LWN: Comments on "US Supreme Court rules for Google over Oracle"

US Supreme Court rules for Google over Oracle

Wol — Wed, 28 Apr 2021 22:29:47 +0000

Yup, I know the normal spelling is Domesday. But the practical impact on the Saxons was that they went from being land-owners under Saxon law to vassal tenants under the Norman conquerors. Their lands were effectively seized and rented back.

Cheers,
Wol

US Supreme Court rules for Google over Oracle

Wol — Wed, 28 Apr 2021 22:25:58 +0000

The Magna Carta guaranteed the right to "a trial by a jury of ones peers". But it only guaranteed it to the nobility, the Peers.

Basically, it was a reigning in of the absolute powers of the King, and making him formally answerable to the Baronage. This is the birth of Parliament and the House of Lords, I guess... the King could no longer sling a Lord he'd fallen out with into the brig without at least the permission of the other Lords.

All part of the everlastng political struggle between those in power, and those who want to be in power.

Cheers,
Wol

US Supreme Court rules for Google over Oracle

nix — Wed, 28 Apr 2021 16:13:01 +0000

> (Spelt that way because, for the Saxon landowners it recorded, it WAS doomsday.)

Well, it was spelt "Domesday", but that meant the same thing. Wikipedia says (and my undergrad-level profs also told me) that this is from OE and Norse "doom" meaning "law, judgement". The implication is that this was a book of unalterable law: this suggests that the book was considered ancient uncontradictable canon of sorts, and unsurprisingly this term is first found centuries after the book itself was composed, in the 13th century. The books' compilers didn't call them that at all! The survey the books record was semi-informally known simply as the Great Survey (the first of a number of Great Surveys in English history), and the record as any of a number of things, most commonly simply as (in Latin) the "enrolling". i.e. it was named practically, for what it was, an enrolling of properties and households, not metaphorically.

The meaning "doom" as in "disaster" is many centuries newer. The OED says that while "doomsday" in the meaning of "the day of judgement" is ancient, the related expansion of the term "doom" is much newer: 16th--17th century, and so, probably, is the implication that doomsday is in any way disastrous. (I have seen older uses of "doom" to mean "disaster" than the OED cites, which I should probably pass on if I can track them down again, but there's only about 50 years in it, certainly not centuries.)

US Supreme Court rules for Google over Oracle

nix — Wed, 28 Apr 2021 15:57:30 +0000

> Some states have been more honest about this all by claiming everything belonged to them and was just temporarily granted to inhabitants

The opposite ("allodial title") is *extremely* rare these days and generally grandfathered in where possible at all with no way to place anything new under that title, most likely because almost all modern states want the right not only to tax property but also to go and do things on what you consider your property in some rare circumstances (putting out fires, arresting criminals etc).

US Supreme Court rules for Google over Oracle

nix — Wed, 28 Apr 2021 15:50:50 +0000

> The definition varies, but it is English Law that pre-dates the concept of a "Jury of one's Peers" - peers is capitalised because it means a Lord of the Realm.

Where does *this* come from? Peers is capitalized because nouns were frequently (and irregularly) capitalized in written text back then. It's got nothing to do with lordship, and juries were never composed entirely of peers of the realm (who are called that because they are distinct from *other sorts* of peers).

US Supreme Court rules for Google over Oracle

nix — Wed, 28 Apr 2021 15:49:03 +0000

Classic example: enforcement of the law in regard to things we would now consider the obvious preserve of the state like murder was a family matter in ancient Rome: the state did not get involved (except in the very rare case that the Emperor himself got interested) and it seems that matters like evidence were almost always much less important than the relative standing of the victim's family and the accused's. But things like what clothes you wore, or whether you lived in excessive luxury? That was a matter of state, and a serious one. (Though even there any role for anything like evidence was... marginal.)

The way societies are constructed is a social construct, obviously enough. The law is an aspect of the way societies are constructed, so it's a social construct. It sits on biological underpinnings shared with many other animals driving emotions governing things like perceived fairness, retribution, etc, but it is still a social construct.

