Hacker News Reader

Comment by doesnotexist | Hacker News Reader

Should Oracle win this case, I predict that Free Software licenses will emerge as the clear market choice. Licenses that enshrine interoperability along with the freedom to modify or reimplement in part or wholesale will suddenly provide significantly more value over non-Free software which will be a proprietary minefield and the value will not be for technical but purely from legal liability. Anyone setting out to do something new would be wise to build on top of software they are at least assured will not be used against them in court should they find success building ontop of it.

View on HN

The Google lawyer should be fired. I can’t believe you’d put the best of the best in front of the SC and the guy didn’t even sound like he could describe what an API actually is in non-technical terms. I don’t see how he represented the best legal representation Google could buy.

Indeed. This was bad representation to the point that I'm wondering if it was intentional. Google might have recently realized that, as a monopoly, fighting for API freedom perhaps isn't in their best interest.

was thinking exactly same thing

That would have been an expensive realization, but it sounds like a reasonable explanation.

If they wanted to lose, wouldn't it have been much easier (and possibly cheaper) to settle?

I can't imagine Oracle would have a problem with getting to say they were right, and getting a boatload of cash, without all the bother of the Supreme Court.

And doesn't intentional bad representation have weird and costly externalities, like reputation damage for the lawyer?

I love the image of Google and Oracle behaving like gangsters, but I'm having trouble imagining it being this sloppy.

A Supreme Court ruling is a lot better than settlement if they've decided they want to pivot to sueing independent implementations of their APIs. I wouldn't find it compelling to follow if Google took the other side of a settlement they were involved in to go and pursue others. Maybe they settled because they figured out they could make more money that way, regardless of the question of law.

Interesting! Leaves me wondering, what independent implementations of Google APIs are out there?

One of the more important in my opinion is the microG project (https://microg.org/), a free re-implementation of Google’s proprietary libraries for Android that many apps on the Play store depend on.

> If they wanted to lose, wouldn't it have been much easier (and possibly cheaper) to settle?

Well, it would certainly be much safer not to present such a comprehensive written argumentation and hope that you can somehow sabotage all that (and all the written arguments from friendly amici, etc., by a clumsy presentation at oral argument.) I think "highly successsful Supreme Court lawyer doing his best has a bad day at oral argument, perhaps in part due to the unusual format" is more likely than some kind of "deliberately throw the case by trying to screw up oral argument" thing.

Google would permanently lose the tech world if they don't appear to fight this battle. Give away the whole concept of API's free of copyright? The reputational damage would be immense and probably spell the beginning of the end for Google.

I had the same thought (that Google is intentionally throwing the case) when I read over the Google lawyer's approach, there's no way the guy is that out of touch.

The Google of today is not the Google that started this case all those years ago and the Google of today is working alot harder to create its own walled gardens. They have alot more to gain from suing the pants off anyone who tries to interoperate with them, than they stand to lose from a few $billion payoff to Oracle.

???

Dunning-Kruger effect, exemplified in this thread

> Dunning-Kruger effect, exemplified in this thread

How does the effect where everyone feels they are slightly closer to about the 70th percentile in performance at a task than they are (and more the farther away from that their actual performance is) have any bearing here?

Or do you mean the D-K effect of popular legend where people who are completely incompetent think they are hypercompetent, rather than the actual effect of that name?

>Indeed. This was bad representation to the point that I'm wondering if it was intentional. Google might have recently realized that, as a monopoly, fighting for API freedom perhaps isn't in their best interest.

That might well be true. However, I'd be more inclined to apply Hanlon's Razor[0] to this, as it's usually closer to the truth.

[0] https://en.wikipedia.org/wiki/Hanlon%27s_razor

> This was bad representation to the point that I'm wondering if it was intentional. Google might have recently realized that, as a monopoly, fighting for API freedom perhaps isn't in their best interest.

Is your speculation that Google had this sudden realization between the filing of the various written arguments (including the standard of review for fair use memo on August 7) after fighting this case for years (all of which they've had a similar position), or do you think their written arguments are consistent with this theory?

