They've tried to (NaCl, Pepper, etc) and are still trying to.
Native Client isn't a proprietary format, though I suppose it uses some in that it uses is x86 or ARM machine code. But Native Client is a transitional technology on the route to Portable Native Client whose format is LLVM bitcode.
Pepper is an API which is non-proprietary API. Its currently Chrome-only, but that's different than proprietary -- Google hasn't erected any barriers to other browser vendors implementing Pepper, and in fact encourages it.
'reinterpreting' standards such that third party options no longer interoperate properly
You mean like the WebKit-only gmail offline support using a proposed standard that the W3C decided was not going to become an actual standard because it would be bad for the web?
No, not like that, becuase that didn't "reinterpret" a standard, or cause third-party options to no longer interoperate propertly.
Adding functionality using non-standard-but-open-specification technology when there isn't, at the time, any standard method, in a way which doesn't harm any existing third-party product integration is very different than misapplying a standard in a way which breaks existing third-party product integration.
Does MicroSoft seriously think people make purchasing decisions based on a 1% price difference?
No, it thinks the patent licensing fees for the features involved go to Microsoft competitors, so it is trying to reduce the amount of money those competitors make distributing fewer units that require paying those fees. Separating out the features and making the consumer pay the licensing fees (that's virtually all of the involved "marginal cost" to Microsoft of the features) is likely to be a fairly effective way of doing that.
The cost to Microsoft of including a few extra megabytes on the install media (or on the disk of a machine sold with Windows) is zero.
Since it takes extra machine time to create and transfer larger files (and they distribute both physical and ISO versions of install media, so both are relevant), the marginal costs of manufacturing are not zero, though they may be very very small.
More significantly, the marginal cost of the per-unit patent licensing fees necessary to legally provide DVD playback in, e.g., the United States is, however, vastly greater than the marginal manufacturing cost, but still must be paid.
As for the appeal, IANAL but to the best of my knowledge, if the jury comes back that even if the APIs are copyrightable Google didn't infringe them, Google walks away. I'm not sure of any grounds for appealing the verdict of a finding by a jury when a verdict of not guilty is entered, and it generally takes gross malfeasance on behalf of some member of the court even in a guilty verdict.
"not guilty" and "guilty" verdicts are criminal verdicts, for which the standards may be somewhat different from civil cases and are, in any case, irrelevant here.
A civil jury verdict, whether it finds liability or finds no liability, can be set aside by the trial judge because that judge finds that the jury could not reasonably have found as it did given the evidence before it (this is particularly key in the case at issue on the Google and the APIs, since the jury instructions all but direct a finding on that before considering the "fair use" defense, using almost exactly the same language in characterizing what Google has admitted as is used in setting out the relevant standard for infringement), and it can be set aside on appeal for that reason as well as others, including any error of law affecting the evidence presented in the case, or the jury instructions, if that error would, in the eyes of the appeals court, likely have effected the decision of the jury.
Oracle can't complain about those instructions to the Jury
Wrong. Purported errors in jury instructions are a frequent basis for appeal of civil judgements.
That said, the US system of giving jurors $5/day is hideously stupid, and the process for disqualifying jurors and allowing jurors to avoid jury duty leads to bad trial outcomes. When I got called up last year I had no opportunity to get out of it(though I could delay it for up to 6 months), and I got paid my full salary by my employer who was reimbursed by the state
The last time I got called for jury duty (which was a little over a year ago), I had no opportunity to get out of it (aside from asserting that I was a non-citizen, incompetent in English, or otherwise legally barred from jury service) though I could delay it for up to six months, and I got paid my full salary by my employer who was the state. This was, btw, in the US.
It is a bit more subtle than that. The higher courts do not rule on issues of fact.
Its actually considerably more subtle than even that, but the short version is that you're wrong. See, US Constitution, Art. III, Sec. 2: "...the Supreme Court shall have appellate jurisdiction, both as to law and fact..."