US Supreme Court rules for Google over Oracle

Cyberax — Fri, 16 Apr 2021 00:30:44 +0000

I can design a very beautiful API to run a quantum computer. But it would be totally worthless without the said quantum computer.

US Supreme Court rules for Google over Oracle

mathstuf — Thu, 15 Apr 2021 22:47:35 +0000

How is their value 0? A good API still requires a usable implementation to stand behind it. As an example, the Windows API is quite valuable, but look at how fast WINE has to move just to stand still to try and keep up. Or Microsoft with the Linux API in WSL1.

Also note that this ruling was decided on fair use. It doesn't say APIs are protected by copyright. Also doesn't say that they aren't. It says that *regardless* of the status, *Google* is protected by fair use for its use in *Android*. Not for use in GCS, not Amazon for AWS.

US Supreme Court rules for Google over Oracle

clicea — Thu, 15 Apr 2021 21:23:10 +0000

I'm all for allowing re-implementations but I do not know how to square it with the fact that good, intuitive APIs are hard to design and we're, with this ruling, setting their value to 0. No idea how to balance this.

US Supreme Court rules for Google over Oracle

mina86 — Sun, 11 Apr 2021 16:21:31 +0000

The creators were compensated in the form of salary which they agreed to.

US Supreme Court rules for Google over Oracle

mina86 — Sun, 11 Apr 2021 16:05:00 +0000

> there are certain circumstances where infringement of copyright is allowed

This is a bit of technicality, but if one operates within fair use doctrine, they do not infringe copyright. In other words, fair use is not a defence in copyright infringement cases, it’s a right everyone has and if it applies than there’s no infringement.

US Supreme Court rules for Google over Oracle

Nemo_bis — Sun, 11 Apr 2021 14:58:39 +0000

Breyer always puts such juicy bits in his footnotes!

Copyright != property

Wol — Sat, 10 Apr 2021 22:26:25 +0000

Sounds like the fun we had with joyriders in the UK :-)

I believe theft is defined as "permanently depriving the owner of their property". Which copying a song most definitely does not.

Amd when joyriders nicked a car over here, I believe their defence council argued "it's not theft, they got their car back". So they had to bring in a new law of "TWOCing" - Taking WithOut Consent.

Cheers,
Wol

Copyright != property

Max.Hyre — Sat, 10 Apr 2021 20:36:39 +0000

I was once in a movie theater which played the RIAA propaganda equating copying a song with smashing a car window to steal a CD. Afterwards someone actually shouted out ``It's not theft, it's copyright infringement.''

US Supreme Court rules for Google over Oracle

khim — Thu, 08 Apr 2021 18:36:28 +0000

If Android would have actually supported J2ME then it would have been an open question.

As it is it's pretty obvious that Sun's mobile efforts were largely irrelevant.

Java was used in Android because it was taught in colleges, essentially, not because J2ME was ever viable.

US Supreme Court rules for Google over Oracle

smurf — Thu, 08 Apr 2021 17:02:12 +0000

No, assigning copyright to somebody else is not necessary by any means. There are jurisdictions (Germany for instance) where that simply is not possible. I create a work, I (or the company employing me for that purpose) hold the copyright; after I die, my heirs do. End of story.

So what happens instead? I sell an exclusive and perpetual license to exercise those rights to somebody else. Same effect as a copyright transfer but an entirely different legal theory behind it, which matters a lot e.g. when the work is translated to a new medium that didn't exist when the sale took place. Lawyers tend to have a lot of fun trying to guess the intent behind those contracts.

US Supreme Court rules for Google over Oracle

k8to — Thu, 08 Apr 2021 04:20:07 +0000

Sadly, Thomas crafting opinions out of his personal ignorance is not new.

