Earth to Google: Here’s Why APIs Need to be Copyrightable

“Essentially, Google wants to drive on Oracle’s eight-lane, super-expressway instead of utilizing the slower side roads—and it wants to do so without compensating Oracle, the creator and maintainer of the expressway, for the privilege.” January 6, 2020, Google submitted its brief in Google v. Oracle, kicking off the Supreme Court case that many are calling the “copyright case of the decade.” The suit pits the search engine platform controlling 93% of the worldwide search market against Oracle, the owner of the ubiquitous Java program, which submitted its response brief last week. After attempting and failing to secure the rights to Java, Google decided to cease negotiating and instead replicated 37 API packages from the copyrighted program, a decision that precipitated the years-long lawsuit.

Making Google Hum

An API, or application programming interface, is what allows different parts of a computer program to communicate with one another seamlessly. APIs are like the turbocharger in a car engine. They speed up the process of retrieving Google’s search results and enhance the user’s experience. To users of a search engine, speed and accuracy of search results are the most important factors affecting their choice. Oracle’s APIs essentially make Google’s search engine hum.

Essentially, Google wants to drive on Oracle’s eight-lane, super-expressway instead of utilizing the slower side roads—and it wants to do so without compensating Oracle, the creator and maintainer of the expressway, for the privilege.

Google claims that, because of the “fair use” doctrine, it shouldn’t have to — APIs shouldn’t be copyrightable, and requiring programmers to pay for the APIs they use would “harm innovation.”  This argument is like stating Uber users should not have to pay for their rides because if Uber were free, people would travel more.

Currently, fair use only permits replication of a copyrighted work if the result is transformative in nature—in other words, the original work has been changed so drastically that it no longer functions as the original creator intended. But Google’s verbatim copy clearly wasn’t transformative at all.

To make up for this fact, Google argues that APIs aren’t individual, creative expressions, but are instead more akin to essential roads one has no choice but to take to reach a destination and thus cannot be subjected to federal copyright law. The Federal Circuit disagreed, however; hence, the appeal to the Supreme Court.

Federal Circuit: Hit the Road, Google

In its ruling, the court found that there were “unlimited” creative ways the code could have been constructed to perform the same function. For Google, the APIs it took from Oracle weren’t the only highway into town. Rather, they were individualized creations, unique to Java—precisely the sort of thing that digital copyright is designed to protect. As such, the Federal Circuit ruled that Google cannot drive its Rolls Royce for free on Oracle’s highway and must either abide by its licensing terms or take the side roads instead.

There are plenty of examples that undermine Google’s claim that APIs are meant to be open source. In the airline industry, for example, airlines utilize proprietary APIs to allow their reservation systems to function. These companies have sometimes permitted chosen third-party agencies to copy and reuse their APIs so the airlines’ systems can operate more efficiently, just as Oracle has always allowed app producers and non-competing platform developers to utilize its APIs without a license. But at the end of the day, the creators of these API get to decide who gets access (and who does not).

SCOTUS Must Not Kill the Drive to Innovate

Ultimately, copyrights are essential; without them, programmers would lack the drive to innovate. If they are neutered, companies that rely upon these copyright protections, either for competitive or security purposes, would no longer have the protections they need to justify spending on innovation. Without copyright to protect their proprietary APIs from poachers, software companies would essentially be forced to give away their work. In the case of Google v. Oracle, Oracle would be giving its proprietary APIs to a company eight times larger—one that can afford to pay a reasonable price for it.

Any programmer who dreams of working for themselves (or a smaller company) to create something instead of being a cog in a trillion-dollar monopoly machine needs to sit up and take notice. A ruling for Google could expand the exceptions to copyright beyond all reasonable measures, swallowing the rules of the road that have made the United States one of the most innovative countries on Earth.

Image Source: Deposit Photos
Image ID: 6575394
Copyright: Krisdog 


Warning & Disclaimer: The pages, articles and comments on do not constitute legal advice, nor do they create any attorney-client relationship. The articles published express the personal opinion and views of the author as of the time of publication and should not be attributed to the author’s employer, clients or the sponsors of Read more.

Join the Discussion

39 comments so far.

  • [Avatar for angry dude]
    angry dude
    March 1, 2020 12:18 am

    “To promote the progress of science and useful arts, by securing for limited times to authors and inventors the exclusive right to their respective writings and discoveries.”

