Follow Slashdot blog updates by subscribing to our blog RSS feed

 



Forgot your password?
typodupeerror
×
Google Oracle Red Hat Software IBM Open Source The Courts

Red Hat and IBM Jointly File Another Amicus Brief In Google v. Oracle, Arguing APIs Are Not Copyrightable (redhat.com) 42

Monday Red Hat and IBM jointly filed their own amicus brief with the U.S. Supreme Court in the "Google vs. Oracle" case, arguing that APIs cannot be copyrighted.

"That simple, yet powerful principle has been a cornerstone of technological and economic growth for over sixty years. When published (as has been common industry practice for over three decades) or lawfully reverse engineered, they have spurred innovation through competition, increased productivity and economic efficiency, and connected the world in a way that has benefited commercial enterprises and consumers alike."

An anonymous reader quotes Red Hat's announcement of the brief: "The Federal Circuit's unduly narrow construction of 17 U.S.C. 102(b) is harmful to progress, competition, and innovation in the field of software development," Red Hat stated in the brief. "IBM and Red Hat urge the Court to reverse the decision below on the basis that 17 U.S.C. 102(b) excludes software interfaces from copyright protection...."

The lower court incorrectly extended copyright protection to software interfaces. If left uncorrected, the lower court rulings could harm software compatibility and interoperability and have a chilling effect on the innovation represented by the open source community... Red Hat's significant involvement with Java development over the last 20 years has included extensive contributions to OpenJDK, an open source implementation of the Java platform, and the development of Red Hat Middleware, a suite of Java-based middleware solutions to build, integrate, automate and deploy enterprise applications. As an open source leader, Red Hat has a stake in the consistent and correct determination of the scope of copyright protection that applies to interfaces of computer programs, including the Java platform interface at stake in this case.

Open source software development relies on the availability of and unencumbered access to software interfaces, including products that are compatible with or interoperate with other computer products, platforms, and services...

This discussion has been archived. No new comments can be posted.

Red Hat and IBM Jointly File Another Amicus Brief In Google v. Oracle, Arguing APIs Are Not Copyrightable

Comments Filter:
  • Jointly? (Score:4, Insightful)

    by javab0y ( 708376 ) on Saturday January 18, 2020 @10:20PM (#59633872)
    Isn't Red Hat the same company as IBM?
    • Re:Jointly? (Score:5, Informative)

      by Aighearach ( 97333 ) on Saturday January 18, 2020 @10:48PM (#59633914)

      Nope. IBM bought all the RedHat shares. That doesn't automatically make them a single legal entity. They could do that if they wanted, but it would be more normal to keep it as a subsidiary. It is usually only direct competitors where one succeeded and the other failed and got bought where the victor disbands the loser. Like Microchip eating Atmel and turning off their website out of spite.

      This was a friendly merger by companies with different strengths and products, but with shared long-term goals and strategies. If they want to roll it into one, it will take years to do so because they wouldn't want to cause any damage. But both brands have value, so it is unlikely.

      Just like if I owned two business that were both impacted by something, I might submit a joint amicus brief that represents both of them.

  • copyrighting APIs (Score:5, Insightful)

    by lkcl ( 517947 ) <lkcl@lkcl.net> on Saturday January 18, 2020 @10:38PM (#59633904) Homepage

    i've mentioned this before, however it's worth repeating: the implications for copyrighting of APIs goes far deeper than most people realise.

    examples include actual instruction sets. instruction sets - the actual binary bits of executables and the means and method of "interpreting" those - are an "API". the instructions *are* an "interface" (to the hardware - the actual core that executes those instructions) which is "programmable" by "applications" (aka "binaries").

    thus, llvm, gcc, microsoft's visual studio *and* sdcc *and* every single other compiler ever written would, in the eyes of US Law, overnight become criminally-infringing of every single "copyright" holder's instruction set for which they offer compiler support. Intel, ARM, MIPS, POWER, OpenRISC, SPARC - all of them - because the "instructions" (the assembly code, aka an "API") would *overnight* become copyrighted material for which NO PRIOR STATEMENT on the license of that copyrighted material had ever previously been stated.

    language translators (.NET, WASM) would similarly become criminally-infringing because every *language* is an API.

    pypy would become criminally-infringing because they implemented the python software foundation APIs *without permission that they never previously required* and on which the python software foundation never predicted that they would ever need to make a clear and explicit copyright statement on.

