Minions Finally Lose
Did Hell freeze over? I thought they would rule for the Minions and Leisure Suit Larry would get another island.
The US Supreme Court on Monday ruled in a 6-2 decision that Google's limited copying of Oracle's Java APIs in its Android operating system constitutes fair use under US law. The ruling puts an end to a case that troubled the software industry for more than a decade and narrows the scope of copyright law as it applies to …
What was decided is that since the code was different you could copy the looks. But clone Apis share some code like reference names for compatibility reasons.
That was why it took so freaking long, do some similarities in code for the sake of compatibility count as a breach of copyright or not?
Apparently the answer is no cause is fair use.
Well this isn't the first time the US supreme court got it wrong. Google blatantly copied Java, there really isn't any argument there. And they used it outside the permitted use according to the license. I know most people didn't want Oracle to win, but really they should have.
"Google blatantly copied Java,"
No. Google blatantly copied a list of Java function names. They didn't copy the JVM. They didn't copy the implementations of those functions. They also skipped some of the functions available. The question is whether those names are as protected as all the important code that actually runs is. The answer appears to be no.
You aren't as wrong as people seem to think, but you aren't right, either.
US copyright law says both that computer programs cannot constitute works subject to copyright, and that, as a matter of direct legislative fiat, they are copyrightable.
It really is a mess. The SC sidestepped that issue here, I suspect because it would otherwise have had to rule definitively (although in passing) that no copyright interest can exist in open-source software, which is not what we'd like to see.
While I'm not absolutely convinced by the majority opinion, it's far more persuasive than the dissent, not least because the majority demonstrates clear understanding of the programming technicalities, while the dissent displays misconceptions.
While I'm not absolutely convinced by the majority opinion, it's far more persuasive than the dissent, not least because the majority demonstrates clear understanding of the programming technicalities, while the dissent displays misconceptions.
The dissent is Alito and Thomas. What did you expect?
To add to the above now I'm better caffeinated, it's unfair and entirely unjustified to accuse SCJs of being intellectually dishonest. However much I may disagree with a given opinion, I wouldn't suggest for a moment that it is sophistry contrived to support a predetermined outcome.
If you read the dissent, it isn't unpersuasive. I disagree about the nature of 'declaring code', as does the majority opinion, but it's unquestionable that if one accepts the opposing view in that one regard, the dissent is a correct presentation of the logical consequences.
At first, you say:
However much I may disagree with a given opinion, I wouldn't suggest for a moment that it is sophistry contrived to support a predetermined outcome.
Then, roughly 20 words later...
[...] but it's unquestionable that if one accepts the opposing view in that one regard, the dissent is a correct presentation of the logical consequences.
So, the dissent makes sense if you're driving to a predetermined outcome. Nice.
You sure that was only caffeine you imbibed?
No, the dissent makes complete sense with one different premise. I'm astonished you're struggling to understand that.
There is certainly no politics involved in the two interpretations. Arguably a technical misconception makes the premise wrong, but not indisputably so.
11,500 lines of code which, to put it in perspective for a non-techie, at 30 lines a page* means 380-odd pages.
Not War and Peace but not what I would call selective, incidental use either.
But out of how many lines total?**
*Yes I know some lines are very short but each has a purpose. yes?
** Oh. 2.8 million. OK. Scrub the above remark.
"11,500 lines of code which, to put it in perspective for a non-techie, at 30 lines a page* means 380-odd pages."
60 lines per page would be a more reasonable number. The real question as to whether 190 pages is a significant number is when you take that code as a proportion of the entire "work".
A million grains of sound may seem like a lot, but if you take that many grains from a beach you'll not be able to discern the difference.
> Years ago (1990s) didn't Lotus sue M$ over the latter's copying of the 123 menus? And wasn't it decided then that you could do this? Weird how the same thing can come back and earn lawyers millions more.
It was Lotus vs Borland, and the written decision in this case cites that case several times. And, interestingly, Breyer, the writer of this decision, was one of the Justices on that case as well.
It's not that odd. The case was an attempt at a cash grab by Oracle, so it was simply a relatively high-risk investment. It didn't pay off, but now that it's over, Oracle's in a more stable position: It won't be spending more on this investment, and there's no possible exposure from setting a potentially damaging precedent. To the extent that anything which happens in the stock market is at all rational, this likely improves the value of Oracle a bit.
Oracle, in a statement emailed to The Register, decried the ruling.
One bully decries that the other is a bully. Then decries that the one bully is right and the other bully is wrong.
Well, at least we don't yet see a copyright tax on APIs just yet, but the final words may not have been said. Unfortunately, we still may see a lot of bullying from the parties to come. I would have liked to see a ruling where both bullies are prevented from any further bullying in the future.