4D chess time: Google wants to steal market share from Oracle, and Oracle offered a perfect opportunity by trying to shoot itself in the foot by making Java forks risky (thus ending innovation from the open-source world). I expect Google to invent a new bytecode language that looks and feels like Java, but isn't Java. Oh and it'll come with a promise not to sue if you copy the API, not like that big bad Oracle.

He also argued that there was only one way Oracle's Java libraries could have been designed, which is clearly false, and the justices picked up on that.

really astonishing. wonder what google saw in him. the oracle lawyer was far more prepared and technically cogent.

After a (long) while it became clear it meant "there is only way that we might make the full use of the current pool of Java programmers", which is an odd argument to advance — why would they be entitled to draw from that pool of programmers. I think it would have been more compelling if he had focused on the interoperability with the ecosystem of Java libraries.

And the justices picked up on that. I thought they did a good job assessing the case from both ends. At first I thought the justices were against Google and pro Oracle, until I saw them address the Oracle lawyer and it was clear they were asking good questions on both sides.

Giving a lawyer an unwinnable case does not take away from the fact that he is one of the most successful SC lawyers in the world.

This case is not unwinnable. And lawyers can specialise in specific areas. This lawyer seems unprepared at best, and simply not appropriate for this at worst.

I'm not saying he's a bad lawyer, I'm saying he's not the right lawyer for this case.

>I'm saying he's not the right lawyer for this case.

Feel free to name anyone who has a better record in front of the SC.

His claim that there is no creativity involved in designing an API is plainly absurd as well, and from the transcript I don't think the justices bought that at all. And I agree with the Justice Department lawyer that the "no other way to do it" argument he pushed at the beginning is circular.

Yes. They have had -years- to come up with a defense and a crack legal team and then this guy shows up. Maybe they wanted to lose?

Google... really isn't that good at what they do besides sell Ads. They're just big so whatever they choose becomes the standard because 'doing what google does' not only won't generally get you fired but also impresses people who don't actually have to build difficult things. It's a testament to the power of an entrenched market position.

How do you know the Google lawyer isn't representing the wishes of his client? A little voice in the back of my head wonders if maybe google's position has changed over the years...

Hanlon's razor may apply here: "never attribute to malice that which is adequately explained by stupidity"

albeit in a different form.

I find Hanlon's razor a wishful thinking at best. I do not think there is any evidence backing it up.

My least favorite part of the arguments was when the US government representative used the existence of open source licenses that explicitly allow copying declarations to claim that it's not commonly believed that such rights don't exist.

"There is a prevalent practice of licensed copying of declarations and often that is done through what is called open-source licensing" ...(6 minutes later) "sometimes, often the license terms don't include the payment of money, they simply include a requirement like whatever improvements you make have to be given back to the programming community. Have to be made known to other potential programmers. But the copyright holders authority to impose and enforce those licenses obviously depends on the proposition that the code is copyrightable to begin with. And so those licenses would be a pointless gesture otherwise and the very fact that they are offered with such frequency tends to dispel the idea that there is a common understanding in the relevant community that this material is not copyrightable at all."

it's a good point though

I agree that part of the reason open source software is so popular (particularly for programming language implementations) is so that there is no ambiguity about the fact that people can make adaptations/alternatives without worrying about copyright lawsuits. However it seems very perverse to use the fact that so many people think it would be a bad idea if these interfaces were constrained by intellectual property laws to justify constraining them by intellectual property laws.

> licenses would be a pointless gesture otherwise and the very fact that they are offered with such frequency tends to dispel the idea that there is a common understanding

He's right. The entire premise of open source licensing is that the work is copyrightable, so it uses the copyright system to enforce the authors distribution requirements.

It still seems like an overstatement to argue that use of open source licenses indicates a belief that method declarations are copyrightable.

Open source licenses have often included terms that were for other jurisdictions, or because the status of something is actually unknown, or just in case some future law changed the accepted belief.

I don't have to believe in Santa Claus to appreciate Christmas presents.

I don't believe that software should be subject to copyright so I put the GPL on my code to protect it from those that do. Fight fire with fire, eh?