It would have been more correct for the grandparent post to have said that jury verdicts cannot be appealed directly.
Well, this is true in the extent that jury verdicts aren't a final act and, as such, aren't normally subject to appeal (and aren't generally subject to the exceptions that allow interlocutory appeals.)
Oracle would have to say that the judge erred on a matter of law that led to the incorrect jury decision.
That's true as a matter of form (since it is the judgement entered rather than the verdict supporting it which is appealed), but not as a matter of substance. Among the errors of law that can form the basis of the appeal is the error of not throwing out the jury verdict to enter a judgement notwithstanding the verdict because the facts presented in the case do not support the verdict the jury returned.
OTOH, if the judge makes the decision that the APIs are copyrightable that decision creates precedence(as does the reverse of course).
Actually, it doesn't. Trial court decisions may be "persuasive authority" -- the same category of legal authority as scholarly articles on the law, decisions of courts of foreign jurisdictions, and, say, references to non-legal works like dictionaries -- but they are not binding precedent.
Now, the almost-inevitable appeal may create precedent, but that's a different issue.
If on the other hand, the jury decides that, even if they were copyrightable Google isn't guilty, then the whole problem goes away and the judge doesn't have to decide and no one has to appeal the decision.
Oracle's grounds for appeal in that case are previewed in their objections to the jury instructions. To think that there is any possible trial outcome in this case that doesn't result in an appeal is naive. In fact, unless all the separate issues are resolved in favor of the same party, its quite likely that both parties will appeal on different grounds.
What sending the fact question related to the disputed legal issue to the jury before deciding it does is reduce the risk that, if the near-certain appeal overturns the trial court decision, the necessary result of that will be starting over, empaneling a new jury, and holding a new trial.
No. Oracle claimed that Google copied some code, that Google showed that was some code that they contributed to Java.
Actually, literal exact copying is still an issue in the case, and is part of the charge to the jury. If you read the jury instructions rather than relying on what other people tell you is the issue, you might be better informed.
Oracle also claimed that the APIs themselves are copyrighted, and that is what is being resolved now.
The issue that has been characterized imprecisely in popular media as "the APIs themselves" being copyrighted (which isn't the only issue in the charge to the jury to resolve) is not about "the APIs themselves" being copyrighted. The actual legal issue is about "sequence, structure, and organization" of the of the source files that are undisputedly part of Oracle's copyrighted Java implementation being protected.
Modded +5 Informative, and yet dead wrong (at least, as it relates to the present case.)
You can't appeal the finding of the jury
It is true that in the US legal system, you can't appeal a jury's not guilty verdict in a criminal case.
However, you can certainly appeal a jury verdict in all other situations (e.g., a guilty verdict in a criminal case, or any verdict in a civil case.) And, even before getting to appeal, the trial judge can set aside the jury verdict (again, except in the case of a not-guilty verdict in a criminal case.)
Ah, Jury Nullification. Something that should happen far more than it does. Many juries are even told that they can't rule on the legality of the law in question - total bullshit, but that's the system we live in.
While it may be total bullshit in a criminal case, it isn't in a civil case, since a jury verdict in either direction can be tossed out if the court (either the trial court or an appellate court) finds that the verdict is unsupportable given the facts in the case.
Nullification only works to acquit defendants in criminal cases.
I am not sure what jurisdiction you are referring to,
I'm pretty sure JNOV's can be returned in civil cases in every jurisdiction in the United States.
but if that is the case I would certainly hope that he would have to support his supposition with some kind of evidence.
The "evidence" is the actual evidence presented to the jury. A JNOV is a finding by the judge that the jury could not have reasonably reached the conclusions it did on the basis of the evidence before them in the case.
Given the Berkeley-BSD/AT&T case, and the Thread-X/Green-Hills case, I think there's already precedent on the copyrightability of API's.
