Because it doesn't matter whether it was a contractor or not.
It certainly does in the context of your earlier suggestion that the decompiled files support some inference related to why Google chose to do it. Since it was a contractor doing something Google expressly forbade them to do, any speculation about why Google chose to do it rests on a fatally false premise.
Google is ultimately responsible for abiding by the law with the products they release
As has been pointed out multiple times in this subthread, most recently in GP -- to which the quoted material above is in reply -- no one is arguing against that point. So why do you keep raising it?
Your post ws posted at 2:57pm on Wednesday, May 16; the article you link to was posted on (and transcribes court discussions that occurred on) Tuesday, May 15. So, "today" probably isn't the word you are looking for.
More importantly, you've confused Oracle's lawyer with Google's. I've broadened the excerpt that you provided to include the part that shows who is talking (material that was excluded from your excerpt is bolded):
Oracle: I'm not an expert on Java -- this is my second case on Java, but I'm not an expert, and I probably couldn't program that in six months. Let me come back to rangeCheck after I've reminded the Court about the test files.
These test files were created by decompiling the Oracle code -- Sun code. That was not an accident, and not something they did by mistake. They did it intentionally, and they did it for the purpose of saving money, or time, or both. You can't look at those decompiled files and say there's no meaning in that; it accelerated Android, that's why they did it.
[PJ: Let me remind everyone that the testimony was that a contractor, Noser, did it, and it did it contrary to specific instructions from Google.]
Let me come back to rangeCheck. Every time I talk about this, I feel like I'm either saying it's important or it's not important. That's a false dichotomy. No one can say it's a big thing. But it was something that was copied, and it was important to TimSort, which had a significant performance advantage.
Well, he certainly claims not to be a "Java expert"; whether he is a coder or not is, I don't think, a point on which there is much basis to make a conclusion, nor is it particularly relevant in any case.
the idea is that Google copied the original java to get an idea on how the VM worked.
No, its not (well, not in the copyright portion of the trial) because ideas aren't protected by copyright, and not even Oracle's lawyers -- as close to it as they come with the API SSO issue -- are going to try to outright make "they stole our idea" the basis for a claim for copyright liability, since even the most clueless federal judge imaginable would have to throw out that argument.
(Obviously, there claim is that specific ideas were stolen in the patent phase, but that's a whole different ball of worms. And, no, the mixed metaphor is not accidental.)
They also aren't making the argument you suggest since, aside from ideas not being so clearly outside the scope of copyright protection that not even BS&F would try that argument, the decompiled files at issue aren't part of the VM at all, but part of Sun's compatibility test suite.
The 9 lines of code, by the way, aren't even the same issue as the decompiled files -- the 9 lines of code are code that a Google employee wrote and contributed both to Android and OpenJDK, and have nothing to do with the 8 files decompiled by Noser while under contract to Google.
PS = IMNAL but IRTFA:
Assuming the obvious interpretation of the acronyms there, I must point out that "read" and "understood" are two different things.
Only a Google employee in the sense that he worked for a 3rd party firm (Noser Engineering AG) Google hired to help create Android.
No, not at all. Josh Bloch worked for Google as Chief Java Architect (not for Noser) when he wrote rangeCheck (and still works for Google in the same role.) You seem to be confusing rangeCheck with the 8 decompiled files, which were done by Noser Engineering AG. Those are the two independent issues on which copyright infringement has been found in Oracle v. Google.
If the judge's statement is even a half-truth, the code in question checks a number, and passes it somewhere if it's a good number.
It does both more (checks more numbers) and less (doesn't deal with actually passing the numbers anywhere) than that: it actually checks two numbers against an array, and throws an exception if either of the numbers is a bad number. Passing the numbers somewhere else is the responsibility of the calling code.
I'm not talking about rangeCheck specificially but the entirety of code that the judge ruled was decompiled and copied from Java.
rangeCheck isn't even part of that body of code, which consists of 8 test files that were never distributed on handsets, but which part of the Android source distribution.