    any code on any git repository that implements an API (kubernetes, google cloud APIs, pyro, DCE/RPC, MSRPC, SOAP, php libraries, *EVERYTHING*) now has to make a clear statement as to how and when licensees may use that (now newly copyrighted) material.

    the list goes on and on. copyright law is very clear: if you don't make a statement (don't issue a license), then potential licencees *must* contact you each and every time to request permission.

    suddenly changing the way as to what is copyrightable to include something as massive as "APIs" where *nobody* - in the entire world - has ever issued a license or a statement about the APIs - is a total nightmare.

    the disruption that this will cause *to Oracle's own business* gives you some idea of how stupid and greedy they really are.

    • Re:copyrighting APIs (Score:5, Interesting)

      by suutar ( 1860506 ) on Saturday January 18, 2020 @11:35PM (#59633954)

      seems like it would also mean that IBM owns the copyrights to the SQL language... that would be a kick in Oracle's pants.

    • Re: (Score:3, Insightful)

      by Keill ( 920526 )
      One of the problems is that I've never really seen anyone, anywhere, use the true description for what is trying to happen: to copyright rules of language, rather than when a language is specifically applied. This is what an API is - a set of rules of how to apply such existing communication used with computers. The basic rules of the English language is an API, just for humans rather than computers, and trying to copyright that, is the equivalent of what is happening here - in addition to any other addit
      • I think this is because understanding copyright requires highly abstract thinking that most people simply are not capable of. Attempting to form an argument by extension to absurdity may not be 100% valid, but it is important to note where things could lead, albeit however unlikely that may be. I have attempted to point this out myself: https://slashdot.org/comments.... [slashdot.org] so it isn't that it goes entirely unrecognized, but rather merely seemingly insignificantly so.
    • That isn't entirely true, some of your examples implicitly or explicitly allow copying of any/all parts of the API.

      The problem here is that Google actually copied Java wholesale which was under a very murky "open" license and then relicensed it and partially rewrote it later on.

      Google is trying to spin this into "we just copied the API" but actually also copied the documentation. This wasn't a "Compaq vs IBM" clean room implementation of a black box, this was "copy, paste including the bits we don't actuall

    • That doesn't sound weird at all to me.

      Yes, you'd either need a license to target a language or for its author to explicitly publish its specifications under an open license.

      In practice, most processor manufacturers would just do the latter.

  • by Anonymous Coward

    The legal department at IBM must be in a constant state of orgasmal bliss over this one.

    If oracle does not win, it stays business as usual.

    If oracle does win, they will owe IBM royalties for SQL backdated to the mid 80s, on top of all the $150,000 copyright violation fines per oracle SQL license ever sold.
    SQL is just an API after all, one owned by IBM that oracle is trying their best to prove they have infringed.
    Oracle is setting themselves up for hundreds of trillions of dollars in legal fines by winning,

    • by jmccue ( 834797 )

      If oracle does win, they will owe IBM royalties for SQL

      I would not consider SQL an API, unless you are taking about how it accesses Rational Databases, which I think IBM also came up with.

      I expect a very narrow ruling if Oracle wins, but if not good time to become a Corp. Lawyer

      • by Anonymous Coward

        I would not consider SQL an API, unless you are taking about how it accesses Rational Databases, which I think IBM also came up with.

        For what it's worth, the wikipedia page for APIs considers it an API, along with accessing any database that is separated from the application.

        They list databases specifically, but SQL also falls under being a communication protocol, a data structure abstraction, and a client/server model.
        They also consider html, soap, rest, json, and a web server backend APIs, plus list ODBC explicitly.

        Then there is lower down the OSI stack to consider, which I didn't in my previous post.

        TCP, named pipes, IPC, all are APIs

  • by Sebby ( 238625 ) on Saturday January 18, 2020 @10:49PM (#59633916)
    Goes something like this:

    "Oracle is a law firm with a software side business."

    (Wish I could find the original quote so that I could properly attribute it).

  • This one is definitely headed the Supreme Court's way... both sides have plenty of money for lawyers, and there's significant interest on both sides as well.