While I agree with the sentiment, a charge of barratry isn't in the cards, I don't believe any of these lawyers are commissioned officers. The only reasonably well known case of barratry in history was the "mutiny" on the Bounty, led by the first officer and that case resulted in some changes in law regarding what a captain could and couldn't do plus some regulations under which a captain can be forcibly replaced without breaking the law.
How many millions did Oracle
waste spend on this stupid lawsuit?
Oracle destroyed a nice company - Sun Microsystems - in their deluded pursuit of greed. They only reason they bought Sun was for the Java lawsuit. They never cared about anything else.
Yes, Sun had its wad of problems - major ones - but at least Sun consistently put a value on innovation and new ideas. And it could have been fixed and brought back to health.
When was the last time Oracle produced something original - besides lawsuits? 20+ years ago?
Under normal circumstances, Oracle's Board should now fire Larry Ellison, Safra Catz, Dorian Daley and Edward Screven. The architects of this toxic, idiotic lawsuit.
But we all know they won't.
> Oracle just bought up what was left after Jonathan Schwartz ran it into the ground.
Have you ever worked at Sun?
If yes, you are entitled to your opinions about Sun and Jonathan Schwartz.
If no, STFU. You have no clue what you're talking about.
Speaking as someone who worked at Sun for quite a while and survived the Oracle acquisition.
Just because what Suns managment did or didn't do wasn't great . . . - that doesn't make what Oracle do good.
The two things aren't connected.
Sun was a bunch of genius techies with little business sense.
Oracle was, and is, a corporate behemoth intent on extracting all your licence obligations and if you drop the ball they will turn up and audit and hit you with a £500k bill for your non-prods and devs you didn't know needed a license. That wasn't in your budget and you are fired!!
Oh and last time I looked Oracle don't provide you with an automated way to deal with your Oracle licensing. They just let you screw up and come along for the money later.
Things may have changed in the 7 years since I last had to deal with the c<nts.
The thing that really grated at the time was that alot of the companies exposure to the audit was being overlicensed in Prod environments. Paying for Enterprise when Standard fitted the bill, etc. Given that Oracle where embedded in the projects that implemented those systems and never called out the wrong licenses and then turned up 5 years later to claim their $$$$$$$s . . I think I will never buy anything off them or deal with them if its my decision to take.
I didn't work there for long, but it was a mess when I left 3 years before M Schwartz took over. It would be easy to say hubris, incompetent management, self aggrandizement, or delusion was their downfall, but it was all of them, and probably a few more.
> [...] [Sun] was a mess when I left 3 years before M Schwartz took over.
The root of Sun's demise was McNealy's terrible, no-good decision in 2002 to abandon Solaris on x86. Right when the rest of the world was moving to x86 and Linux.
That was the year - 2002 - when I got a ThinkPad (IBM was still making them back then) for $2500 and installed RedHat Linux 7.3 on it (RedHat's old version numbering scheme). I had been running Linux for a while back then on PC's, but not on a laptop.
The ThinkPad had a dedicated ATI video card with a whopping 32MB RAM. State of the art for laptops, back then.
The $2500 ThinkPad with a whopping 8GB RAM and a 2nd generation 32-bit Pentium 4 (1.2GHz) running Linux, was, on average, 3 times faster than the $30,000 Ultra 60 workstation with dual UltraSPARC-II CPU's and 32GB RAM I had in my office.
The Ultra 60 also had some super-duper Sun video card that didn't support XRender. The OpenGL doughnut demo was 4 times faster on the ThinkPad than on the Ultra 60 - counting the number of FPS.
I have nothing against the SPARCv9 ISA. It's a very elegant ISA. But Sun didn't deliver on price, or performance, and they kept falling behind.
Nothing to do with Jonathan Schwartz. If anything, Jonathan understood the consequences of McNealy's catastrophic mistake, and tried to do something about it - namely Solaris 10 on x86_64.
I worked at Sun from the early 90s and well through the Oracle acquisition, and the OP isn't wrong. McNealy's flip-flopping over Solaris/x86 (and the same over Sun Linux) was short-sighted, but also short-lived. When our then-VP ordered my team to stop the work on the Solaris/x86 version of our product we ignored him, correctly assuming it would soon be back (which was more than could be said for him).
What really hurt Sun was McNealy's reluctance to swing the big axe when the bubble burst (and I say that knowing my non-HQ office would probably have been closed if he had done). He tried to minimize the immediate damage, but instead crippled Sun so that it didn't have the funds to survive in the long run, and condeming us to the annual "death by a thousand cuts" layoff model that led eventually to Oracle.
McNealy's succession plans were crimped when Ed Zander left, leaving the woefully inadequate Schwartz to flounder, well out of his depth. His "if you build it, they will come" approach to opensourcing everything was never realistic. Computer science departments training the next generation were already well involved with Linux, if Schwartz wanted to change that he should have been wooing them with free systems & active help, instead of just scattering a few community-based "campus ambassadors" around on the assumption that people would flock to them.