That leaves the question of whether the contractor copied other parts of the Java that are not as easy to identify or if they decompiled and copied then rewrote to make it look like it wasn't copied.
What parts of Android the contractor was involved in is, I believe, part of the trial record; it is, at any rate, information Oracle could have subpoenaed and addressed had it any interest in doing so.
Maybe they didn't do that with this function because it was trivial.
They didn't do anything with this function (rangeCheck) because they had nothing to do with it. It wasn't part of the decompiled files, the author of the function was a Google employee (who wrote it outside of the scope of his employment) who both included the file in Android and contributed it to OpenJDK.
Google pretty much admitted it was copied - claiming it was accidental (the same guy wrote it for Oracle and later, Google).
Actually, its not the "the same guy wrote it for Oracle and later, Google". It wasn't written twice.
Its that the Google employee who wrote it on his own time both included it in the Android source tree and contributed to to the OpenJDK (including the required copyright assignment.)
The question is whether it is worth anything. Google says no. Oracle's own expert witnesses said no. The judge - who has apparently revealed he is a programmer - says no. Oracle are arguing it's worth millions.
Actually, Oracle's pretty much given up on that, too. In a joint stipulation filed today, Oracle has agreed to waive both jury trial and its claim for actual damages and infringers profits on the copyright claims for which liability has already been found if the API SSO issue isn't resolved in its favor, and accept statutory damages for rangeCheck and the decompiled files.
It may be the contractor's fault that Sun's IP was violated but Google did reap the benefits of that violation. If it was the contractor's fault that the Sun's IP was violated, Sun now Oracle still has a valid claim for damages.
No one is arguing against that point, so I don't know why you keep arguing for it. My objection was your use of reference to the decompiled files and Google's supposed decision to copy them (which is refuted by the fact that the decompilation was done by a contractor under express direction from Google not to decompile any third-party files) as part of the backing for your post asking why, if it was so simple, did Google choose to copy?
It doesn't really matter who decompiled the files.
It doesn't matter to copyright liability (well, actually, it does, but it doesn't matter to Google's copyright liability.)
It does matter, however, when someone uses quotes about the decompiled file to backup an argument of the form "if its all so easy, why did Google copy it?"
Since, in fact, Google didn't copy it, and in fact the person who did copy it was doing something Google had expressly directed them not to do.
My elite math skills tell me they are spending $30 million dollars per year on Facebook, where none of that $30M can be accounted for by paid ads.
Some of it, presumably, is content development for stuff that gets distributed as not-ads on the facebook page. If the "paid ads" is only the amount they are paying for the advertising space, some of the $30million may be content development for the ads themselves.
Assuming it costs $50k/year for GM to pay someone to upload pictures of their cars, type status updates ("Looking forward to tomorrow's release of car X!" or "OMFG car X is sooo beautiful and fast, I don't even care what it costs!") I can't help but imagine they're paying 600 people to do that kind of work.
I would expect that in addition to paying people to do that via the Facebook user accounts or their company Facebook page, they are also paying people to write and edit/review approved "status updates", design and manage a coherent overall essaging strategy, take and edit professional photographs for distribution via Facebook, etc. And that many of those people cost GM more than $50k/yr (workers cost employers a lot more than the workers gross salary.)
"Judge William Alsup ruled that evidence presented during the trial had shown that Google infringed on Oracleâ(TM)s copyrights by decompiling eight Java files and copying them in their entirety for use with Android."
The decompiled files are a different issue than rangeCheck. Google didn't decompile them, Noser -- a contractor for Google -- did, against the express prohibition in the Google-Noser contract, which required all original work or open-source code, and expressly prohibited copying or decompiling proprietary code.
"Judge William Alsup ruled that evidence presented during the trial had shown that Google infringed on Oracleâ(TM)s copyrights by decompiling eight Java files and copying them in their entirety for use with Android."
Look Ballmer is a douche, no doubt. But worst CEO, compared to the putzes who ran almost every bank, Chrystler, and GM into bankruptcy? Compared to Scott Thompson? Jerry Yang?