The issue isn't actually the copyrightability of APIs qua APIs, its whether the sequence, structure, and organization of the code in what is undisputedly copyrighted code files is subject to protection under copyright. The effect of finding that they are would be that implementing APIs by copying the sequence, structure, and organization from copyright-protected source files without a license would be copyright infringement, but it would not, per se, be a finding that the API itself was protected by copyright.
They are essentially free add-on software that is supported by the free OS vendor, but they are no more part of the OS than Microsoft Office is part of Windows.
The judge told the jury to ASSUME that it is copyrightable and that the actual choice of copyrightability was his decision to make. This makes it so that appeals court will have an easier time based off his decision by not having to bring the jury back.
In practice, IIRC, the jury wouldn't be brought back. What would happen if there was a necessary question of fact that hadn't been heard by the jury after the appeals court had made determinations of law and remanded the case for further proceedings would be that a new trial would be ordered, because the old trial jury would have been dismissed (and would no longer have the benefit of immediate temporal proximity to the original trial presentation.)
Note also that the benefit here only applies if the judge ends up ruling that the sequence, structure, and oranization of the APIs aren't subject to copyright protection. If he rules that they are, then whether he does so before or after the jury gets its instructions is immaterial. But if he rules they are not, then doing so before the jury instructions would make the questions legally irrelevant.
So, while its possible that the judge really is punting this and doesn't have a leaning on a legal issue that has been briefed by both sides repeatedly in the back and forth over the jury instructions, it seems to me more likely that the judge has a pretty good idea which way he is going to decide the legal issue, and it isn't the way lots of people here are afraid it might be.
But registration increases the damages that you may claim should you bring a case to court.
This is somewhat misleading. Registration (in US law, at least) is a prerequisite for bringing most copyright claims (I don't think there is any cause of action for which registration increases the available damages -- mostly, its completely binary.) But the registration doesn't have to precede the alleged violation, it just has to precede filing the lawsuit.
In C-style languages, the && operator is short-circuited, so you put the easier/less-computation bit first.
Yeah, and that's where the analogy to C-style languages fails -- the court in this case has separate specialized processing units (one unit called the "trier of law" or "judge" and a 9 others collectively called the "trier of fact" or "jury".)
One branch of this logical-and operation can only run on the "judge" processing unit, and the other has to be run as part of the "jury deliberation" task which involves the other 9 processors working together. Serializing them is not efficient.
This whole thing seems rather silly. First of all, yes, clearly Google copied the Java API. How can they even begin to deny that?
The API in abstract terms is not the subject of the issue here, though some of the media sources make it seem that way. The actual code (both compilable and comment) in the source files providing the API is the issue.
From p. 9, lines 6-9: Among other things, the registered copyrights generally include the compilable code and documentation for the Java API packages. The main issues you must decide concern these two general types of material contained therein, namely “compilable code” and “documentation.”
That's actually one of Google's defenses. They didn't copy the entire Java API, just a portion of it. So no, if the ruling is in Oracle's favor (which is unlikely but not impossible), then you can't get away with fair-use.
There are three different questions that have to be resolved in Oracle's favor for a verdict in Oracle's favor on this point: 1. (legal) Are the APIs subject to copyright protection, 2. (factual) Did Google infringe the exclusive rights of Oracle as copyright holder of the APIs, 3. (factual) Was Google's infringement outside the scope of the exception carved out of the exclusive rights under fair use.
Each logically only requires a response if the preceding question is answered in the affirmative. #2 is presented to the jury as a question they must answer as if #1 were answered in the affirmative, though the court has not ruled on it; #3 is presented to the jury as a question they must answer if and only if they answer yes to #2.
A ruling in Oracle's favor on the legal issue (#1) would not obviously not mean fair use was not available as a defense. Nor, obviously, would a finding of fact in Oracle's favor on #2.