The GPL, in particular, goes further than just refusing to exercise your copyright over a piece of software. The GPL restricts the usage of your software, most notably by forbidding its use as part of any proprietary software. Without copyright no one would have to respect that restriction.

The situation you are describing is closer to a project with an "unlicense," e.g. SQLite, which explicitly makes the project public-domain and places no restrictions whatsoever on its use.

Yes, the GPL actually depends on strong copyright.

who knows, maybe strong APIs will prevent embrace/extend/extinguish?

Glad to see this take, it was in line with what I was thinking. Namely, that APIs have always been an extremely poor excuse for true openness, more or less a way to get that half-freedom of "letting other people do your work for you" without always being required to give back.

AWS licenses it's client library as under the Apache License. (https://github.com/aws/aws-sdk-java/blob/master/LICENSE.txt) Doesn't that mean anyone can build an interoperable service? I can then benefit from other cloud providers implementing the same API. I don't need access to AWS' source code, I wasn't going to deploy a private cloud anyway. All I want is robust competition.

Yeah, I think this is what Digital Ocean does? I believe the AWS S3 Python library works out of the box for DO.

This has already happened to an extent. For example, there are bunch of API compatible implementations of s3, both oss and proprietary and most of them suggest you use the AWS sdk as a client.

Including, amusingly enough, Oracle:

https://docs.cloud.oracle.com/en-us/iaas/Content/Object/Task...

Wow, that's wild to read. In comparison (and complete lack of contrast) to:

https://docs.aws.amazon.com/AWSJavaSDK/latest/javadoc/com/am...

What the hell kind of comparison is that? You linked a human-doc-writer written documentation to auto-generated Java SDK documentation...

I would guess that it doesn't work out that way. The Apache license has the patent clause, but it doesn't have a comparable "API copyright" clause. Though perhaps another consequence of an Oracle win is that we end up with an Apache3 license.

They've given a license to use and extend the copy-written software, including its method signatures.

It's copyright, not copywrite

https://en.wiktionary.org/wiki/copywrite

Oracle has done that, too. They release the full JDK under the GPL.

The law around these sorts of things can get pretty hair-splitty. My guess is that the situation here is that Amazon client libraries and OpenJDK are distributed with a license to create derivative works that are based on their respective products, but that these licenses do not necessarily grant a license to create a new thing that works the same out of whole cloth. Should Oracle win the case, that would seem to imply that the Court believes they do not. Alternatively, the fact that open source licenses do not appear to have even been brought up in the course of these hearings would perhaps imply that the existence of open source libraries that implement these APIs is legally irrelevant.

OpenJDK did not exist when Android was developed and initially released. Google did move to using it 2016 and may not have liability after that date.

Andriod was First released in Sept 2008,

OpenJDK was first released May 2007...

Perhaps more legally relevant, Android's first public beta dropped in late 2007. But still, technically after OpenJDK's release.

I don't follow. If someone can determine the API by referring to code released under the Apache licence, what copyrights could they be infringing by building a different implementation of the API using the Apache-licensed code as a reference?

For a copyright infringement to have taken place, there generally needs to be an unauthorised instance of recording or of duplication, of some copyrighted work.

I put generally as, I believe, precisely recreating someone else's photo can still count as an infringement of their copyrights, despite that you haven't copied the image itself in the usual sense of making duplicates. Singing someone else's song can also infringe on their copyrights over the song. I don't think this would apply here though.

(Disclaimer: I'm not a lawyer, I could well be missing something obvious.)

The same sort of API copyright that Oracle is trying to establish in this case.

How things actually shake out would depend a lot on the specifics of the Court's ruling, and I am not a lawyer either. But, if the court rules that Oracle owns a copyright to the Java APIs, and that this means they can prevent others from implementing their own versions of those APIs, and this right remains in effect even though they release a full implementation of them under GPLv2, then I can't see any particular reason to expect that things would work differently for Amazon's APIs and Apache2.