  • It would be terrible if the Java API were copyrighted. IBM and Redhat are arguing that the harm done to society outweighs the benefit of allowing a collection of APIs to be copyrighted. Note: an individual API isn't copyrightable, like a recipe or a dictionary entry most countries require a copyright to be more creative than that. However an entire collection of recipes, or a dictionary or a set of APIs is a creative work and therefore worthy of copyright protection. I hope Oracle loses but the law as i
    • by suutar ( 1860506 )

      A recipe (specifically an ingredients list and a minimal list of operations) is not copyrightable, but a book of them is, true, because someone decided which recipes to put in, added descriptions, pictures, more depth to the operations, etc. However there's nothing stopping you from making your own book with some or all of those recipes as long as you don't copy the creative parts. "pizza dough - 1/2 cup milk, 1.5 cup flour, 1 packet yeast, 4 tablespoons butter. Melt butter, add milk, warm milk/butter mixtu

      • Google did both though. They copied the API as well as its composition within their own product (Android), modified it and called it their own while the license, although open, doesn't allow them to just go about and modify the Java API

        Microsoft did the same thing to Sun in the 90s (there was a thing called MS Java) and lost to Sun because they didn't conform to the "open" license. It was then decided the only way Java could be reimplemented is by a cleanroom implementation that isn't associated with the Ja

        • > It was then decided the only way Java could be reimplemented is by a cleanroom implementation that isn't associated with the Java brand name.

          Is Google advertising anything associated with the Java brand name? Pretty sure they called "their" language something completely different, which frees them from trademark law. And Oracle has no copyright claims (other than trying for the API), because the implementation was copied from the firmly open Apache Harmony re-implementation.

      • Tell that to the tetris company

  • SCO (Score:4, Insightful)

    by timeOday ( 582209 ) on Saturday January 18, 2020 @11:26PM (#59633940)
    Did 13 years of SCO suing IBM over Unix APIs teach us nothing?

    (Also, why am I the first person who referenced this? What has ./ come to?)

    https://en.wikipedia.org/wiki/... [wikipedia.org]

  • by Todd Knarr ( 15451 ) on Saturday January 18, 2020 @11:31PM (#59633948) Homepage

    This one ought to be brought up during the case. Oracle themselves do exactly what they are claiming is copyright infringement: they created their own server implementation of AWS's S3 APIs for the express purpose of allowing customers with tools that worked with S3 to work with Oracle's Object Storage product without any code changes. That's exactly what Oracle's claiming Google did, create an implementation of an Oracle API that was compatible with that API. Let Oracle explain to the judges why it's OK for them to do it and not Google.

    • With the amount of lawyers involved in this case I gotta imagine that will be one of the arguments they put forward.
    • "An Oracle spokeswoman said that the S3 API was licensed under an Apache 2.0 license. She pointed us to the Amazon SDK for Java, which does indeed come with an Apache 2.0 license."

      *shrug*

      • That would be a problem for Oracle, though. OpenJDK, which is the open-source standard for the Java APIs, is licensed under the GPL. That would make Google's implementation eminently legal. Except for the small fact (which still hurts Oracle) that the SDK isn't the API and certainly isn't the server-side part of S3 which is what Oracle implemented. At most, having Amazon's SDK licensed under Apache 2.0 would grant Oracle the right to develop their own SDK to interact with Amazon's S3 servers. It wouldn't gr

      • Thing is, when Google did it, it was legal because Sun openly encouraged companies to use Java. Oracle's argument is that the owner of the API can choose to revoke permission to use an API which they'd previously allowed. So it really doesn't matter that Oracle has licensed the S3 API. If they win, Amazon can simply say they're revoking the permission granted by that license, and Oracle will need to pay them a gazillion dollars if they wish to continue using it.
  • by jonwil ( 467024 ) on Sunday January 19, 2020 @12:30AM (#59633994)

    I wonder if IBM could claim (under "API copyright" as Oracle is defining it) that Oracle violated IBM copyright when it copied IBM's Structured Query Language to produce the first version of the Oracle database management system. (I doubt IBM gave Oracle explicit permission to use SQL back then)

  • IMHO, it can have potentially far reaching implications if APIs are protected. Applications use system APIs to run on top of Operating Systems.

    It will be possible to lock specific applications and prevent them from running...

    Oracle has traditionally never supported open systems.

Love may laugh at locksmiths, but he has a profound respect for money bags. -- Sidney Paternoster, "The Folly of the Wise"

Working...