But even a verdict in Oracle's favor (meaning that all three issues were resolved in Oracle's favor) wouldn't mean that fair use wasn't available, it would just mean that the jury in this particular case didn't find Google's particular use of Oracle's copyright-protected APIs was fair use. Fact findings are not, however, precedential or binding on future triers of fact, only legal rulings are. (And the only legal ruling related to the availability of fair use in this case is that, at a minimum, it is an available defense, because if it wasn't, question #3 wouldn't be put to the jury at all, or would be put to the jury with the same restrictions as #2 in that it was waiting on a legal ruling on the applicability of the defense.)
The article claims Mono could be sued for copyright infringements, but actually Microsoft's public statements would count against any legal actions they could attempt against Miguel De Icaza.
Sure, but that's somewhat tangential to the issue. That estoppel would be available as an offense isn't the point, the point is that affirmative defenses like estoppel (or, as asserted by Google in the Oracle v. Google case, fair use) would be necessary.
...which means it's one of the few languages/APIs that could survive unscathed...
Original languages and APIs, of course, escape unscathed, as do reimplementations licensed from the original. As do languages and APIs dedicated by the creator to the public domain.
I don't think its all that few.
which means Oracle's attempts to control Java could end up sending EVERYONE, including the GNU/Linux community who, thus far, have given.NET the cold shoulder, into the embrace of its earnest rival, destroying Java completely.
Wow. That's more hyperbolic than the headline of TFA/S. Which is pretty amazing.
Seems a bit non-sensical to me: "You, spend the next several weeks doing a bunch of work based on an assumption that I may invalidate, but only after you are done."
Since the jury will have ruled on the premise that the APIs are protected by copyright, and since whatever the judge rules on the legal question is bound to appealed by one side or the other, the question of fact will have been resolved by the jury if it turns out to be relevant either based on the trial judges decision or based on an appellate decision.
Because the trial jury will have decided the issue, the likelihood of further proceedings being necessary after appeal will be reduced.
One might read this as an indication that the judge is going to rule against APIs being protected, since nothing would be saved by this if he was going to rule in favor of them being protected. But if he issued that ruling before the questions were put to the jury, then the question couldn't be included in the jury instruction, since it would no longer be a relevant question of fact.
The actual jury instructions make this clear: what is assumed copyrightable in those instructions is both the compilable (non-comment code) and documentation (comments) in the source files implementing the 37 APIs for Java that are at issue.
Does this mean that we can copyright an API in a broad generic manner? i.e. a "web service that returns a JSON object of apps in an app store".
No.
Or is the specific manner in which the API fetches the information the part that is copyrightable?
Would this also possibly imply that a programmer needs to get a license to code in a specific language or to utilize a specific API?
I think it would take another stretch of the law to reach that conclusion beyond what is in the jury instructions in this case (except in the case of, e.g., statically linking the code implementing the API, but that that requires a license is, if not necessarily established law, at least a fairly common understanding of the present law -- one on which, for instance, the GPL already is premised.)
Would this also possibly imply that a programmer needs to get a license to code in a specific language or to utilize a specific API?
Nothing stops them from adding that as a licensing term on the JDK whether or not it requires a license under current law. Since using the JDK itself requires a license, nothing prevents that license from being conditioned on the user agreeing to refrain from acts which they would have free reign to do if they weren't restrained by the license.
So, Jython, IronPython, and PyPy for Python; JRuby, IronRuby, and Rubinius for Ruby; Mono for C# and VB; possibly C++ for C, GCC for C and C++ and Objective-C; and so forth. And of course, all the various browsers that use JavaScript might owe royalties to the acquirers of Netscape's intellectual property.
Many of these are wrong, because many of the examples there are not done "without a license". Many of them are licensed reimplementations of open source originals that share more than just APIs with the original. AFAIK, Rubinius, JRuby, and IronRuby aren't clean-room reimplementations of Ruby, they are parallel open source projects that use code (including standard library code) from the mainline Ruby project (and I'm pretty sure that code moves both directions.)
Slashdot Mirror
Native Client isn't a proprietary format, though I suppose it uses some in that it uses is x86 or ARM machine code. But Native Client is a transitional technology on the route to Portable Native Client whose format is LLVM bitcode.