US Supreme Court rules for Google over Oracle

k8to — Thu, 08 Apr 2021 04:16:26 +0000

Indeed, fair use is evaluated in each use. It has no bearing on the copyright itself. You could maybe construct some where case where most uses of a copyrighted work would be permitted, or more narrowly most expected uses, but that does not make the work any less copyrighted, just it may limit the practical power of it regarding restricting use.

US Supreme Court rules for Google over Oracle

micka — Wed, 07 Apr 2021 16:27:45 +0000

Well... That's actually pretty recent :)

US Supreme Court rules for Google over Oracle

Wol — Wed, 07 Apr 2021 15:35:45 +0000

> In fact I’m pretty sure the earliest recorded property registers were actually tax registers.

The Doomsday book, for example (1086).

(Spelt that way because, for the Saxon landowners it recorded, it WAS doomsday.)

Cheers,
Wol

US Supreme Court rules for Google over Oracle

nim-nim — Wed, 07 Apr 2021 14:49:37 +0000

Property is 100% made up by whatever state you live in. Property is basically what the state is ready to defend with police and army, and what it wants to tax in return. In fact I’m pretty sure the earliest recorded property registers were actually tax registers.

Some states have been more honest about this all by claiming everything belonged to them and was just temporarily granted to inhabitants (most recently, the Soviets, but that is still the case for some kinds of “property” in some parts of the world, with foreign investors treating such grants as “property” because they effectively behave exactly the same way).

“Natural rights property” is another fiction invented by people that want to convince states they do not owe any taxes. It only works on very corrupt administrations.

US Supreme Court rules for Google over Oracle

plugwash — Wed, 07 Apr 2021 12:59:10 +0000

> Sun tried from the very beginning, to get java into mobile phones. And failed. Miserably.

There was a time, before the iphone and andriod took over the world, when J2ME support on phones was not uncommon.

Would android still have taken over the smartphone world if it had not used the Java programming language and APIs? If not would J2ME still be a thing? I doubt anyone really knows.

US Supreme Court rules for Google over Oracle

nim-nim — Wed, 07 Apr 2021 08:17:24 +0000

BTW, if you want to dissect the judgment, when the other judges agreed on the transformative character of Android, as one reason it was all fair use, they basically agreed existing Java positioning on mobile was crap and ripe for disruption. Which also means they disagreed with Oracle’s inflated claims of Java ME valuation.

US Supreme Court rules for Google over Oracle

nim-nim — Wed, 07 Apr 2021 08:08:20 +0000

Java ME valuation was always going to crash with or without Android, it was an astonishingly bad product constrained by pre-smartphone design decisions, had Android tanked WebOS would have had the same effect on its valuation (the whole point of Java ME was to pretend handhelds were not full-blown computers and did not need a full-blown execution environment).

However, because only one judge bought Oracle’s argument that Java ME was a great product betrayed by Google, no one bothered to disprove it.

US Supreme Court rules for Google over Oracle

gfernandes — Wed, 07 Apr 2021 05:51:46 +0000

Everyone has their own perspective /opinion. Having said that, there is no way you can regard the use of a java like language and API for programming Android, as competing with java.

Sun tried from the very beginning, to get java into mobile phones. And failed. Miserably.

So the value add of Android is the transformative use of the java API, clubbed with a Linux kernel, and a custom user land layer.

Don't forget that.

Java, by itself, only ever managed to conquer the server.

US Supreme Court rules for Google over Oracle

rgmoore — Wed, 07 Apr 2021 01:23:46 +0000

"the fair use of a copyrighted work […] is not an infringement of copyright." means—it flat out says—that copyright does not cover fair use.

No, that's not what it says. What it says is that fair use is an exception to the general rule that you aren't allowed to copy a work that's protected by copyright. Very importantly, the determination of fair use depends on the details of the derivative work, not the original work. One derivative work may make fair use of a copyrighted work and thus not be a violation of copyright, while a different derivative work may use the exact same part of the original but not be fair use.