Many of those examples aren't, any more, current CEOs of publicly traded American companies, largely because of the things you think make them bad CEOs. They are, therefore, ineligible for he competition.
That's okay, if he's not on the ballot come November I'll write in his name anyway.
In most (perhaps all) states, write-in votes in general for candidates who haven't filed papers as a "write-in candidate" are exactly equivalent to non-votes and are not counted. This is particularly true in the case of Presidential general elections, where you aren't actually voting for the candidate whose name is printed on the ballot (or written in) but for a slate of state-level Presidential Electors pledged to vote for the candidate.
So, even as a protest, this is particularly pointless.
In a couple of years time, there will be a proliferation of different, incompatible versions of CLang/LLVM that will be increasingly expensive to maintain.
Or they'll be a mainline version that is maintained by an active community of both independent and paid corporate developers contributing code, and corporations that want to maintain proprietary extensions will do so in a modular way, contributing any necessary core support back to the mainline project.
There's plenty of widely-used non-GPLv3 (public domain, BSD, Apache, GPLv2, etc.) free software that doesn't have the problem you suggest in practice. SQLite (public domain) comes to mind.
Taken at face value, this means if Google infringed even a single line of code (let alone 9), Oracle is entitled to ALL of Google's gross revenue unless Google proves that each dollar was either 1) Used for expenses or 2) Attributable to some factor other than the infringed work
And you probably thought statutory damages were the most unjust part of copyright law.... turns out actual damages are even worse.
Two things:
First, what you are actually pointing to as "worse" is the infringer's profits that are available in addition to actual damages, not the actual damages themselves.
Second -- and more importantly -- your reading of Section 504(b) conflicts with the 9th Circuit precedent that is binding on the trial court here. See, among other cases, Polar Bear Productions v. Timex, 384 F.3d 700 (2004):
Section 504(b) sets forth the evidentiary burdens for recovery of profits: "The copyright owner is entitled to recover... any profits of the infringer that are attributable to the infringement.... In establishing the infringer's profits, the copyright owner is required to present proof only of the infringer's gross revenue, and the infringer is required to prove his or her deductible expenses and the elements of profit attributable to factors other than the copyrighted work." 17 U.S.C. [Section] 504(b). Thus, [Section] 504(b) creates a two-step framework for recovery of indirect profits: 1) the copyright claimant must first show a causal nexus between the infringement and the gross revenue; and 2) once the causal nexus is shown, the infringer bears the burden of apportioning the profits that were not the result of infringement.
(BTW, Slashdot's failure to handle many HTML entities in comments -- like § -- is annoying.)
When the RIAA sues, typically they sue over a large number of different copyright protected works; with up to $150,000 in statutory damages available per work without proof of actual damages or infringers profits, they are able to rack up large statutory damage awards this way.
In the charge at issue here, Oracle has gotten a verdict on Google infringing a single work for which they have provided no evidence of actual damages or infringers profits. With up to $150,000 in statutory damages available per work, that gives them $150,000 in statutory damages available.
Its worth noting that the judge isn't informing Oracle that their damages are minimal. He has informed Oracle that they didn't bother presenting evidence on damages or infringers profits from the infringement of the work at issue, and since they didn't do that, there is no evidence in the case on which to find anything other than statutory damages.
The difference here is not a problem with the court system (I'm not saying that the court system does not, in many ways, favor the wealthier litigant, but the difference in the damages available in the two kinds of cases at issue in this subthread isn't actually that kind of issue.)
If there is a problem, its with the way copyright statutory damages work (either in being too generous in the kinds of cases the RIAA brings or being too stingy in the kind of case Oracle has brought.) But its not the the people targetted by the RIAA have succeeded less well in cases where the facts at issue were parallel to those in Oracle v. Google, its that the legal rules provide larger awards without proof of actual damages when lots of works are at issue than when fewer works are at issue.
Oracle can go after infringers profits, but in doing so it has to give up on statutory damages.
The Judge has pointed out that they haven't submitted any evidence supporting that Google has any profits associated with the rangeCheck method that is at issue, so this may not be a wise course of action.
