In my view, the Oracle v. Google case may be headed in the same direction.
Judge Alsup has let the case go to the jury to determine whether the "structure, sequence and organization" (SSO) of the Android platform infringes either the compilable code or the documentation of 37 of the 166 total API packages constituting a part of the Java platform. But the judge has reserved for his exclusive determination, after a verdict is entered, whether such SSO is even protectible under U.S. copyright law. In instructing the jury, the judge merely told the jurors to assume for purposes of their deliberations that the SSO was protectible under copyright and to decide the case based on that assumption. This means that, if Oracle loses on the verdict, its copyright case is dead. If it wins, then and only then will the judge need to address the ultimate legal question of whether the SSO can be protected by copyright at all. If the case does reach that point, he will be able to decide the question based on a full presentation of evidence by both parties and that in turn will make any decision he renders far less vulnerable to being overturned on appeal.
I am not saying that the judge will necessarily support Google in its claim that APIs should not be covered by copyright. I do say that Judge Alsup is nobody’s fool and is playing this smart, letting Oracle play out its case in full and only then stepping in to say whether its case is even legally supportable.
Of course, the judge could also have rendered a definitive ruling one way or the other in the pretrial phase. But there are some areas of law where the principles are pretty clearly articulated but get knotty in their application and this is one of them. Yes, U.S. copyright law, like EU law, specifies that something that is functional or a system or a method of operation, etc. cannot be protected by copyright. But what exactly does that mean in the factual context of APIs and programming languages? Maybe those extremely well-versed in what APIs are and what they do can answer that clearly for others who are also well-versed. But in a court of law, a lot can get lost in translation and, in the absence of a clear factual record, skilled lawyers have all sorts of room to confuse and confound the issues as they apply abstract legal rules to a murky factual record. Because of this, the judge is letting it play out through a full jury verdict while reserving the right to rule definitively on the issue until after the verdict.
In this context, Oracle has plenty of reason to be nervous. If it loses before the jury, it loses; if it wins, it still may lose. And, as for the case being a bad precedent, that remains to be seen depending on how the judge and any appeals court may ultimately rule on this issue. For the moment, U.S. law remains sane on this point and nothing bad has happened as yet apart from an overbearing company attempting to pour vast resources in hopes of gaining short-term advantage to the long-term detriment of the developer community.
Note also that even if Alsup decides that APIs are not copyrightable, and knows now that this will be his decision, he needs to consider that a higher court may disagree.
If that happens, and the case comes back to him, and he had not sent infringement to the jury in this trial, he'd have to seat a new jury and conduct a new trial just on infringement. Yuck. By letting the jury consider the factual aspects now, Alsup is ensuring that one trial will be sufficient. This saves a lot of time and effort, at the slight cost of having the present jury consider something that may turn out moot.
> This means that, if Oracle loses on the verdict, its copcyright case is dead. If it wins, then and only then will the judge need to address the ultimate legal question of whether the SSO can be protected by copyright at all.
Isn't this bizarre though? I'm not a lawyer so I guess I am ignorant of how US courts work, but shouldn't it be clear whether there is even a relevant law for something before a jury is asked to make the effort of figuring out if a law applies in a particular case? It won't be just a matter of SSO is copyrightable yes/no, it will be more complicated. But the jury won't hear any of that before saying yes or no.
Shouldn't what the law is be first determined, and only then whether it applies to a particular case?
In the case where APIs are not protected, it doesn't matter what the jury thinks, there was nothing for Google to infringe. So there's no need to ask the jury about that and you can have them just assume that APIs are protected, because that's the only time when their verdict will matter.
Does this mean that if Oracle loses the case, the question of whether APIs are copyrightable will not be answered in this case? Or will the judge still give a ruling on that once the verdict is entered? The former being a big win for Google only, the latter being a big win for the industry.
This is ultimately more by luck than judgement. Because the main legal perspective is utterly, fundamentally flawed. The notion that there is a principled difference between idea and expression simply does not work with software. In software everything is an abstraction, which means everything has some generality. There is no dividing line anymore.
The idea/expression rule of thumb only made sense because it was a simple proxy for actual market effect. And there is the nub of the matter: no-one really wants to measure what the real market effects are for any of these rules. The law is unhinged from reality. It is just stupid.