    Do API function declarations qualify as “writings” ? Sure they do !

    SCOTUS is about to trash The Constitution of the United States of America once more !

    Only now it’s copyright’s turn (They trashed inventors and their patents back in 2006 with Ebay decision)

    Where is Special Prosecutor ???

  • [Avatar for angry dude]
    angry dude
    February 29, 2020 02:23 pm

    Markus Mobius @37

    “If it it would be all so obvious then the Supreme Court would not have taken this case.”

    Sure, sure…

    Senile scotus justices with liberal arts degrees from like 50 years ago and their absolutely horrible Constitution-breaking track record sure know better what API function declarations are and what protections they have to have…

    Comments like this make me very angry

  • [Avatar for Markus Mobius]
    Markus Mobius
    February 28, 2020 11:26 pm

    @angry due

    Well, we’ll see whether API function declarations have the same level of protection as implementation code. If it it would be all so obvious then the Supreme Court would not have taken this case.

  • [Avatar for angry dude]
    angry dude
    February 25, 2020 07:42 pm

    Markus Mobius @33

    “it’s just that nobody cared to sue and abandon this implicit social contract … until Oracle came along”


    If you or someone else copies 600 pages of my code (API function declarations are code) without my permission I will be very angry
    If that someone is a trillion dollar corporation I will be not just angry but mad

    “implicit social contract” is something like not missing a urinal when using public restroom
    Got it ?

  • [Avatar for angry dude]
    angry dude
    February 24, 2020 04:46 pm

    Tiburon @34


    We are talking copyrights here, not patents

    Are you ok with trillion dollar company copying 600 pages of code (API function declarations are code) to save on programmers cost ?

  • [Avatar for Tiburon]
    February 24, 2020 04:14 pm

    No, we cannot allow copyrighting of API’s – it will strangle and kill small software startups.

    There are literally thousands of small software startups that are thriving in today’s world without relying on patent protection. For example, look at and the source code to this very web page and you’ll see some components used to construct it:

    WordPress: 0patents
    jquery: 0patents
    OpenTracker: 0patents
    FormSwift: 0patents
    Feedburner: 0patents

    These startups (and future ones like them) must be allowed to survive and thrive.

  • [Avatar for Markus Mobius]
    Markus Mobius
    February 23, 2020 11:45 am

    Oracle’s law suit was a patent case with some copyright claims thrown in – when it started back in 2010 Florian Mueller, for example, viewed the copyright claims secondary. They only became primary after Oracle lost all patent claims.

    Oracle’s APIs have nothing to do with Google’s search engine – it’s about Android and reducing the switching cost of programmers as Kyle says. Companies create APIs in order to allow external developers to write code for a platform or system. Importantly, they do not own this third-party code. However, it is extremely expensive to rewrite third-party code for a different API – even for the original third-party developer. It is for that reason that all competitor’s to Google AdSense mimic Google’s API for submitting campaigns or that Amazon’s and the upcoming Huawei app store mimic the Google Play APIs: otherwise it would be extremely expensive for developers to submit their Google Play games etc. to the Amazon or Huawei stores.

    In other words, if APIs are copyrighted and not covered by fair use then the platform owner would have huge power over the code of all third party developers: they could always change the terms of the primary platform to extract more rents from existing third party code.

    For this reason, it has long been accepted among programmers that re-implementing a runtime (such as new Java runtime) is not violating copyright. The SQL language, the C standard library, the Unix Posix calls etc. are all examples of APIs that have been widely adopted and evolved in subsequent system (even the Java library has many similarities to the C standard library). This all happened dispite the fact that these APIs had no explicit permissive license – it’s just that nobody cared to sue and abandon this implicit social contract … until Oracle came along.

    Despite its importance, this case won’t have a significant long-time effect, however. Recent successful APIs (Kubernetes, Docker etc.) are all explicitly licensed with permissive licenses and I doubt that after this law suit will adopt an API without it. Hence, while this case is very significant for Google and Oracle it will not change the status quo of having APIs that can be modified by third parties.

  • [Avatar for Anon]
    February 20, 2020 10:02 am

    I find it interesting that in the back and forth here of IP practitioners (who often HAVE the technical chops) and the technical practitioners (who often lack the legal chops) that there is a massive glossing over of the plain fact that an item may well have aspects that may properly obtain different legal IP protections under the different legal IP realms.

