Google’s copyright win against Oracle is in danger on appeal
arstechnica.com
arstechnica.com
I think designing APIs is an art, and I think it is creative. But ultimately I think APIs are much more like protocols, file formats, or languages than they are like fixed works. Copyright API documentation all you like, but the day that people can control who is allowed to speak a language that they invented is the day that you cede control of everything to incumbents.
What if SCO could really have sued IBM on the basis that they owned the actual APIs to UNIX? What if you couldn't actually say "#define ENOENT 1" in your own project without violating SCO's copyrights?
There is also the counteracting market force that single source illiberally licensed API's are unappealing to users so in many cases granting rights to other implementations can be commercially desirable for the creator of the API. Mono's existance encourages .Net and C# adoption.
Also there is the distinction between copying an API to duplicate it and USING the API to develop a program against (this would the speaking a language aspect). I would argue that using the API would never be infringement but would in all circumstances I could imagine would be fair use.
On one hand, haven't we all worked with (or built) APIs which seem remarkable in their structure and simplicity? The best of the best seem like unambiguously creative works.
On the other hand, I don't want to work in a world where simply conforming to an interface specification could be illegal. What a nightmare that would be.
Had there been strong copyrights, would Sun have made better date/time APIs? Would Microsoft not have ended up creating code that requires:
String.Split(s, new char[] { ',' }, StringSplitOptions.RemoveEmptyEntries)
I can't see any positive difference in the world by copyrighting APIs.
Do remember that if you independently come up with the same API for a function without copying that is not infringement (copyright is not patents).
But I'm nervous that there are at least a few ways to make the argument, to a layperson (like a judge or jury), and that Oracle might succeed eventually.
Being able to copyright APIs would just mean that everyone would be incentivized to switch to open-source APIs with clear licensing terms (BSD, GPL, etc). I don't see that as a bad thing.
However, thesauruses are hard. See http://www.itworld.com/cloud-computing/379566/don-t-go-progr...
The other thing is that it gives a incentive to avoid platforms that have liberal API licenses.
Specifically in the Java case Google could I think have used the API under the GPL but chose not to.
They have also taken APIs from free software and written proprietary implementations. They would need to release entire huge projects under GPL.
First, SQL is a standardized language. It's not copyrighted, and has an ISO standard. Anybody can implement a SQL database and not run into any copyright problems. They might run into problems if they implemented proprietary Oracle extensions in their new DB, but even that seems unlikely.
Next, even if APIs are copyrightable, it doesn't imply Oracle would have to release anything as open source. For one, I'm not sure any open source projects would sue over it. And even if they did, Oracle could simply stop distributing the software.
What does being standardized have to do with it? MP3 and h.264 are standardized, that doesn't prevent them from being heavily patented. Likewise being standardized wouldn't prevent an API from being copyrighted.
> It's not copyrighted
If APIs were copyrightable, they would automatically be copyrighted when they were written down in a fixed form. That's the way that copyright works; you don't have to take special action to copyright something, it is automatically copyrighted when you create it. Registering a copyright puts you in a stronger position legally but it is not required to actually have a copyright.
See the danger now? If APIs were copyrightable, every API would automatically be copyrighted. Implementing an existing API would be copyright infringement unless you had a license.
...
SQL became a standard of the American National Standards Institute (ANSI) in 1986, and of the International Organization for Standards (ISO) in 1987.
So does time of infringement matter in copyright cases? Is the fact that it became an ISO standard later relevant?
Unless Oracle was psychic, they could not have foreseen SQL becoming standardized, right?
Pre-standard implementations had copyrights. Under this ruling the standard itself would violate those. There might be some sort of promissory estoppel type thing if all of the owners of the implementations also participated in the standardization process.
In the US post-1978 works don't even have to be registered to receive copyright protection--good luck tracking every significant bit of API lineage down and not getting hit by a troll that acquired the assets of some dinky company.
I always hoped it'd resume someday, or rise as a phoenix from its ashes. Where did the community go after the sign-off?
... this made me laugh. I find Javas APIs horrific and muddled. Surely not one of the "most appealing" parts?
If the appeal finds for Oracle I'll laugh... Google at least hastened Sun's demise by being too cheap to pay for Java, then they were too cheap to buy Sun when it was for sale. If they end up having to pay Oracle for it, that's karma at work.
And how much work did Tim Berners-Lee put into HTTP/HTML, the Unicode Consortium into Unicode, the W3C into XML, etc? The Java ecosystem depends heavily on these and many more. What kind of licensing agreement does Oracle have with the W3C? If you can own copyright on an API, I can't imagine what argument you would use to justify implementing a protocol without licensing it first.
Some parts of the Java API implement APIs that others have designed. Should Oracle need a licensing agreement with the W3C to implement the DOM?
Some parts of the Java API are thin wrappers around existing APIs. Is the Java API a derivative work of CORBA? Does Oracle need a licensing agreement with OMG?
Tim Berners-Lee specifically granted permission and encouraged people to adopt the WWW and create their own implementations. The W3C allows people to use its standards freely too (I'm not sure if there is a formal license grant or not).