To give a concrete example, consider one short scene from a movie. If a professor of film studies makes a lecture that uses that scene, that lecture may very well qualify as fair use. It is for non-profit educational use, it is transformative, it doesn't compete with the original work, and the excerpt is small in relation to the whole lecture. By all the prongs of the fair use test, it should qualify, and the copyright holder would likely lose if they sued the professor.

But now imagine if someone took the exact same scene from the movie, put a frame around it with some ads, and posted it to YouTube. It was done for profit, wasn't transformative, does compete with the original movie, and the excerpt is large in relation to the whole. A court would undoubtedly rule the use was infringing. That would be true even if the same court had ruled in favor of the film professor. Finding that one use didn't infringe doesn't invalidate the copyright; it just says that specific use was allowed.

In this case, the court ruled that Google's use didn't infringe. That doesn't touch on the deeper issue of whether the files Google copied could have been copyrighted in the first place. It's possible that header files, by themselves, are not sufficiently expressive to be copyrightable. A lot of people have been pushing for the courts to rule exactly that. The ruling in this case hints in that direction, but it doesn't say it outright. As of today, those files are still under copyright, and Oracle could sue the next company that tries to copy them.

US Supreme Court rules for Google over Oracle

bauermann — Wed, 07 Apr 2021 00:16:56 +0000

I only have a very superficial understanding of law, but my understanding is that some legal concepts aren't considered to be made up, but rather a reflection of reality (e.g, natural rights IIUC).

Some people would put property in that category, though that is controversial. In any case, the fact that one concept is thousands of years old and another is a few centuries years old is IMHO relevant.

In any case, I just brought this up because conflating property with copyright is a strategy that has been used to argue for stronger copyright protections without bringing up the necessary discussion about the benefit such change should bring to society. Therefore it's important to be aware that it's an analogy with important flaws.

US Supreme Court rules for Google over Oracle

bauermann — Wed, 07 Apr 2021 00:04:49 +0000

I thought it was important to mention that copyright isn't property because this imperfect analogy has been used in the past to argue for stronger copyright protection while avoiding to discuss any corresponding benefit to society.

It's an argument Lawrence Lessig made in the Free Culture book far better than I ever could. 🙂

US Supreme Court rules for Google over Oracle

rodgerd — Tue, 06 Apr 2021 22:53:58 +0000

Even Scalia, who routinely relied on Thomas' vote, had nothing good to say about him.

US Supreme Court rules for Google over Oracle

jthill — Tue, 06 Apr 2021 20:11:51 +0000

I'm not sure. I believe there's a sound argument in support, so I posted, that is about as sure as I get on this stuff, but I don't play in the same league as supreme court justices, so me thinking my argument's sound doesn't make me feel any too very sure, no.

"the fair use of a copyrighted work […] is not an infringement of copyright." means—it flat out says—that copyright does not cover fair use. Remember that §107 was codifying centuries of common-law precedent, cases judges saying ~no, copyright doesn't cover that (or does), go away~. The justices walked through every factor they could find that would weigh on whether Google's use was fair (as they point out, the statutory factors are noted but not exclusive), and in Google's use they all came out in Google's favor. So *this* case is basically the poster child for a fair-use slam dunk. Copyright authority cannot be extended to interfere with fair use. Pointing out that copyright covers, that holders are still granted the authority to prohibit and get the state to enforce penalties for, unfair use, doesn't have any bearing on whether it covers fair use: it doesn't. It can't. That's beyond the time-honored and now statutory limit of copyright ablility.

So the question that remains is whether reimplementing API's is going to *necessarily* be fair use. In any real use, I think so.

For starters, as the opinion spends some effort elaborating, it's incorrect in this context to regard the entries in an index as some sort of independently-copyrighted work. What Google copied was 0.004 of the only clearly copyrightable work in evidence. I dare say that shouldn't surprise anyone here, I wouldn't expect anyone here to credit any notion of a market for some "Java API header-files-equivalent" work, nor the equivalent argument that there's some plausible transaction where having just the headers would be any viable substitute for an entire original work. I think they pretty much put paid to the silliness of the "but they copied all of what they copied!"-vuvuzela-blowing cohort. Substantiality will necessarily favor reimplementers.