Pepper is an API which is non-proprietary API. Its currently Chrome-only, but that's different than proprietary -- Google hasn't erected any barriers to other browser vendors implementing Pepper, and in fact encourages it.
No, not like that, becuase that didn't "reinterpret" a standard, or cause third-party options to no longer interoperate propertly.
Adding functionality using non-standard-but-open-specification technology when there isn't, at the time, any standard method, in a way which doesn't harm any existing third-party product integration is very different than misapplying a standard in a way which breaks existing third-party product integration.
No, it thinks the patent licensing fees for the features involved go to Microsoft competitors, so it is trying to reduce the amount of money those competitors make distributing fewer units that require paying those fees. Separating out the features and making the consumer pay the licensing fees (that's virtually all of the involved "marginal cost" to Microsoft of the features) is likely to be a fairly effective way of doing that.
Apparently, much better than you do.
Since it takes extra machine time to create and transfer larger files (and they distribute both physical and ISO versions of install media, so both are relevant), the marginal costs of manufacturing are not zero, though they may be very very small.
More significantly, the marginal cost of the per-unit patent licensing fees necessary to legally provide DVD playback in, e.g., the United States is, however, vastly greater than the marginal manufacturing cost, but still must be paid.
Juries ignoring jury instructions in civil cases are very rarely done.
Well, yes, where "nothing" means "they can directly throw it out".
No, actually, one of the benefits of the time I spent in law school is that it is very hard to confuse me with "lawyer speak".
"not guilty" and "guilty" verdicts are criminal verdicts, for which the standards may be somewhat different from civil cases and are, in any case, irrelevant here.
A civil jury verdict, whether it finds liability or finds no liability, can be set aside by the trial judge because that judge finds that the jury could not reasonably have found as it did given the evidence before it (this is particularly key in the case at issue on the Google and the APIs, since the jury instructions all but direct a finding on that before considering the "fair use" defense, using almost exactly the same language in characterizing what Google has admitted as is used in setting out the relevant standard for infringement), and it can be set aside on appeal for that reason as well as others, including any error of law affecting the evidence presented in the case, or the jury instructions, if that error would, in the eyes of the appeals court, likely have effected the decision of the jury.
Wrong. Purported errors in jury instructions are a frequent basis for appeal of civil judgements.
The last time I got called for jury duty (which was a little over a year ago), I had no opportunity to get out of it (aside from asserting that I was a non-citizen, incompetent in English, or otherwise legally barred from jury service) though I could delay it for up to six months, and I got paid my full salary by my employer who was the state. This was, btw, in the US.
Its actually considerably more subtle than even that, but the short version is that you're wrong. See, US Constitution, Art. III, Sec. 2: "...the Supreme Court shall have appellate jurisdiction, both as to law and fact..."
Well, this is true in the extent that jury verdicts aren't a final act and, as such, aren't normally subject to appeal (and aren't generally subject to the exceptions that allow interlocutory appeals.)
That's true as a matter of form (since it is the judgement entered rather than the verdict supporting it which is appealed), but not as a matter of substance. Among the errors of law that can form the basis of the appeal is the error of not throwing out the jury verdict to enter a judgement notwithstanding the verdict because the facts presented in the case do not support the verdict the jury returned.
Actually, it doesn't. Trial court decisions may be "persuasive authority" -- the same category of legal authority as scholarly articles on the law, decisions of courts of foreign jurisdictions, and, say, references to non-legal works like dictionaries -- but they are not binding precedent.
Now, the almost-inevitable appeal may create precedent, but that's a different issue.
Oracle's grounds for appeal in that case are previewed in their objections to the jury instructions. To think that there is any possible trial outcome in this case that doesn't result in an appeal is naive. In fact, unless all the separate issues are resolved in favor of the same party, its quite likely that both parties will appeal on different grounds.
