to OpenJDK. Then release that as Android 5 or something.

Slashdot Top Deals

You are always doing something marginal when the boss drops by your desk.

Close