Pushing the "Solaris is better than Linux" approach was also doomed to fail, just as that style of campaign had failed decades earlier when Ken Olsen described Linux as "snake oil", only to watch Sun eat DEC's breakfast. Solaris was better for very large systems and workloads (still is, which is why Oracle hasn't been able to kill it despite the best efforts of Larry's team), but that made it slower on more lightweight hardware. Schwartz would have been better to embrace Linux as a "little brother" to Solaris, offering an evolution path, instead of simultaneously demonizing it while trying to copy it.
McNealy's mistakes may have started the slippery slope, but Schwartz kept it well-greased.
I had 5 years at Sun and loved it. The company went from 24,000 employees to 42,000 employees in that time. When the dotcom bubble burst and the revenue shrank Sun should have shrunk the numbers down to match. I left and watched my colleagues spend the next 10+ years constantly under threat of redundancy. That's not very motivational.
SUN's were pure magic when they came in the late 80's. Remember doing graduation work on SUN 2/50 and 3/50 with MC68010 and MC68020 cpu's running 4.1BSD. First time I worked on real graphics workstations. What a difference was this compared to Unix V7 on a pdp-11 and BSD 4.2 on a VAX 11/750.
SUN contributed so much to the Unix world, things like NFS and RPC were invented by SUN, others like IBM just used these things and never gave anything in return.
Maybe SUN made two mistakes, hanging on to the anemic SPARC architecture which didn't have enough scale to enable it to compete with others and the adoption of Solaris, which was actually AT&T Unix System V Release 4.0. This was a overly complex OS which makes systemd seem like the best invention ever. I also found it weird Motorola didn't invest in keeping its 68000 CPU's relevant, they were the only kid on the block in the late 80's.
Always thought SUN management gave Oracle permission to use Java: https://www.cnet.com/news/former-sun-ceo-says-googles-android-didnt-need-license-for-java-apis/
hanging on to the anemic SPARC architecture
A modern 5GHz SPARC M8 still runs rings around anything you can do with an Intel chip. I have access to a 1024-core SPARC M8 system with 2TB of RAM, it is far from anæmic, especially when doing crypto with the on-chip accelerators.
> ... Motorola didn't invest in keeping its 68000 CPU's relevant
ISTR Sun approached Motorola to collaborate on a new architecture which would have superseded 68000, but were summarily rejected ... so went it alone with SPARC instead.
And yes, at the time, SPARC did indeed run rings round contemporary Intel offerings.
Motorola got on the RISC bandwagon with the 88000.
What I remember - and I'm sure this is inaccurate in one or more (maybe all!) ways - was that there was a Motorola/Intel/Apollo VLIW effort, which led separately to Motorola 88K, Apollo 88K (DN10000), and Intel i860.
This post has been deleted by its author
Yes, there are a lot of new projects started using Java these days. In the enterprise space, Java won the battle over .NET years ago, out of openness and availability and in spite of the incredibly complex (and mostly unnecessary) J2EE architecture.
Enough useful systems have been built with Java and .NET to survive and transition to the flatish "useful" part of the hype curve. Yesterday's top of the curve (Ruby/Rails) has not been so lucky I'm afraid and has been replaced by Node.js which I think will neither survive to reach the "useful" part of the curve and will be replaced by Python or something else that is not promoted via the usual hype cycle.
"Does anyone use Java for new projects these days?"
I think this is an important point. The big loser in this case is Java. Since the case has been going on, the uncertainty has meant that other languages such as kotlin and swift have come about and diluted the space Java once held.
without this case Java could of become the general purpose language the Sun envisaged, but a combination of legal doubts and Oracle's contract power grabs has meant Java is far less attractive than it once was
That seems to be occupying the space that Java was originally intended for - interactive applications delivered over the web, and via Electron, multi-platform desktop applications.
It's not entirely Java's fault mind you. The hardware running Electron apps is mindbogglingly powerful compared to what was available 20 years ago. The dreadful and jarring look and feel of Java applications though.. that definitely didn't help.
A someone who was tangentally connected to mobile Java while at Sun, I'd say this ruling fell on the side of justice. Sun fumbled with mobile Java and handset SW in general (remember Savaje?) and Oracle decided it wasn't worth their time and money to pursue. When Android came out Larry and his swarm of lawyers saw blood in the water, and deep pockets. Oracle's nothing more than a patent troll and customer shakedown outfit these days. R.I.P Sun.
If Oracle had won it would be a body-blow to the whole concept of open systems and API commonality.
The reason is of course that even if Google had coded their own version of the Java API headers, the function signatures & object declarations would necessarily be the same as those defined by the Java libraries.
According to the article, one major factor the Supreme Court judges considered was the size of the part borrowed under fair use, in relation to the whole. In this case, it was tiny.
Taking a whole program, making a few changes, and then declaring it proprietary would be a completely different case.