The trick is in closing the feedback loop. Not all projects are software projects, where quality is highly subjective and unmeasurable.
Subjectivity of quality isn't the main problem (and, arguably, isn't the main problem) in the case of software.
The biggest problems are: 1. There's a pretty big mass of research showing that systematic reward mechanisms in work are usually demotivating rather than motivating, and a new rash of such studies usually shows up shortly after each new fad using reward mechanisms under a different name (which is what "gamification" is), and 2. The measures and premises on which gamification in work is based are almost always examples of micro-optimization which assume that processes are perfect and the problems are with individuals. This is an idea that has a natural attraction to the leaders who are responsible for the process, but in reality process is very often the area that actually needs the improvement.
You want to make things better, you get a process performance metric that measure real value delivered by the process -- and use them with statistical sophistication so you can understandwhether the process is in control or out of control -- and engage the team that is responsible for executing the process in the work of deciding how to improve the process to better deliver on the metric.
There are too many other things insurance pays for besides hitting another car.
What is being discussed as the end of is (though perhaps not clearly enough identified as) automobile liability insurance. This is, in practice, often bundled with other forms of insurance that also relate to automobiles, but usually only the liability part is mandated for operating on public roadways (and its usually, by far, the most expensive part), and the mandatory liability coverage is used as the wedge to sell the other coverages.
Sorry, that's incorrect. All those are examples of robots being used to deliver exactly the "speech" some human has decided should be delivered in some particular set of circumstances.
This would be true if "speech" meant simply "people speaking."
Its true, period. Corporations are a legal fiction that have no real existence. All so-called corporate speech is actually the speech of individual people, whether its "speech" by "speaking", or "speech" in the form of mass commercial advertisement.
Slashdot Mirror
It certainly does in the context of your earlier suggestion that the decompiled files support some inference related to why Google chose to do it. Since it was a contractor doing something Google expressly forbade them to do, any speculation about why Google chose to do it rests on a fatally false premise.
As has been pointed out multiple times in this subthread, most recently in GP -- to which the quoted material above is in reply -- no one is arguing against that point. So why do you keep raising it?
Misreading on several points.
Your post ws posted at 2:57pm on Wednesday, May 16; the article you link to was posted on (and transcribes court discussions that occurred on) Tuesday, May 15. So, "today" probably isn't the word you are looking for.
More importantly, you've confused Oracle's lawyer with Google's. I've broadened the excerpt that you provided to include the part that shows who is talking (material that was excluded from your excerpt is bolded):
Part of it is, part of it isn't.
Well, he certainly claims not to be a "Java expert"; whether he is a coder or not is, I don't think, a point on which there is much basis to make a conclusion, nor is it particularly relevant in any case.
No, its not (well, not in the copyright portion of the trial) because ideas aren't protected by copyright, and not even Oracle's lawyers -- as close to it as they come with the API SSO issue -- are going to try to outright make "they stole our idea" the basis for a claim for copyright liability, since even the most clueless federal judge imaginable would have to throw out that argument.
(Obviously, there claim is that specific ideas were stolen in the patent phase, but that's a whole different ball of worms. And, no, the mixed metaphor is not accidental.)
They also aren't making the argument you suggest since, aside from ideas not being so clearly outside the scope of copyright protection that not even BS&F would try that argument, the decompiled files at issue aren't part of the VM at all, but part of Sun's compatibility test suite.
The 9 lines of code, by the way, aren't even the same issue as the decompiled files -- the 9 lines of code are code that a Google employee wrote and contributed both to Android and OpenJDK, and have nothing to do with the 8 files decompiled by Noser while under contract to Google.
Assuming the obvious interpretation of the acronyms there, I must point out that "read" and "understood" are two different things.
I've never heard of the subtroutine concept! What's that, when a bigger fish calls a smaller fish to perform a function required by the bigger fish?
Bzzt. Wrong.