The difference is that the way Google used the APIs is not compatible with code use under the GPLv2 (in particular, they didn’t GPLv2 license it). If the original code implementing the APIs is Apache licensed however, you have a very free hand in changing the original code so you probably wouldn’t care about your working being considered a derivative of it. I’m not sure that absolves you categorically but it definitely sounds like a better position to be in.

But the API must be defined in the client library, and that has been released under an open source license.

Using a client is different than reimplementing compatible classes.

The idea is that a client generally contains a copy of the API (it's calling it so it should).

So you get a copy of the API with Apache license and are free to build your own implementation of it.

To me it's not related, the SDK is under the apache licence but it's just an implementation of the API, the API itself is not defined or under the copyright of that SDK.

Yes, but we'll get a major bit rot problem with existing programming languages and probably even file formats (PDF anyone?). And that's before we get to open-source drivers for mainstream hardware, whose APIs are surely not soon going to be released from copyright?

Maybe after the dust settles. But you can bet lots of people will reenact the SCO saga, and now they won't have to lie on every claim.

I would love to see IBM try to sue Oracle for the entirety of its history of database revenues over its use of their SQL API...

SQL is an ISO standard. IBM would only have grounds if any part of Oracle's implementation implemented a proprietary portion of an IBM database implementation. This is the positive thing about standards, it removes ambiguity around what people can/should implement and present to users of the systems.

The IBM implementation of SQL far predates the ISO standard. Does something being standardized somehow relinquish the copyright? The idea of APIs being copyrightable is new, has IBM somehow relinquished its copyright claims for SQL to ISO? This article discusses the possibility at length https://arstechnica.com/tech-policy/2020/03/before-it-sued-g...

I think you have to have demonstrated that you have taken steps to protect your copyright. I’d imagine submitting to a standard would relinquish that. Patents on the other hand ...

I have heard that a lot for trademarks, do you have reference for copyrights?

You absolutely do not have to protect a copyright. Even if a book is out of print, you can't start selling it--or even giving copies away for free. However, although IANAL, I assume that submitting something to a standards body and having it accepted allows those standards to be used in accordance with the terms of the standards body--which BTW may not require that use be free in all cases.

> Even if a book is out of print, you can't start selling it--or even giving copies away for free.

That's not exactly true. You can't photocopy it and start printing your own.

But if you have legal copies of it, you can definitely sell them or give them away.

First-sale doctrine of physical artifacts. But not copies of those. Correct.

Yeah but what did Oracle copy here? They made the language compatible, but did they copy the API too? Cleanrooms have been a thing for a long time.

Also note that even for established patents, a ton of standards standardized patented technology, and implementing the standard requires a patent license! This is a huge part of how Qualcomm makes its patent revenue from 3G/4G implementors.

From what I heard, a large part of the politics in 3GPP standardization are about big companies wanting the prescribed methods to use their patents.

I would like to see Amazon sue Oracle for their S3-compatible API for Oracle cloud.

Or proprietary software could just publish their interoperable APIs under free license.

I look forward to all the unnecessary type casting we will get to do so public interfaces don't infringe. System.Out.Println(byte []) is my new favorite function on Android.

Would a tool to do this be illegal?

This tool would -- sort of -- implement an API.

Would make for an interesting case ...

When I learned that Java is open source, I had assumed that it was free. I guess I was wrong.

There's open source and really open source, it seems. Just wondering if the ruling will put a bomb under the Java edifice? Maybe it will even pave the way for Kotlin to become the dominant JVM language. One can only dream.

Yes, for the last decade or more commercial companies have been polluting the term "Open Source" to mean "this software takes away all your freedoms, but at least you can read our source code."

This is why the FOSS licenses enshrine Compilation, modification and redistribution also.

Google isn't complying with the GPL license, so it doesn't really matter that it's open source.

Also, I forget if openjdk or this infringement happened first, but if the latter was first than it wasn't open source at the time.

> Should Oracle win this case, ... Free Software ...

Hmmm... First Free Software project that will dye after this is WINE that implements Windows API ( https://www.winehq.org/ )

Maybe just in the USA. But also Microsoft is on Google's side here. Also doesn't Microsoft's WSL also have this problem as a "reverse WINE" ?