They got API's right:

The nature of the work at issue favors fair use. The copied lines of code are part of a “user interface” that provides a way for programmers to access prewritten computer code through the use of simple commands

[and not the prewritten code itself] so "nature of the work" will always point to fair use.

"purpose and character" won't necessarily do that. Google's use clearly does, as does say WINE's. I think it's good that cases are decided on the facts before them, and that courts don't spend time adjudicating cases that don't actually arise. So I really don't care about the contrived ten-lines-at-a-time effort. The law isn't about imagining every conceivable bad behavior and launching preemptive strikes, because there'd be no end. It's enough to deal with behaviors people actually engage in, eagerly anticipating more is just borrowing trouble, wasted effort.

So of the four statutory factors, three (substantiality, nature, market effect) will always land squarely in fair use territory, and "purpose and character" would only be troubling when considering a direct competitor with some clearly detrimental tendencies. I'm having trouble imagining such a thing, implementation is A Lot Of Work; and remember that IBM got hit with anti-trust for trying to conceal their interfaces (this was before copyright was codified), copyright is a created artifice, instituted to promote progress, not to entitle permanent possession.

US Supreme Court rules for Google over Oracle

NYKevin — Tue, 06 Apr 2021 19:19:21 +0000

My favorite paragraph in the majority opinion:

> JUSTICE THOMAS’ thoughtful dissent offers a very different view of how (and perhaps whether) fair use has any role
> to play for computer programs. We are told that no attempt
> to distinguish among computer code is tenable when considering “the nature of the work,” see post, at 10, even
> though there are important distinctions in the ways that
> programs are used and designed, post, at 18 (“The declaring
> code is what attracted programmers”). We are told that no
> reuse of code in a new program will ever have a valid “purpose and character,” post, at 16, even though the reasons
> for copying computer code may vary greatly and differ from
> those applicable to other sorts of works, ibid. (accepting
> that copying as part of “reverse engineer[ing] a system to
> ensure compatibility” could be a valid purpose). And we are
> told that our fair use analysis must prioritize certain factors
> over others, post, at 9, n. 5, even though our case law instructs that fair use depends on the context, see Campbell,
> 510 U. S., at 577–578.

Breyer is basically saying "See how wrong Justice Thomas is? I barely even have to respond to his arguments, he practically refutes himself." All cloaked in a respectful and collegial tone, but damn, this is some savage legal writing.

(Disclaimer: I work for Google and wanted Google to win, but those two facts are mostly unrelated.)

US Supreme Court rules for Google over Oracle

zuki — Tue, 06 Apr 2021 18:55:34 +0000

> It is still interesting to look at the dissenting opinion.

It is!

>>> Declaring code defines the scope of a set of implementing code and gives a programmer a way to use it by shortcut. Because declaring code incorporates implementing code, it has no function on its own.

I think Justice Thomas gets this crucial fact wrong. I think any programmer would say that the opposite is true: "declaring code" does not "incorporate implementing code", and very much has a function of its own. As an example, in test-driver-development we are encouraged to create "declaring code" without any implementation, and then use it, first without any implementation… And obviously there are many examples of implementations created independently to an API specification.

>>> The functionality of both declaring code and implementing code will thus typically rise and fall together.

This is also not true.

And the rest of his argument seems to be based on this (false, imo) understanding.

He also summarizes the majority opinion nicely:

>>> The result of this distorting analysis is an opinion that makes it difficult to imagine any circumstance in which declaring code will remain protected by copyright.

Yay!

US Supreme Court rules for Google over Oracle

rgmoore — Tue, 06 Apr 2021 17:46:26 +0000

Licences are not "transfers" of copyright, not even temporarily: only the copyright holder has the right to licence the work.