No, not at all. Josh Bloch worked for Google as Chief Java Architect (not for Noser) when he wrote rangeCheck (and still works for Google in the same role.) You seem to be confusing rangeCheck with the 8 decompiled files, which were done by Noser Engineering AG. Those are the two independent issues on which copyright infringement has been found in Oracle v. Google.
It does both more (checks more numbers) and less (doesn't deal with actually passing the numbers anywhere) than that: it actually checks two numbers against an array, and throws an exception if either of the numbers is a bad number. Passing the numbers somewhere else is the responsibility of the calling code.
rangeCheck isn't even part of that body of code, which consists of 8 test files that were never distributed on handsets, but which part of the Android source distribution.
What parts of Android the contractor was involved in is, I believe, part of the trial record; it is, at any rate, information Oracle could have subpoenaed and addressed had it any interest in doing so.
They didn't do anything with this function (rangeCheck) because they had nothing to do with it. It wasn't part of the decompiled files, the author of the function was a Google employee (who wrote it outside of the scope of his employment) who both included the file in Android and contributed it to OpenJDK.
It actually has Chrome, Firefox, and Opera tied with 8.5 out of 10, with IE (8.0) and Safari (7.8) trailing.
Actually, its not the "the same guy wrote it for Oracle and later, Google". It wasn't written twice.
Its that the Google employee who wrote it on his own time both included it in the Android source tree and contributed to to the OpenJDK (including the required copyright assignment.)
Actually, Oracle's pretty much given up on that, too. In a joint stipulation filed today, Oracle has agreed to waive both jury trial and its claim for actual damages and infringers profits on the copyright claims for which liability has already been found if the API SSO issue isn't resolved in its favor, and accept statutory damages for rangeCheck and the decompiled files.
No one is arguing against that point, so I don't know why you keep arguing for it. My objection was your use of reference to the decompiled files and Google's supposed decision to copy them (which is refuted by the fact that the decompilation was done by a contractor under express direction from Google not to decompile any third-party files) as part of the backing for your post asking why, if it was so simple, did Google choose to copy?
rangeCheck was not from one of the decompiled files, its a different issue.
It doesn't matter to copyright liability (well, actually, it does, but it doesn't matter to Google's copyright liability.)
It does matter, however, when someone uses quotes about the decompiled file to backup an argument of the form "if its all so easy, why did Google copy it?"
Since, in fact, Google didn't copy it, and in fact the person who did copy it was doing something Google had expressly directed them not to do.
Some of it, presumably, is content development for stuff that gets distributed as not-ads on the facebook page. If the "paid ads" is only the amount they are paying for the advertising space, some of the $30million may be content development for the ads themselves.
I would expect that in addition to paying people to do that via the Facebook user accounts or their company Facebook page, they are also paying people to write and edit/review approved "status updates", design and manage a coherent overall essaging strategy, take and edit professional photographs for distribution via Facebook, etc. And that many of those people cost GM more than $50k/yr (workers cost employers a lot more than the workers gross salary.)
Because it goes directly to the claims that it infringing it was the source of a substantial quantity of Google's profits.
The decompiled files are a different issue than rangeCheck. Google didn't decompile them, Noser -- a contractor for Google -- did, against the express prohibition in the Google-Noser contract, which required all original work or open-source code, and expressly prohibited copying or decompiling proprietary code.
I tried to submit this over the weekend but it wasn't selected. http://www.wired.com/wiredenterprise/2012/05/google-oracle-decompile/
"Judge William Alsup ruled that evidence presented during the trial had shown that Google infringed on Oracleâ(TM)s copyrights by decompiling eight Java files and copying them in their entirety for use with Android."
Many of those examples aren't, any more, current CEOs of publicly traded American companies, largely because of the things you think make them bad CEOs. They are, therefore, ineligible for he competition.
In most (perhaps all) states, write-in votes in general for candidates who haven't filed papers as a "write-in candidate" are exactly equivalent to non-votes and are not counted. This is particularly true in the case of Presidential general elections, where you aren't actually voting for the candidate whose name is printed on the ballot (or written in) but for a slate of state-level Presidential Electors pledged to vote for the candidate.