@MacroRodent: "... the part borrowed under fair use, in relation to the whole. In this case, it was tiny. Taking a whole program, making a few changes, and then declaring it proprietary would be a completely different case."
Yes, and the case will be taken sometime soon. The question now is "How much is too much?"
If only we had a large corpus of jurisprudence on what constitutes "fair use" in the US... oh, would you look at that.
I think Breyer's opinion is pretty clear. This decision clarifies, not clouds, the status of open-source software in the US. It's not an ideal decision – the court ought to have found APIs not subject to copyright, as Alsup did, and instead left that question open – but it acknowledges the issues with copyright law for software, and draws a pretty clear line regarding fair use.
The whole point of an API is that its a *public* facing interface to a 'black box' code subsystem. Of necessity sometimes that API is going to be expressed as source code but this is more as a form of documentation than active code.
To me this entire lawsuit was moot from Day One. I'd guess that just about everyone on this site knows exactly what an API is, its not something that needs to be debated by lawyers and give some kind of spin by a sophisticated public relations campaign. Oracle tried to pull the wool over the non-technical community's eyes in the search for a monster payoff without any thought for the consequences for the industry as a whole should they have got away with it. This is, IMHO, corrupt behavior -- thoughtless, destructive and self-centered, telling us a great deal about the company.
A good day for developers.
Microsoft launched MS-DOS by copying the CP/M interface - I'm not going to complain because it help me write a bunch of MS-DOS programs early on and Gary Kildall seemed to have no concern that they were copying his work, but that was decades ago and in those days we were all running around getting things working - with very little thought about lawsuits.
Actually MS simply bought it from a company that ripped off CP/M 86. It was far more than the interface. CP/M 86 was easy to produce from the 8080 CP/M (later called CP/M 80) because the 8088 / 8086 wasn't a real 16 bit CPU. It used 64 K blocks and segment registers. Very similar to 8080 / 8085, so Intel had an assembler translator. Very little work to port say Supercalc or Wordstar from CP/M 80 to PC-DOS/MS-DOS.
They did base MS Basic on Dartmouth BASIC.
So MS original fortune based on two products, one a simple port and one a bought in knock-off.
Actually the Java issue was caused by the fact that full fat Desktop Java was essentially free but not allowed to be used AT ALL for Mobile. Symbian used a cut down licensed Mobile Java. Android was bought in by Google and Sun or Oracle refused to license the full version Java. So they had to write their own, and use Davik instead of the cross platform PC full JVM.
So Oracle has only themselves to blame.
CP/M 86 was designed as a feeder product for Concurrent CP/M and was a re-write of CP/M. Seattle Computer Products DOS was a simple program loader + shell with just enough OS for a demo. IO.SYS and MSDOS.SYS were tiny. Microsoft got the OS gig from IBM because they pitched Xenix (Unix on a Micro). MS BASIC and BASICA (Basic Advanced) used the syntax on basic (including READ for in-line data cards).
No, it was a DOS that ran Wordstar and supercalc that attracted IBM. They wanted minimalist. The original PC had no graphics, no realtime clock, no HDD (Xenix would have needed an HDD) and the Floppy was even then not only priced separately but only 360K. The 8" drives were 1M then and other companies had higher capacity 5.25" floppies.
I could be wrong, but Xenix was brain dead on an 8088, though it existed in 1980. It needed a 286 to be at all viable. Minimum was the later XT (a HDD)
IBM didn't want competition with "real computers" hence choosing the 8088 and calatogue parts. IBM had a choice of real 16 bit CPUS and real OSes and deliberately chose MS-DOS. IBM did buy Xenix from MS for other products.
Nope , the 5150 (the IBM PC model number) was the way it was because of the spectacular failure of the 5100. IBM's first "PC" which was completely closed system and proprietary. I still have a manual for it around somewhere.
Don Estridge was given an impossible schedule for the new model, which is why it used an existing 8088 motherboard from another product (not MC68000 as planned), and the decision to use third party OS's and bundled software was mostly driven by this schedule and Estridge belief that the success of the S/100 systems and Apple II could be replicated by being as non-proprietary as possible. The failure of the DEC Rainbow the following year proving that Estridge was correct. The DEC Rainbow was a typical piece of beautiful hardware by Digital but was closed system enough to kill it in the market. You could not even format floppies initially. You had to buy expensive per-formmated ones from DEC..
The huge success of Compaq and Victor that year showed how it could be done. Non proprietary and more bang for the buck than IBM
The 5100 series wasn't a complete failure. It was the only portable APL machine around (as far as I know), and so had some fans in finance and some other sectors. I knew one or two people back in the day who had worked on 5100-series machines (employer-supplied, of course; in today's money those things would be around $50K - $100K).