Yes and no. An exclusive license looks a lot like a temporary transfer of the property, especially if it comes with sublicensing rights. I think this gets at a bigger reason our legal system tends to look at things like copyright as forms of property: it's a useful abstraction. Just as Unix treats all kinds of things as files, even when they aren't, the legal system treats all kinds of things as property even when they aren't. Our legal and business systems are built around the concept of property, and they've spent centuries learning how to reason about it. Treating more and more abstract kinds of things- contracts, obligations, debts, government granted privileges, or what have you- as property lets lawyers and business people apply those centuries of case law and business processes to them. Even if the abstraction is imperfect so there need to be new principles added, like fair use for copyright, starting from well established principles means they aren't inventing everything from scratch.

US Supreme Court rules for Google over Oracle

amacater — Tue, 06 Apr 2021 16:47:41 +0000

Strictly, 6th July 1189 - the accession date of Richard I. It' s a cut-off date: claims for possession of land reference that date - anything earlier can't be overturned, effectively. [Legal history appreciated here :) ]

US Supreme Court rules for Google over Oracle

Wol — Tue, 06 Apr 2021 15:20:02 +0000

> And I'm sure there are (historic) societies that do not have one or more of the legal concepts we now take for granted.

Time Immemorial?

Which is a real thing. The definition varies, but it is English Law that pre-dates the concept of a "Jury of one's Peers" - peers is capitalised because it means a Lord of the Realm.

So basically, "time immemorial" is stuff that predates either Magna Carta, or the accession of Henry III, whichever date you choose as they're both roughly the same.

Cheers,
Wol

US Supreme Court rules for Google over Oracle

rgmoore — Tue, 06 Apr 2021 15:03:07 +0000

The fair use test is actually a complex one. It's not a simple question of whether the copying falls into a bucket of allowed conduct. Instead, the court is supposed to consider multiple factors and decide how they balance each other. Those factors include:

The amount of copying, both absolute and relative to the whole work
How it's being used, i.e. commercial or non-profit
Whether the copied work is likely to compete with the original
Whether the copying is transformative

In this case, the amount of copying is small, both in absolute amount and relative to the total size of the project, it's used commercially, the court determined it's not likely to compete with the original, and the use was considered to be transformative. Based on that, the court decided it was fair use. But they easily could have decided otherwise if they felt it was likely to compete with Java. I'm not sure the decision was correct- I think it really is competing with Java- but deciding on fair use let the court sidestep the question of whether the code was copyrightable at all.

US Supreme Court rules for Google over Oracle

danielthompson — Tue, 06 Apr 2021 13:00:32 +0000

I particularly the comment in the syllabus about: "Unlike many other computer programs, the value of the copied lines is in significant part derived from the investment of users (here computer programmers) who have learned the API’s system.".

This recognizes that the value of an API arises, in part, from skilled people learning to use it rather than through the effort of the author and that this shared investment by all parties weighs in favour of fair-use.

US Supreme Court rules for Google over Oracle

dottedmag — Tue, 06 Apr 2021 12:33:58 +0000

One of the major points discussed was the fact that Android does not compete with Java SE, being different platforms, so other kinds of API copying may be found to be not covered by "fair use".

US Supreme Court rules for Google over Oracle

farnz — Tue, 06 Apr 2021 11:44:52 +0000

Not quite; the court made no decision about whether or not the code in question is copyrightable, merely that fair use would cover the alleged infringements, and thus there's no reason to think about whether or not the code is copyrightable, because either it is, but fair use means no infringement, or it isn't and thus copyright cannot be infringed by Google's actions. Either way round, the result is that Google has obeyed the law.

US Supreme Court rules for Google over Oracle

sabroad — Tue, 06 Apr 2021 10:35:02 +0000

> And if you're going to allow people to transfer the copyright temporarily in exchange for a fee, it would be strange not to let them transfer it permanently.

Licences are not "transfers" of copyright, not even temporarily: only the copyright holder has the right to licence the work.

The reason for copyright assignment is not because licences exist but that work-for-hire is the norm.