So, even as a protest, this is particularly pointless.
Or they'll be a mainline version that is maintained by an active community of both independent and paid corporate developers contributing code, and corporations that want to maintain proprietary extensions will do so in a modular way, contributing any necessary core support back to the mainline project.
There's plenty of widely-used non-GPLv3 (public domain, BSD, Apache, GPLv2, etc.) free software that doesn't have the problem you suggest in practice. SQLite (public domain) comes to mind.
Two things:
First, what you are actually pointing to as "worse" is the infringer's profits that are available in addition to actual damages, not the actual damages themselves.
Second -- and more importantly -- your reading of Section 504(b) conflicts with the 9th Circuit precedent that is binding on the trial court here. See, among other cases, Polar Bear Productions v. Timex, 384 F.3d 700 (2004):
(BTW, Slashdot's failure to handle many HTML entities in comments -- like § -- is annoying.)
When the RIAA sues, typically they sue over a large number of different copyright protected works; with up to $150,000 in statutory damages available per work without proof of actual damages or infringers profits, they are able to rack up large statutory damage awards this way.
In the charge at issue here, Oracle has gotten a verdict on Google infringing a single work for which they have provided no evidence of actual damages or infringers profits. With up to $150,000 in statutory damages available per work, that gives them $150,000 in statutory damages available.
Its worth noting that the judge isn't informing Oracle that their damages are minimal. He has informed Oracle that they didn't bother presenting evidence on damages or infringers profits from the infringement of the work at issue, and since they didn't do that, there is no evidence in the case on which to find anything other than statutory damages.
The difference here is not a problem with the court system (I'm not saying that the court system does not, in many ways, favor the wealthier litigant, but the difference in the damages available in the two kinds of cases at issue in this subthread isn't actually that kind of issue.)
If there is a problem, its with the way copyright statutory damages work (either in being too generous in the kinds of cases the RIAA brings or being too stingy in the kind of case Oracle has brought.) But its not the the people targetted by the RIAA have succeeded less well in cases where the facts at issue were parallel to those in Oracle v. Google, its that the legal rules provide larger awards without proof of actual damages when lots of works are at issue than when fewer works are at issue.
Oracle can go after infringers profits, but in doing so it has to give up on statutory damages.
The Judge has pointed out that they haven't submitted any evidence supporting that Google has any profits associated with the rangeCheck method that is at issue, so this may not be a wise course of action.
Subjectivity of quality isn't the main problem (and, arguably, isn't the main problem) in the case of software.
The biggest problems are:
1. There's a pretty big mass of research showing that systematic reward mechanisms in work are usually demotivating rather than motivating, and a new rash of such studies usually shows up shortly after each new fad using reward mechanisms under a different name (which is what "gamification" is), and
2. The measures and premises on which gamification in work is based are almost always examples of micro-optimization which assume that processes are perfect and the problems are with individuals. This is an idea that has a natural attraction to the leaders who are responsible for the process, but in reality process is very often the area that actually needs the improvement.
You want to make things better, you get a process performance metric that measure real value delivered by the process -- and use them with statistical sophistication so you can understandwhether the process is in control or out of control -- and engage the team that is responsible for executing the process in the work of deciding how to improve the process to better deliver on the metric.
What is being discussed as the end of is (though perhaps not clearly enough identified as) automobile liability insurance. This is, in practice, often bundled with other forms of insurance that also relate to automobiles, but usually only the liability part is mandated for operating on public roadways (and its usually, by far, the most expensive part), and the mandatory liability coverage is used as the wedge to sell the other coverages.
Sorry, that's incorrect. All those are examples of robots being used to deliver exactly the "speech" some human has decided should be delivered in some particular set of circumstances.
Its true, period. Corporations are a legal fiction that have no real existence. All so-called corporate speech is actually the speech of individual people, whether its "speech" by "speaking", or "speech" in the form of mass commercial advertisement.
No, though they may sometimes be tools used by people in the making of speech.