You know, I have vague memories that something else at the time could also run APL. I remember some weird ZX81 like micro that had an APL EPROM. Not APL/360 but a micro version. Of course not a lot you could do in 4K. There was also the Exidy Sorcerer APL ROM pack which never saw the light of day. The first actual APL I saw running in the wild was on a IBM PC/XT and there was a version kicking around on the Mac by 1985. To me APL was always the ultimate write-only language. It made LISP code look self-documenting. Just like in LISP you could write code that did in a few lines what would take pages in other languages. But just like LISP try remembering how it worked three month later....
SCC who were selling 8086 cards got tired of waiting for Digital Research to ship a stable version of CM/86 so wrote their own. No ripping off involved. It was just a monitor really. Not a lot of code.
MS heard about it as SCC were a local company across the lake and through some very sneaky lying got SCC to sign off on an opened ended license. Which MS turned around and relicenced to IBM. Years later SCC successfully sued MS for contract fraud and there was a 8 figure settlement (+ legal fees) wrapped up in a typical MS legal settlement NDA. MS have signed off on several hundred of those over the decades. MS's legal strategy has always been to try to bankrupt the other party through legal costs (or make them larger than any likely settlement) because otherwise MS always will lose the case. Always.
MS even tried that with the state of Washington when they defrauded the state out of billions in sales tax. Even though MS had successfully got a special rate they still defrauded the state. After more than a decade of litigation MS eventually settled by writing check for over a billion. Maybe 30% of the total fraud.
As for the port to CP/86 dont know about SuperCalc but the Wordstar code should have been a simple port. It was all in assembler but the system monitor calls were all pretty isolated. That codebase ran on a lot of different S/100 etc machine at the time.
And the story of Java on mobile is the story of the failure of J2ME. Which was many different versions none not quite compatible with the others. So Nokia S40/S60 had the biggest market share of all the versions but still maybe only 30% max. SymbianOS was a C++ weirdo but you could get yet another version of J2ME that ran on top. Thats why there was no real J2ME app ecosystem. Totally balkanized versions of the platforms.
The legality of API's usage has been clear for almost 50 years. Going back to the days of plug compatible IBM 370 clones. Oracle were the first big company in decades to try to overturn this vital legal protection. And all because the warped diseased mind of Larry was convinced that if Java was not locked down propriety one of the most lucrative Oracle products at the time might not generate such eye watering high margins. It was this story that convinced me that Larry was just milking Oracle dry for hundreds of millions every year before he died. I would not be very surprised if Oracle collapsed financially, an empty shell, once Larry is dead.
Would not be the first time. Jack Tramiel did pretty much the same with Commodore in the late 1970's/ early 1980's. No knowing how much was siphoned off at the time through the Bermuda office. What was left lingered for almost another decade but it was just a zombie company by that stage.
Back when Microsoft started, it wasn't even clear yet if software was covered by copyright, never mind APIs.
Gates famously published an open letter to computer enthusiasts, pleading them to not copy his BASIC, https://en.wikipedia.org/wiki/Open_Letter_to_Hobbyists
I'll take MS and raise you Oracle.
Oracle and Larry have made billions implementing SQL, created and developed by IBM.
If Oracle had won this it would of been the ultimate hypocrisy that the company which literally came about by copying a language API's restricted others from doing similar
Based on his re-tweeting: No.
Alongside others of "I created something (possibly trivial) and my descendants should live off that IP for eternity" brigade.
The ruling certainly opens up more potential for "fair use" claims, but I suspect far less than the some of the hyperbole currently going around would suggest.
And in many cases they've gone against the expectations of those who nominated and supported them. Souter, for example, was appointed by Bush I, but is generally considered to have voted "liberal" in most of the decisions while he was on the court. (This Wikipedia article, which should of course be taken with a grain of salt, shows Souter as ranging from 54% to 71% liberal in various categories.) Kennedy, appointed by Reagan, was often the swing vote and in his day generally considered the hardest to handicap. Scalia, definitely right-wing (but famously a very skilled thinker, jurist, and writer, and friend of RBG), skewed left on federalism and taxes, according to that same Wikipedia page.
SCOTUS judges do what SCOTUS judges want to do. They're pretty tough to influence, except in the approved manner (i.e. well-founded legal arguments grounded in the legislation and jurisprudence of the US).
I never thought that Google would win this before the Supreme Court. Too many of the Justices make me look young and none are known for being particularly computer savvy. I applaud their clerks who were able to explain things to them. I applaud Stephen Breyer for his opinion which could have come from one of us (albeit less elegantly) despite the fact that he was born before electronic computers existed and graduated from Harvard Law before UNIX was developed.
I don't usually offer pints to lawyers but I will today.
I'd be interested to know what the dissenting opinions were. Clarence Thomas, despite being pretty well the most right wing and useless SC judge on the majority of issues, only saving grace is that he is is supposed to be the court expert on copyright, and is the one issue where the others defer.