    There is no such thing as an “either/or” patent or copyright question.

    This is NOT to say that there are no nuances such as expressions that become absorbed into functionality (copyright law is more nuanced in that regard).

    This IS also to say that the irony is NOT lost on those of us in the legal community to hear from technical practitioners trying to deny expressive protection based on functionality, when (and admittedly perhaps not from the same exact people – but indeed from the same exact same philosophical camps) ANY type of patent protection for software has been wholesale criticized as somehow inherently improper.

    Here on this thread, there are those that use analogies, then “smell a Tr0ll” when those same analogies are shown to be improper. Here’s a hint for talking with legal folk who are NOT f00led with such name-calling: don’t attempt the improper analogies in the first place and don’t get upset when the limits of analogies are put on the table. Your “opponent” is NOT “tr011ing” you in setting you straight, and the reflexive name-calling makes YOU appear to be the one “tr011ing” the topic as you appear to NOT be grasping all the issues (technical AND legal) that are in play.

  • [Avatar for Night Writer]
    Night Writer
    February 20, 2020 09:30 am

    @29 Kris Fox

    Not really accurate. The analogy might work if you included hose sizes and functional aspects of the turbocharger. There is some functional organization based on an API.

  • [Avatar for angry dude]
    angry dude
    February 19, 2020 11:49 pm

    Ternary @28

    So they are done with patents and moved on to destroy copyrights ?

    If Google is allowed to steal 600 pages of someone’s creative work from another megacorp then what about the rest of us, bovines ?

  • [Avatar for Kris Fox]
    Kris Fox
    February 19, 2020 11:15 pm

    An API is not like the turbocharger on an engine. It is merely the measurements of the bolt holes and orifice size of a turbocharger. Google didn’t copy Oracle’s turbocharger, they just made a totally new one with the same bolt holes and orifice size, which could be used in place of Oracle’s.

  • [Avatar for Ternary]
    February 19, 2020 08:35 pm

    Another Latin saying that applies here:
    Quod Licet Iovi, Non Licet Bovi.

    Gods (Jupiter) may do what bovines may not.

    No doubt here who the God is and who the bovines in these IP stories. My feeling is that also in this case the decision will be in favor of G. Probably with some convoluted explanation why APIs are “merely” functional statements.

    I realize I don’t have to feel sorry for Oracle. Larry Ellison will still be able to buy some mega-yachts. But it is discouraging to realize that if Oracle cannot win, the chances for the bovine are really non-existing.

  • [Avatar for angry dude]
    angry dude
    February 19, 2020 05:39 pm

    Bill @19

    “API is a set of strings and expected parameters”


    That’s what dictionary is – strings (words or sentences) and expected parameters (words and sentences explaining the meaning in same or foreign language)

    I suggest you copy Merriam-Webster or some other dictionary and put it online under your own name
    See what happens next (You WILL be sued for copyright violation)

    Same with Google… only Google is trillion dollar monopoly .. which needs to be broken up into small pieces asap but won’t be of course… because
    “Pecunia non olet”

  • [Avatar for Ternary]
    February 19, 2020 05:18 pm

    In its most basic form this case deals with copying of 11,330 lines of code representing 600 pages of text as source code. The above discussion seems to focus on what an API is. But the functional aspects should be irrelevant. 600 pages of text was copied by Google. The defense is that rewriting these lines (which everyone seem to agree to, is possible, but involves time, manpower and perhaps loss of adaptation) would be prohibitive compared to other source code that was modified in creating Android.

    That seems strange, a company like Google stating that it was prohibitive to create new code. Perhaps the “technical weaklings” at Google indeed were unable to create something new, which seems unlikely. However, defending their direct copying of 600 pages of source code as “fair use,” seems a bit much.

    There is quite some work involved in naming classes, methods and their declarations. As such, I would consider that creative work. Google did not have to do that work because it literally copied it. Sounds like copyright infringement to me.

  • [Avatar for angry dude]
    angry dude
    February 19, 2020 05:13 pm

    Pro Say @24

    “All your books belong to us”

    …. (some 25 years later)

    “All your foods belong to us”

    – Amazon

    Congress, break them up already, pleeeeaase !!!

    They won’t, of course.. because .. “pecunia non olet”

    Well, “the doc said ‘to the morgue” – to the morgue it is!”