What sending the fact question related to the disputed legal issue to the jury before deciding it does is reduce the risk that, if the near-certain appeal overturns the trial court decision, the necessary result of that will be starting over, empaneling a new jury, and holding a new trial.
Actually, literal exact copying is still an issue in the case, and is part of the charge to the jury. If you read the jury instructions rather than relying on what other people tell you is the issue, you might be better informed.
The issue that has been characterized imprecisely in popular media as "the APIs themselves" being copyrighted (which isn't the only issue in the charge to the jury to resolve) is not about "the APIs themselves" being copyrighted. The actual legal issue is about "sequence, structure, and organization" of the of the source files that are undisputedly part of Oracle's copyrighted Java implementation being protected.
Yeah, there is. Its called a judgement notwithstanding the verdict (or judgement non obstante veredicto, or JNOV.)
The only case where absolute jury nullification is possible in the US system is when the jury acquits despite the law in a criminal case.
Modded +5 Informative, and yet dead wrong (at least, as it relates to the present case.)
It is true that in the US legal system, you can't appeal a jury's not guilty verdict in a criminal case.
However, you can certainly appeal a jury verdict in all other situations (e.g., a guilty verdict in a criminal case, or any verdict in a civil case.) And, even before getting to appeal, the trial judge can set aside the jury verdict (again, except in the case of a not-guilty verdict in a criminal case.)
While it may be total bullshit in a criminal case, it isn't in a civil case, since a jury verdict in either direction can be tossed out if the court (either the trial court or an appellate court) finds that the verdict is unsupportable given the facts in the case.
Nullification only works to acquit defendants in criminal cases.
I'm pretty sure JNOV's can be returned in civil cases in every jurisdiction in the United States.
The "evidence" is the actual evidence presented to the jury. A JNOV is a finding by the judge that the jury could not have reasonably reached the conclusions it did on the basis of the evidence before them in the case.
The issue isn't actually the copyrightability of APIs qua APIs, its whether the sequence, structure, and organization of the code in what is undisputedly copyrighted code files is subject to protection under copyright. The effect of finding that they are would be that implementing APIs by copying the sequence, structure, and organization from copyright-protected source files without a license would be copyright infringement, but it would not, per se, be a finding that the API itself was protected by copyright.
Or, more importantly, the base install.
They are essentially free add-on software that is supported by the free OS vendor, but they are no more part of the OS than Microsoft Office is part of Windows.
In practice, IIRC, the jury wouldn't be brought back. What would happen if there was a necessary question of fact that hadn't been heard by the jury after the appeals court had made determinations of law and remanded the case for further proceedings would be that a new trial would be ordered, because the old trial jury would have been dismissed (and would no longer have the benefit of immediate temporal proximity to the original trial presentation.)
Note also that the benefit here only applies if the judge ends up ruling that the sequence, structure, and oranization of the APIs aren't subject to copyright protection. If he rules that they are, then whether he does so before or after the jury gets its instructions is immaterial. But if he rules they are not, then doing so before the jury instructions would make the questions legally irrelevant.
So, while its possible that the judge really is punting this and doesn't have a leaning on a legal issue that has been briefed by both sides repeatedly in the back and forth over the jury instructions, it seems to me more likely that the judge has a pretty good idea which way he is going to decide the legal issue, and it isn't the way lots of people here are afraid it might be.
This is somewhat misleading. Registration (in US law, at least) is a prerequisite for bringing most copyright claims (I don't think there is any cause of action for which registration increases the available damages -- mostly, its completely binary.) But the registration doesn't have to precede the alleged violation, it just has to precede filing the lawsuit.
Yeah, and that's where the analogy to C-style languages fails -- the court in this case has separate specialized processing units (one unit called the "trier of law" or "judge" and a 9 others collectively called the "trier of fact" or "jury".)
One branch of this logical-and operation can only run on the "judge" processing unit, and the other has to be run as part of the "jury deliberation" task which involves the other 9 processors working together. Serializing them is not efficient.