Except Linux is FOSS so Microsoft has a license to do it.

Only under Linux's GPL license, WSL is most definitively not under GPL

Probably related to this https://spdx.org/licenses/Linux-syscall-note.html

Is it possible to make WSL a GPL software without making entire Windows a GPL software?

Yes, GPL defines some boundaries but as long WSL acts as stand-alone software and Windows doesn't link against it that's allowed.

Not at all. Please read page 14 of the amicus curiae brief in support of Google by Microsoft (https://www.supremecourt.gov/DocketPDF/18/18-956/128381/2020...). Here is the relevant paragraph:

""" In another example from the 1990s, an open-source developer created a program called WINE, which al- lowed developers to enable Windows applications to run on computers that used the Linux open-source system, without explicit authorization from Microsoft. Gratz & Lemley, supra, at 611. To create WINE, the developer “use[d] the same hierarchy of function names” of various Windows APIs. Id. at 612. Years later, Microsoft created “the inverse of WINE,” reimplementing the structure of certain Linux APIs to create the Windows Subsystem for Linux, a program that allowed Linux programs to run on Windows. Ibid. The Windows-Linux experience shows that reuse of functional code is a “two- way street” that benefits both the original creator and the follow-on developer—and ultimately the consumer. """

Not to mention there's no monetary chest at the end of that fight (and the creation of WSL wouldn't bode well for them if they were to make a case at all).

Companies don't bother with principled consistency, and so I wouldn't expect any from Microsoft if this goes Oracle's way. It's totally possible that Microsoft thinks APIs being non-copyrightable is good for them on net, but that if SCOTUS decides otherwise that they might as well go crush WINE anyway since they now have the means.

It will depend a lot on how wide a decision SCOTUS makes, and that fair use is interpreted in a very different way. Wine is not a windows replacement nor can Microsoft really show any negative impact on the market for windows license sales because of Wine. Wine (recursive backronym for Wine Is Not an Emulator) is "simply" a compatibility layer that convert between Windows ABI calls into native linux system calls. Wine is not an operative system replacement for windows, and wine user base compared to windows is comparable so tiny that any argument based on market impact should be taken with a lot of skepticism.

In addition I would be a bit surprised if wine has copied 11000 lines of header files from windows, arranged identically to Microsoft Windows source code.

Wine allows running windows applications without windows lincense, directly taking revenue from Microsoft who arguably invested a ton in making documentation and developer tools to make those applications.

It seems you are arguing mainly because the user base is so much smaller it does not count. The law does not work that way.

Windows SDK contains tens of thousands of lines in header files for those DLLs Wine implements. Not saying Wine copied them outright, but the opportunity is clearly there.

"The fourth factor measures the effect that the allegedly infringing use has had on the copyright owner's ability to exploit his original work. The court not only investigates whether the defendant's specific use of the work has significantly harmed the copyright owner's market, but also whether such uses in general, if widespread, would harm the potential market of the original. The burden of proof here rests on the copyright owner, who must demonstrate the impact of the infringement on commercial use of the work."

It would be up to Microsoft to demonstrate that Wine impact the market for Windows license sales. I strongly doubt they can do that regardless how much they have spent on documentation and developer tools. Wine does not replace the need for an operative system, and if the usage ever become widespread enough to harm the market for windows then developers would just support programs natively for Linux, side stepping the need for wine.

Wine is not a market substitute for Microsoft windows, which is the first consideration that courts look at. Courts would then look if there is a potential existence of a licensing market (ie Microsoft licensing out the ABI) which again does not seems very likely. Failing those two tests it would be a rather uphill battle for Microsoft to demonstrate Wine impact on the market for Windows license sales.

Yet, if you pick the right lines out of the WINE source-code, it will look eerily similar to the Windows source-code. And not by coincidence: if Windows' code had looked different, WINE code would have looked different too. And the important thing is that the Windows code could easily have looked different. It was a creative process that led to what it was. WINE is a derivative work.