  • [Avatar for Pro Say]
    Pro Say
    February 19, 2020 03:23 pm

    “All your copyrights are belong to us.”

    “All your trademarks are belong to us.”

    “All your patents are belong to us.”

    — Google

  • [Avatar for Anon]
    February 19, 2020 03:05 pm


    I congratulate you on your last post in which you pulled back (even if but a little) on the emotion and expressed reason (to which I have previously noted as the factual nature that the various ‘wares’ in the computing arts [software, firmware, hardware] are but patentably equivalent design choices for innovation in the computing arts.

    I am looking forward to an equally non-emotive (non-philosophical) FACTUAL acknowledgment from our ‘software friends’ so that the discussion of various forms of innovation protection (copyright for aspects of expression, patents for aspects of utility) can proceed on a common ground of understanding.

  • [Avatar for Bill]
    February 19, 2020 02:53 pm

    I smell a troll. It was an analogy, and trolling 201 is taking analogies too literally to control the flow of conversation (the equivalent statement, Angry, if you want to argue against Oracle next time, would be of course API’s aren’t highways! I definitely wouldn’t recommend driving on them). Well, this is probably not the place I’m going to get an answer to why people think APIs should be copyrighted. Taking Kyle’s approach.

  • [Avatar for angry dude]
    angry dude
    February 19, 2020 02:19 pm

    Bill @13

    “Programs are the books”


    You are actually insulting all hardcore software engineers by this sentence

    My programs are not books – they are precisely engineered mechanisms to take physical inputs (e.g. audio after ADC) and to provide logical and physical (after DAC ) outputs

    Stuff like that used to be done in assembly for each particular processor but nowadays its mostly C/C++ with some stuff in Java or even Python (with e.g. numpy or jit compiler or cython)
    Calling those highly technical software programs for e.g. audio/video processing “books” is just nonsense
    They are digital engineering implementations of what used to be purely analog domain (opamps, transistors, etc) some 30-40 years ago – signal and image processing

    I’m not even sure you understand what I’m talking about – keep reading techdirt and other crap like that

  • [Avatar for Jam]
    February 19, 2020 01:53 pm

    What does “use an API” mean?

    Does it mean a third party developer can write a program that accesses a first system exposing the API?

    Does it mean a third party developer can copy the source files that define the API to make a competing system exposing the same API so that subsequent third party developers can use the competing system instead of the first system?

    It seems that most engineers see red and become irrational when contemplating the first definition and fail to consider the second definition.

    Here it seems that Google copied source code owned by Oracle to create a competitive system and is trying to argue that plagiarism of copyrighted work is okay, because the users of the competitive system should be free to use the original API on the competitive system. Imagine copying 37 chapters of Harry Potter in your novel “to avoid the ‘drudgery’ of recreating the code the backstory”. [see 1]


  • [Avatar for Bill]
    February 19, 2020 01:41 pm

    *sigh* mixed up Kyle’s comment and Night’s. Either way, people here seem to know an API is a set of strings and expected parameters that another application can use to interact with the underlying implementation. Would someone mind going over why they think that IS copyrightable vs. why it’s not not copyrightable? None of the author’s analogies are correct, and they’re all extremely biased toward Google’s use of the API to cheat and speed (so application of circular logic in analogy formation) vs. using an industry standard and making a smart technological choice, so I’m wondering if someone here can do a better job. Thanks in advance.

  • [Avatar for Bill]
    February 19, 2020 01:21 pm

    Angry, I’m not talking about my convenience, I’m talking about what happens in an unconstrained copyright rule applied to APIs. Please try to counter my example of a publication of most combinations of words and parameter types without an implementation behind them. It would constitute an API. It would be useless, but copyright doesn’t require something be useful. I could then start a program to find whenever someone used those random combination of words and parameters on GitHub in something and sue them? If the ruling is constrained to using the same API with the same implementation, I’m on board actually (I think I agree with you more than you think). The problem is that names don’t really represent the hard work of programming.

    I also agree with Kyle on most points. Sorry if you felt that comment was directed at you, Kyle. It was not. Your discussion of unimplemented APIs directly shows you know what an API is. It’s just that confusing APIs and implementations is a recipe for disaster, which I’ve seen happen a lot in this case. I also liked BP’s point (which is also why I said I wasn’t commenting on Google’s infringement of Oracle IP, but the problems with copyrighting an API).