The API in abstract terms is not the subject of the issue here, though some of the media sources make it seem that way. The actual code (both compilable and comment) in the source files providing the API is the issue.
From p. 9, lines 6-9:
Among other things, the registered copyrights generally include the compilable code and documentation for the Java API packages. The main issues you must decide concern these two general types of material contained therein, namely “compilable code” and “documentation.”
There are three different questions that have to be resolved in Oracle's favor for a verdict in Oracle's favor on this point:
1. (legal) Are the APIs subject to copyright protection,
2. (factual) Did Google infringe the exclusive rights of Oracle as copyright holder of the APIs,
3. (factual) Was Google's infringement outside the scope of the exception carved out of the exclusive rights under fair use.
Each logically only requires a response if the preceding question is answered in the affirmative. #2 is presented to the jury as a question they must answer as if #1 were answered in the affirmative, though the court has not ruled on it; #3 is presented to the jury as a question they must answer if and only if they answer yes to #2.
A ruling in Oracle's favor on the legal issue (#1) would not obviously not mean fair use was not available as a defense. Nor, obviously, would a finding of fact in Oracle's favor on #2.
But even a verdict in Oracle's favor (meaning that all three issues were resolved in Oracle's favor) wouldn't mean that fair use wasn't available, it would just mean that the jury in this particular case didn't find Google's particular use of Oracle's copyright-protected APIs was fair use. Fact findings are not, however, precedential or binding on future triers of fact, only legal rulings are. (And the only legal ruling related to the availability of fair use in this case is that, at a minimum, it is an available defense, because if it wasn't, question #3 wouldn't be put to the jury at all, or would be put to the jury with the same restrictions as #2 in that it was waiting on a legal ruling on the applicability of the defense.)
Sure, but that's somewhat tangential to the issue. That estoppel would be available as an offense isn't the point, the point is that affirmative defenses like estoppel (or, as asserted by Google in the Oracle v. Google case, fair use) would be necessary.
Original languages and APIs, of course, escape unscathed, as do reimplementations licensed from the original. As do languages and APIs dedicated by the creator to the public domain.
I don't think its all that few.
Wow. That's more hyperbolic than the headline of TFA/S. Which is pretty amazing.
Since the jury will have ruled on the premise that the APIs are protected by copyright, and since whatever the judge rules on the legal question is bound to appealed by one side or the other, the question of fact will have been resolved by the jury if it turns out to be relevant either based on the trial judges decision or based on an appellate decision.
Because the trial jury will have decided the issue, the likelihood of further proceedings being necessary after appeal will be reduced.
One might read this as an indication that the judge is going to rule against APIs being protected, since nothing would be saved by this if he was going to rule in favor of them being protected. But if he issued that ruling before the questions were put to the jury, then the question couldn't be included in the jury instruction, since it would no longer be a relevant question of fact.
The actual jury instructions make this clear: what is assumed copyrightable in those instructions is both the compilable (non-comment code) and documentation (comments) in the source files implementing the 37 APIs for Java that are at issue.
No.
Not that, either.
I think it would take another stretch of the law to reach that conclusion beyond what is in the jury instructions in this case (except in the case of, e.g., statically linking the code implementing the API, but that that requires a license is, if not necessarily established law, at least a fairly common understanding of the present law -- one on which, for instance, the GPL already is premised.)
Nothing stops them from adding that as a licensing term on the JDK whether or not it requires a license under current law. Since using the JDK itself requires a license, nothing prevents that license from being conditioned on the user agreeing to refrain from acts which they would have free reign to do if they weren't restrained by the license.
Many of these are wrong, because many of the examples there are not done "without a license". Many of them are licensed reimplementations of open source originals that share more than just APIs with the original. AFAIK, Rubinius, JRuby, and IronRuby aren't clean-room reimplementations of Ruby, they are parallel open source projects that use code (including standard library code) from the mainline Ruby project (and I'm pretty sure that code moves both directions.)