Ok, looked at dissenting opinion from Thomas (https://www.supremecourt.gov/opinions/20pdf/18-956_d18f.pdf)
Any statement that starts with
"Oracle spent years developing a programming library that successfully attracted software developers, thus enhancing the value of Oracle’s products."
Is obviously a rewrite of history (Sun developed it, encouraged use. Oracle just bought the company). The rest is just Thomas decrying that the verdict will harm Oracles bottom line, but misses the bigger picture and fundamentals of the case.
In short a typical Thomas opinion based on narrow minded bias, rather than looking at the overall picture
Mike Masnick has a good review of Thomas's dissent on Techdirt. The dissent is rubbish. Thomas doesn't know what he's talking about. He fails to understand what an API is, for example.
Of course that's not surprising, since Thomas is pretty suspicious of anything that happened after the Constitution was ratified. He certainly doesn't think much of the Bill of Rights.
Breyer's the best of the current panel on IP, I think, and this decision confirms it.
"Too many of the Justices make me look young "
Which means that they've had a lot more time than you not only working with the law but also seeing a huge variety of issues pass through the courts. So many people think law courts are remote. Spend time there and you discover just the opposite.
Justice Thomas is well aware of what the law is and is required to apply it even if he does not want to. He can look for ways to creatively interpret how section 230 is written - if a case reaches SCOTUS. He can beg congress to write a new law. He can join the dog-pile on spreading the popular misconceptions about 230. The one thing he - and congress - cannot do is take away the constitutionally granted right of freedom of association.
If Twitter does not want Trump soiling Twitter's platform then Twitter has the right to remove his shit from their website and to stop him from squatting there in future. Twitter was free to associate with Trump when it suited their profits even though he repeatedly broke their terms of service. Twitter was free to stop being associated with Trump for any reason or no reason at all and that is a constitutional right and nothing to do with section 230.
If Trump wants a platform there are plenty of delusional conspiracy theorist web sites that have already invited him. Trump can set up his own web site or he can ask Robert Mercer to fund one for him.
Thanks to section 230 racists can spew their hate on other people's web sites and the site's owner can avoid legal responsibility because responsibility lies with the person who submitted the post. Without section 230 Trump, his supporters and probably everyone else would only be able to put comments on their own websites.
The problem is that facebook, twitter et al are becoming defacto carriers. People turn to them in preference to making a phone call a lot of the time.
Phone companies aren't allowed to kick people off their networks for saying things they disagree with, or even for things like actual incitement to murder.
Facebook, google et al are trying to have it both ways. They want to be the primary means by which people interact with one another - making them in effect a carrier utility - but they also want to curate and control how (or even whether) those people interact with one another, for their own ends.
All I did was describe a problem. I didn't suggest "breaking" anything, or even what I would change to fix that problem. The acrimony about section 230 has emerged from the fact that these private companies are close to monopolising communication between people for most purposes. Getting rid of it isn't the solution, but so much of the discussion has boiled down to "muh private companies!", and all parties deliberately reading malice into everything that disagrees with them, whilst ignoring the real issues at stake, that getting rid of it appears to be the solution.
Can someone who is more versed in US law explain why Oracle continuing to harp on with their statements doesn't constitute libel or slander now?
I mean, to me it seems the highest US court has said "no, Google did NOT break the law and what they did was legal under fair use". If Oracle continues to gripe and make statements that in essence say "Google broke the law because WE say so"... isn't that illegal? They're not saying "in our opinion the SCOTUS got it wrong", if I read this right? Meaning that would be free speech.
IANAL, but I'm pretty sure this wouldn't be contempt of court (in the US). It's just expressing an opinion, which is strongly protected by the First Amendment.
I can proclaim on the street-corner, or publish an editorial in the New York Times,1 excoriating various SCOTUS decisions on questions of the Fourth Amendment – an area where they have Not Done Well. That's not libel (or slander) or contempt. I can, in fact, profess literal contempt for the court;2 that's not "contempt of court" in the legal sense either.
Defying a court decision is contempt. Disrupting proceedings is contempt. Disagreeing isn't.
1Oh, you know they'd love to have it.
2Not that I'm actually contemptuous of the court. But if I were, I could. Well, I suppose I could even though I'm not, if you see what I mean.
The bar for libel and slander in the US is appropriately high, not absurdly low as it is in some other jurisdictions.
Also, of course, there's nothing libelous or slanderous in claiming that a court arrived at the wrong decision. That's a clear matter of opinion, and under US law opinion is never libel or slander.
If Oracle published a statement, that, say, accused Breyer of taking a bribe from Oracle, that would be potentially libelous (though it's such an absurd hypothetical that it's hard to say whether anyone who matters would take it seriously).
Personally, I don't consider that what I can see in the API specs is "code". Which is the total sum that was used by Google. They reimplemented the code from scratch in order to match the API, same as the Wine project reimplemented Windows APIs.