I don't know the status of CORBA but it is plausible that Sun/Oracle did sign agreements to adopt those specifications or that they were properly granted to all comers.
Sum encouraged adoption of Java but was always specific and controlling of the distribution of runtimes and development kits. There were always licenses to agree to and real contracts available to people who wanted to develop their own Java platforms. They were quite strict on compatibility and what versions should be supported on what sorts of devices. They did release the JDK under the GPL. If Google had declared that they were using the GPL licensed code and were obeying the terms then they would have been entitled to proceed but they did not do that. They chose to develop the incompatible Dalvik to deliberately avoid Sun's contractual obligations but still benefit from the API design, documentation and existing developer experience expensively created by Sun.
> I'm not sure if there is a formal license grant or not
Where is the actual license? If formats are truly copyrightable, then general "encouragement" is not sufficient, you would need an explicit license, as every open source project has to allow copying of copyrighted code. Where is the actual document that says you can use HTML without infringing the copyright of Tim Berners-Lee or the W3C?
The answer is that there isn't one, because no one thinks that formats themselves are copyrightable.
> They chose to develop the incompatible Dalvik to deliberately avoid Sun's contractual obligations but still benefit from the API design, documentation and existing developer experience expensively created by Sun.
Writing books about your technology does not mean you somehow own the experience that people gain from reading them. Spending a lot of money does not entitle you to exert however much control you feel you deserve.
"Intellectual property" is a balance between the interests of creators and the interests of society. Giving first movers complete control over APIs, formats, and languages that become popular would enable a new wave of rent-seeking, very much like the patent trolling we see now.
OK I decided to actually look it up. Here it is: http://info.cern.ch/hypertext/WWW/Policy.html
"The definition of protocols such as HTTP and data formats such as HTML are in the public domain and may be freely used by anyone."
So back when originally developed TBL seems to have felt the need to declare the public domain nature (implying to me that it was NOT an obvious default).
> Writing books about your technology does not mean you somehow own the experience that people gain from reading them.
I agree. The creation and design of the API is what allowed Sun to do that. The documentation is merely part of the why Google decided to copy Java and many of it's libraries rather than develop their own API or use a freely licensed one. [Edit: I didn't really mean to agree with "owning the experience" bit but it does give them some rights.]
> Spending a lot of money does not entitle you to exert however much control you feel you deserve.
Agree. Copyright gives specific rights and there are also fair use rights that should possibly be strengthened.
> "Intellectual property" is a balance between the interests of creators and the interests of society.
I largely agree. I would like it the case that I could fully agree but I don't think that balance currently exists (especially on copyright durations and I think for software the source should be published or escrowed and then released when copyright expires). I also don't like discussing "Intellectual property" as that leads to sloppy thinking and overbroad claims, it is better to discuss the specific types, patents, copyright, design patents, trademarks separately. The balance is also set by legislation and the judiciary should not have that much discretion (although they should have on fair use issues).
> Giving first movers complete control over APIs, formats, and languages that become popular would enable a new wave of rent-seeking, very much like the patent trolling we see now.
The trade off is the incentive to create and promote new platforms and the breadth of the right to implement them. For me this is an issue of fair use and competition law plus market solutions of people avoiding platforms where alternative implementations are not possible. There is a useful discussion to be had about fair use rights to copyrighted API's but I'm not sure the incompatible Dalvik would be fair use even if a compatible alternative was, we are deep in the grey area here.
I would add as I have in other places that you need to consider how you would feel if it was old Sun in the Oracle role and MS in the Google role? Would this not be an egregious case of "embrace, extend extinguish"?
The definitions: I'm pretty sure this means that you can copy the documents themselves. This distinguishes it from standards documents by bodies like ISO where the standards documents themselves are copyright.
Here's another example: in the SCO lawsuit against IBM, why did they not attempt to claim copyright infringement over APIs? That would have been a much easier case to make than copyright infringement of the code, as Linux does in fact "copy" the APIs of UNIX.
Also, as another poster pointed out, if APIs were copyrightable then BSD UNIX would never have happened.
> The trade off is the incentive to create and promote new platforms and the breadth of the right to implement them.
No one would be incentivized to create platforms unless they could lock out competitors? WINE/ReactOS hasn't stopped Microsoft from developing the Windows API. Mono hasn't stopped them from developing .NET. Companies that create platforms get natural benefits from developing platforms: it puts them in a position of leadership and ahead of everybody else.
> I would add as I have in other places that you need to consider how you would feel if it was old Sun in the Oracle role and MS in the Google role? Would this not be an egregious case of "embrace, extend extinguish"?
Copyright law is not a tool for preventing anticompetitive business practices. It's not copyright's job to prevent "embrace, extend, extinguish."
But if Microsoft (for example) decided to implement the Dart APIs as their development environment for apps on Windows Phone, I don't see the problem with that.