For copyright to make sense, we should get rid of all the lawyers and remove it from the legal domain. Copyright is really a matter of data measurement and analysis. If the negative effects of restricting and paying for movement of information are outweighed by the positive effects of creating information, let us measure that . . .
We built the internet, but then we want to rule it as if it were a small stack of manuscripts. It is ridiculous.
When two people independently construct the same code with only trivial naming differences, what then? Is that a copyrightable expression or is it an uncopyrightable idea?
The problem I think GP is citing is that it's unreasonable when one piece of code violates copyright for carrying out some function, and another piece of code does not violate copyright for carrying out that function, even though after being run through some automated transformations (for example, an optimizing compiler), the machine code for the two pieces of code is near identical.
One person uses a for loop, another uses do while, another uses recursion. Copyright law doesn't care that they all do the same thing; one could be a copyright violation while the others are not.
If you abstract copyright to apply to the functionality of the code, you turn copyrights into patents. If you don't, anyone can tweak code to be literally different while having the same function.
The code is expression. Neither can enforce the code against the other; both can enforce against third parties (with registration).
Machine code cannot be copyrighted under U.S. law. It is not an issue in the Oracle/Google case.
You cannot abstract copyright. It is a codified legal doctrine (i.e., an expression), not an abstract set of ideas. On that note, however, the abstract idea behind copyright law (protection of intellectual property) is the justification for patent and trademark laws.
Copyright and patents are entirely 2 different issues.
This trial is very important because Oracle is suing for both and because copyright protection is much, much stronger than patents.
The differences: patents expire in 20 years, copyright is basically unlimited; patent infringement is hard to prove, copyright infringement is easy; patents are easily invalidated, while copyright is pretty clear (except in this case); getting a patent is expensive, while copyrighting something is essentially free, etc...
And copyright protection of interfaces to block interoperability has been struck down so many times it's explicitly written into the EU convention (article 4 IIRC)
Oracle could conceivably get a ruling on the 'Java' name but using copyright to block somebody else doing an interoperable implementation has been defeated before
I'm sure Oracle knows this and is simply going to trial in the hope of a random verdict (it happens) or to get Google to reveal lots of documents that can be used in a future patent / monopoly / trade case
Judge Alsup has let the case go to the jury to determine whether the "structure, sequence and organization" (SSO) of the Android platform infringes either the compilable code or the documentation of 37 of the 166 total API packages constituting a part of the Java platform. But the judge has reserved for his exclusive determination, after a verdict is entered, whether such SSO is even protectible under U.S. copyright law. In instructing the jury, the judge merely told the jurors to assume for purposes of their deliberations that the SSO was protectible under copyright and to decide the case based on that assumption. This means that, if Oracle loses on the verdict, its copyright case is dead. If it wins, then and only then will the judge need to address the ultimate legal question of whether the SSO can be protected by copyright at all. If the case does reach that point, he will be able to decide the question based on a full presentation of evidence by both parties and that in turn will make any decision he renders far less vulnerable to being overturned on appeal.
I am not saying that the judge will necessarily support Google in its claim that APIs should not be covered by copyright. I do say that Judge Alsup is nobody’s fool and is playing this smart, letting Oracle play out its case in full and only then stepping in to say whether its case is even legally supportable.
Of course, the judge could also have rendered a definitive ruling one way or the other in the pretrial phase. But there are some areas of law where the principles are pretty clearly articulated but get knotty in their application and this is one of them. Yes, U.S. copyright law, like EU law, specifies that something that is functional or a system or a method of operation, etc. cannot be protected by copyright. But what exactly does that mean in the factual context of APIs and programming languages? Maybe those extremely well-versed in what APIs are and what they do can answer that clearly for others who are also well-versed. But in a court of law, a lot can get lost in translation and, in the absence of a clear factual record, skilled lawyers have all sorts of room to confuse and confound the issues as they apply abstract legal rules to a murky factual record. Because of this, the judge is letting it play out through a full jury verdict while reserving the right to rule definitively on the issue until after the verdict.
In this context, Oracle has plenty of reason to be nervous. If it loses before the jury, it loses; if it wins, it still may lose. And, as for the case being a bad precedent, that remains to be seen depending on how the judge and any appeals court may ultimately rule on this issue. For the moment, U.S. law remains sane on this point and nothing bad has happened as yet apart from an overbearing company attempting to pour vast resources in hopes of gaining short-term advantage to the long-term detriment of the developer community.