In the first place, they'd have taken a license from Sun if Sun had been willing. But Sun refused because they wanted the world to use their crap Java for Mobile API instead...
>And any time you stick in include at the top of a program file you do they same thing
No you don't. Google copied and pasted actual source files from the JDK source tree to the Android/Dalvik source tree, stripping out the copyright headers. This is not the same as a fork or an inclusion or a dependency, in either intent or mode of operation.
If Google had depended upon the existing interface files - which they easily could have done - they likely would have been subject to Oracle licensing whims. The cloning of the JDK was done expressly to avoid Oracle's licensing. The manual copying of the interface definition source files was done to facilitate that. These are the accepted facts of the case and are not in dispute, much as that may trigger inexplicable downvotes.
... are no such thing, they are just indicators of disagreement. You find them inexplicable because you find even the fact of people holding a differing view to be inexplicable, which may explain why some of those downvoters have decided not to bother attempting to advance a counterargument. HTH.
This isn't an area where disagreement should exist. That Google chose to copy + paste literal code files rather than implement-by-spec is an accepted matter of fact by all parties. It is also relevant to the case, because one of the tests for fair use is the "purpose" and the "character" of the work. The fact that Google chose to copy the literal code rather than write it themselves was of material interest to the case; in the opinion of Oracle it further demonstrated that the key purpose of the work to be a byte-for-byte replica of the existing JDK as much as possible rather than deliver something new or innovative in nature.
The area of disagreement is whether this replication was permitted as fair use, not that the replication happened or how it was done. The byte-for-byte replica argument put forward by Oracle would lead one to conclude that the work was derivative rather than transformative; up until this ruling they'd have been right.
"Google did not implement the code in question based on the published API specifications, they copied the 11k LOC directly."
The lines concerned were the API specifications and didn't do anything because they just listed functions. They required separate implementations, which Google wrote without copying. If you compiled those lines of code and tried to do anything with them, you couldn't. Not a single one of them ran. This was just a list of names. You are implying otherwise, but we and you know what they were.
I don't think that's what I'm doing. The court case has been decided on fair use grounds. In the way I like, but using arguments that aren't the ones I would use. That's not at issue. I'm merely disagreeing with the quote "Google did not implement the code in question based on the published API specifications, they copied the 11k LOC directly.". It's not central to the decisions, but it is central to the argument being made by the original post. I think that statement is incorrect in the extreme, and that statement is what I'm arguing against.
That Google copied and pasted source files from Sun to their own product is not in dispute. That is exactly what they did. It was a decision on Google's part that mattered a great deal to this case, particularly in the lower courts.
That the files were interface declarations rather than class implementations is immaterial in the face of copyright law - that's why the justices skipped straight to fair use. The law makes it very clear that all code "directly or indirectly" invoking software functions may be copyrighted. That fact is central to Thomas's dissent and not disputed by the majority; where they differ is determining the degree to which the work is transformative in nature.
That determination that the work is transformative made by the majority is where one of the real difficulties of the case lies, and it is linked to this "just an interface" argument. You can argue (and most would agree) that the 11K LOC were "just" API definitions and that they are insignificant in the face of the other millions of lines of functionality. The majority made this argument in detail. However they also, just a few, slim paragraphs later, articulate how essential this API code is to the platform and how without it the whole platform has no value. The court simultaneously found the code in question to be insignificant to Oracle, but essential to Google. It's utterly bizarre, and frankly doesn't pass the expert smell test.
Anyone with even a passing familiarity of Java will tell you the value of a library lies in its interface rather than its implementation, and likewise will tell you there are many ways to write the same interface in functionally-compatible ways. Google didn't have to copy the interface exactly, and they certainly didn't have to copy the code directly to do so.
"The law makes it very clear that all code "directly or indirectly" invoking software functions may be copyrighted."
It also says the exact opposite. US Federal copyright law is a mess. While it is explicitly stated in the legislation that 'computer programs' can be subject to copyright, without that they would be unquestionably not copyrightable.
Since the legislation doesn't address this contradiction, courts have struggled to resolve it.
Ultimately, it seems the judges took into account that Sun refused to licence the code to Google despite it having a value, if copyright existed, of a few tens or hundreds of thousands of dollars at most, and that Sun were attempting to use copyright in a market-abusing way to monopolistically stifle invention. As the majority opinion said, that's the opposite of the purpose of IP laws.
"Anyone with even a passing familiarity of Java will tell you the value of a library lies in its interface rather than its implementation,"
Well, I have a familiarity with Java and several other languages, and I don't agree. The value of a library lies in both of those things, biased toward the implementation. A nice set of functions is an asset, but it will not rescue a library which is inefficient. Meanwhile, there are libraries in use which have terrible interfaces because their function is still too important to give up. Work to improve the interfaces of such libraries is greatly appreciated because it connects good interface with good implementation.