"Embrace, extend, extinguish" would be if Google took a platform that already had a J2ME ecosystem and started distributing their own J2ME SDK that was targeted at the same developers and devices and introduced proprietary extensions that would make the Google-J2ME apps incompatible with the Oracle-J2ME apps.
Check your BSD history. Various actions and omissions of copyright registration and explicit copyright notices which were then required led to the BSD settlement.
I didn't say there could be no incentive without copyright but that it is part of the balance. .NET doesn't support the argument that much though, it could be seen as a counter reaction to losing the legal action over the J++ embrace, extend, extinguish attempt against Sun's Java.
I see embrace, extend in a wider way but that isn't a key argument.
I'm inclined to think Alsup's basis of "sequence, structure, and organization" not being copyrightable is far too broad a change, with ramifications beyond this specific case. (This is what allows lists, like a phone book, to be copyrightable, even though the numbers are public).
I'd have thought the reverse engineering exceptions were a more apt basis http://en.wikipedia.org/wiki/Reverse_engineering#Legality However, that's where Oracle's argument about the "fan-base" comes in: it wasn't purely interoperability between machines that Google wanted, but the installed base of developers. I haven't followed the exact policy considerations for reverse engineering, but it seems to me that the training of engineers to work on a particular machine is part-and-parcel of interoperability (for example, third-party after-market car parts - not only do they fit, but mechanics already know how to fit them; a more relevant example is IBM mainframes/minicomputers being copied, where programmers already knew the platform).
So, the higher court's consideration of the bigger picture may have serious financial consequences for Google and Oracle. I think Oracle will still lose, just on a different basis.
It's good to be appellate judge.
Microsoft’s Community Promise has made the .Net runtime and C# specifications available to Miguel de Icaza and the Mono project developers.
“If a .Net port to Android was through Mono it would fall under that agreement,” he said.
http://www.techworld.com.au/article/358564/microsoft_won_t_s...
http://blog.xamarin.com/android-in-c-sharp/
And has some cute benchmarks to show off for it.
But go one more level deep: Despite Microsoft's Community Promise, many places and legal teams will still view a potential threat if MS has a copyright on APIs used in Mono (not to mention patents).
It's a great win-win for Microsoft. They get to trumpet their PR and gain face with open source people, as well as deflect some criticism. But they know it'll still not hurt them as much as if they officially blessed Mono with a full, proper, transitive license.
Read the actual promise[1].
The promise only covers patent claims from Microsoft on those technologies, and only if you're not engaged in any patent lawsuit with them. So, if you're some random company that happens to encroach on Microsoft's territory in another area, but you happened to use Mono, they can strike at your usage of Mono due to your work in another area.
Automotive? Health? Entertainment? "Social"? Telecommunications? Advertising? Resource planning? Imaging/mapping? Security? Aeronautics? Anything related to computing?
All those and more are areas that Microsoft works on. I can't imagine any legal team being able to assert that their company will _never_ use Microsoft IP in any area - it's just too vast.
Example: You use Mono to implement some coordination software for networking gear - that sounds "safe". Microsoft then publishes research and obtains patents on their own work on networking (already done) from Azure and MSR (or they buy some large dead company's portfolio). Microsoft now has IP all over your area of choice, and as soon as there's any fight, you lose your Promise over Mono and now have a ton more claims to defend. Oops.
This isn't really even cynicism - it's spelled out very clearly in the promise.
Edit: In the Google Android .NET issue: Yes, Google can go use .NET for Android, no problem. Until Google gets in a fight over a plethora of other patents (say, advertising or other phone related patents). At that point, Google loses its promise for Mono and MS has a lucrative ground to gain.
Also, copyright issues are NOT included in the Community Promise. So MS still wins by making APIs copyrightable. They might extend the CP to cover copyright, but those copyright issues are still there as soon as someone fights MS over anything else. Totally strengthens MS's position.
As far as .NET on Android - OF COURSE, MS won't "stop" that. That'd be a major win for MS and hurt Oracle/Java. The .NET port would fall under the agreement, but I've already demonstrated how quickly that agreement can roll back. Google adopting .NET is a good technical move, but gives MS a massive advantage.
That said, app developers using MonoDroid are probably reasonably safe so long the apps they make are of relatively low strategic value to MS. If all of Google Play's apps were made on MonoDroid, Google would be in a terrible position.
1: http://www.microsoft.com/openspecifications/en/us/programs/c...
So even if the judge's decision is overturned the jury's decision still stands and there was no infringement.
http://es.scribd.com/doc/92725855/completed-verdict-form-in-...
I'm not familiar with the Java ecosystem. I appreciate if you could correct me.
Only a lawyer...
Of course, Sun wouldn't have sued in the first place...
Yes they would, but were out of money.
http://news.cnet.com/8301-1035_3-57423754-94/java-creator-ja...
"Just because Sun didn't have patent suits in our genetic code doesn't mean we didn't feel wronged. While I have differences with Oracle, in this case they are in the right. Google totally slimed Sun. We were all really disturbed, even Jonathan: he just decided to put on a happy face and tried to turn lemons into lemonade."