Copyright law needs to be clarified to allow this without licence. Just like you're allowed to refer to a competitor's brand if you follow the rules in trademark law. It's one of those things where you think legislation should be rushing to fix it to remove uncertainty. Instead, a good chunk of senators is comfortable letting the Ellisons of this world make other people's life harder in negative-sum games.

What would the benefit to MS be to crush w WINE. They have made a big effort to rebuild their reputation with developers, and WINE doesn’t really hurt their bottom line. They have no incentive.

The risk of WINE or ReactOS becoming, or contributing to, competitors. Proton is a fork of WINE by Steam that is used to run Windows games on Linux, and it works very well. One day in the future it might make sense to go after Steam, or CodeWeavers.

You’re suggesting that Microsoft may choose not to crush WINE even if they had the right, because some other incentive holds them back. I think GP agrees, and is merely saying that Microsoft’s current philosophical beliefs would not likely be that incentive.

And what about Xamarin?

Microsoft owns Xamarin (as of February 2016), so that would be weird :b

This would be somewhat concerning if we were talking about Ballmer-era Microsoft where they were going around extorting companies that used Linux. Although nothing was really stopping them then. They had a good system of picking companies one by one, being threatening and making them sign "licensing deals" without ever needing to go to court or publicly say which patents were being infringed. (Probably all FAT stuff. Google eventually had the good sense to stop using FAT on Android, but it was slow going because if you haven't noticed Google manoeuvres about as well as a freight train). With the amount they've changed under Nadella, I'm not sure present day Microsoft even could do that.

But for present-day examples we're going to need to search elsewhere.

What I meant is that Xamarin reimplements the APIs from iOS and Android.

Seems like you've discovered the philosophy of Accelerationism!

Sometimes it does feel that embracing awful system is the best way to topple it.

I don’t understand why software and other IP rights aren’t de facto open-source after 10 or 20 years, if nothing else for archive and legacy purpose, but also to encourage sustained innovation. I’m not even leftist; but even in the capitalist/libertarian theory, the state shouldn’t help guard private companies’ rights in software. If you publish it, the recipient has it, period. The state lending its judges to enforce exclusive rights for a few years is already a big extraction of collective resources for private interests.

Found the source of Windows XP? It’s alright, you shouldn’t be condemned for publishing it, it will only help with bugfixes and security, and as long as the owner refuses to sell it...

> I’m not even leftist; but even in the capitalist/libertarian theory, the state shouldn’t help guard private companies’ rights in software. If you publish it, the recipient has it, period. The state lending its judges to enforce exclusive rights for a few years is already a big extraction of collective resources for private interests.

How is software, something of which there is an infinite supply that doesn’t exist until someone creates it, more of a “collective resource” than my back yard, which was here before I was born and will remain here after I die?

Millions of people can use VI at once without any direct impact on Bill Joy or his even even noticing it.

Millions of people using his back yard at once would be an entirely different matter!

That means a good is non-rival. Why does that make it more or less “communal?” VI wouldn’t exist without Bill Joy, but his backyard would. Doesn’t that make VI more personal and less communal?

What is communal is the protection provided by the state. When the state has to protect houses, it’s an easy job, very deterministic, and costs are known. But states routinely refuse to protect large sums of money and require that you acquire your own protection. Something which is too costly for the commons or to vague to determine will be excluded from public protection. Protecting ideas is excluded for example, because they spread so fast that it is practically impossible to protect. I’m of the opinion that bytes are too costly/too vague to be protected by the state, and today is a good example since we had to use all courts up to the Supreme Court to rule on what sounds like a frivolous detail to most of the people. Ruling on whether an API is copyrightable is maybe a sign that the state attempts to protect too much and therefore can’t even distinguish between “too much” and “beyond” its scope.

Here, you make an argument is about enforceability that undermines your earlier note about "10 or 20 years". The obvious conclusion of your enforceability argument is that we should not have any form of copyright. That's fine, but it would be a tighter thread if you'd spent a few more minutes to form that argument originally.

The "collective resource" and "communal" concepts above seem to be improvised. If you want to succeed at introducing new concepts, they will need to be well-grounded.

Reply on news.ycombinator.com