"and likewise will tell you there are many ways to write the same interface in functionally-compatible ways. Google didn't have to copy the interface"
If they hadn't copied the lines specifying the function names, they would have rewritten files containing the function names. They could bash the names around so they're technically different, but most of the stuff has to stay the same so the interface is functionally compatible. If they did that instead of copying, Oracle would claim that they did the same thing because the functions are all here and named the same.
>They could bash the names around so they're technically different, but most of the stuff has to stay the same so the interface is functionally compatible.
This is simply untrue. They could, for example, have adjusted the function signatures to leverage different-but-compatible types, for example by implementing Android-specific Numeric interfaces to unify disparate Math functions. This would not only have delivered a "Java-compatible API", but would almost certainly have been *a good idea* because Java 5's core APIs were diabolically badly designed. Likewise they could have extended the existing core API with Android-specific extensions, such as core library functions lifted from Guava, which mostly took until Java 9 and Java 11 to make it into the JDK.
Crucially for the case doing any of this would have put clear water between what Oracle were arguing: that the Dalvik JVM was entirelty a derivative work, because all it did was run Java APIs in a new venue.
Google opted not to do that and have been extremely lucky that the court has opted to adopt their interpretation of the law: that running byte-for-byte identical Java, just on Google-branded phones, constitutes transformation.
I think like most I'll be on balance happy that Google won this, but the way the Court has landed at the decision is going to cause issues. Most importantly they've "assumed" APIs can be copyrighted, which leaves this question open for further litigation. Or hopefully some actual legislation.
They've made a slightly weird comparison of significance/value within the case for the determination of fair use. On the one hand they've said that the 11k LOC is only 0.4% of Java so should be deemed insignificant to the whole, but those 11k LOC were essential to Android so we should be looking at the whole of Android in comparison. That should swing both ways.
The way they've determined the work is transformative is pretty challenging too. They say that the intention for Google was to get Java familiarity on a new platform - this is simply untrue. Java already ran on mobiles in both SE and ME form. The intention was for Google to get Java familiarity *without paying Java licensing*. That distiction is glossed over the in case - a big win for Google's lawyers. Similarly they base the idea that the work was transformative significantly on the amount of effort thrown at the problem - hundreds of man-years etc. etc., This ignores Google's express intention for Dalvik to mirror the JDK (right down to the bugs). This opens up scope for "fair use" to be enabled retrospectively simply by throwing manpower at a problem until the violation is too big to fail.
So, hooray, it's a happy day, but we shouldn't be too happy - we have clearly reached the limitations of current US copyright law.
we have clearly reached the limitations of current US copyright law
In fairness, that's basically what Breyer said. And the Roberts court really prefers narrowly-scoped decisions, which is why they left the "does copyright apply to APIs" question open.
Federal legislation to clarify how Title 17 applies to software would be the best remedy, if it's decent. But getting decent legislation written and passed is non-trivial. And this isn't one of those areas where the states have much discretion, so Congress can't rely on the usual legislative process of using the states as a laboratory, then copying what they like.
I thought the court was hinting that we really, really don't want them looking to closely at what is copyright in computer software, because we won't like the answer.
It is entirely possible for a ruling on a matter like this to clarify the position on, say, open source, in passing, such that all open source 'copyright' assertions in the US would fail. (That hypothetical is chosen because I suspect that's what was being hinted at. There is a real problem in US law, when it comes to is.) )
>I thought the court was hinting that we really, really don't want them looking to closely at what is copyright in computer software, because we won't like the answer.
Which is fair cop, but they've wriggled out of this by reintroducing sweat-of-the-brow tests to Fair Use, enabling well-resourced organisations to get out of copyright violations by pointint to how much money they spent doing it at scale. That's going to have consequences I don't think anyone can reasonably predict.
I'm not sure they did, but this is way out of my comfort zone. I think they made an interesting distinction between labour inputs and actual work - the point is that Google invested the programmers' time successfully in creating a major additional 'work', of which the copying was a tiny part. But I'd have to reread that section of the judgement, and ICBA.
It was just lucky that it was a tech giant such as Google which had the deep enough pockets to be able to fight this all the way to the supreme courts. If it had been some tech start up trying to depend this case for a decade, they would have ran out of money to pay lawyers years ago and Oracle could have won the case in the lower courts.
A 2019 post of mine rings true:
"Sun had the beginnings of an Android-like experience when they purchased SavaJe but didn't have the money or the will to pursue it. When Oracle came along they also saw little value in the concept of a java-based universal mobile platform and gave it no love. Oracle is simply expressing their frustration at their inability to spot golden geese when they're presented with them and going after those with enough vision to do it. APIs are simply not copyrightable anymore than anyone can copyright the use of clutch, brake and accelerator pedals in a car."
Biting the hand that feeds IT © 1998–2022