    To angry, A protocol is a series of API calls. If I copyright those APIs, I’ve made the protocol owe me money. Yes, it’s a problem for protocols if the SCOTUS makes a broad claim about APIs vs. Java APIs in this one case of Google using it and going against Java’s standard JVM. Any single company’s claim to the protocols in HTTPS is probably weak? So I doubt there are any real issues to HTTPS right now, but think about trying to develop that standard in a scenario where the APIs to develop it can’t be used because they’re copyrighted. That’s not programmer convenience, that’s consumer safety that’s impacted.

  • [Avatar for Night Writer]
    Night Writer
    February 19, 2020 12:38 pm

    Maybe Scotus will hold the API as functional and not eligible for copyrighting.

    (And yes I know what an API is. I have an advanced degree in computer science and was a developer before becoming an attorney.)

  • [Avatar for Anon]
    February 19, 2020 12:32 pm


    While I do not agree with your views, I urge you to reconsider and not let a “heckler’s veto” dissuade you from discussing the subject.

    Trust me, “angry” does NOT speak for all that may disagree with your views.

  • [Avatar for Kyle E. Mitchell]
    Kyle E. Mitchell
    February 19, 2020 12:19 pm

    Angry Dude @ 11: Raising ageism non sequitur, then putting me down ad hominem as “kiddo” from behind a pseudonym, is not respectful addition to this discussion. I continue to sympathize for the anger you self-consciously feel. But if you’d like further attention and consideration from others like me, you will need to uphold a higher level of respect, and stay on topic.

    My cost-benefit on comments here has tipped. Those interested in following up on the case can reach me directly.

  • [Avatar for angry dude]
    angry dude
    February 19, 2020 11:40 am

    Bill @10


    you convenience as a programmer using familiar API comes last to google’s ability to monopolize markets for profits. Period.

    Your talk is cheap crap: if large industry participants want single common “free as in beer” API they found a consortium to create one, sign some legal agreements including license and copyright attached

    OpenGL and OpenCL are far better example then your HHTPS which is just a protocol
    These are open alternatives to proprietory APIs namely DirecX and Cuda -which are 100% controlled by respective manufucturers – Misrosoft and NVidia
    Everybody can do pretty much whatever they want with OpenGL or OpenCL, but not with proprietary APIs like Cuda and DirectX – you will be sued immediately if you violate those license agreements by e.g. creating and marketing your own extension

    Lats but not least, scotus and other us courts ignorance in basic science, technology and patent law does not preclude them from messing up the foundations of 230-years old US Patent system turning it upside down
    Likewise your own ignorance in copyright and patent issues does not preclude you from posting here
    Perhaps you should follow your own advice and shut up ?

  • [Avatar for Bill]
    February 19, 2020 10:24 am

    I actually agree that implementation should be protected, let’s say that protection is copyright. I actually think not changing the underlying code to the API is worse than using the API and changing how it works to fit a particular need while maintaining compatibility. I see most pro-Oracle people out there saying that programmers just like free and open things. While this is true, because there is utility in that, I’m not arguing for that here. If you want to close-source and commercialize your programming language, cool! You can do that, just like MATLAB (although be prepared to suffer the losses R and Python have caused it, and to a much much much lesser extent Octave). That should be within every company’s right. I also think the hard work of developers should be protected (since I am one, and I like making money). It’s not hard work to call a function that adds two numbers “add” though. It’s not hard work to call an object that works like an integer “Integer”. Copyrighting APIs is copyrighting those words. It’s not protecting the code underneath, it’s preventing anyone from making anything useful because unlike patents, copyrights don’t need to be new, novel, or useful. This ruling could actually make a business model out of copyrighting basic words! When this ruling goes into effect, I could just make a ton of unimplemented code (say, by writing a program to string together words and argument types), copyright it, and prevent anyone from doing work without paying me (again, because copyrights don’t need to have utility, function, or social value).

    Copyrighting API’s is like copyrighting the English language. It’s absurd. Maybe French is more applicable since there’s actually an organization that manages it. Programs are the books, APIs are the words / language.

  • [Avatar for Anon]
    February 19, 2020 09:39 am

    Without diving into the details and deconstructing the “wants” masquerading as “musts,” what we have here is simply a philosophical battle of convenience versus protections of expression (copyright) and utility (patent).

    The parade of horribles is nothing more than hyperbolic mantra of “free-riders” (who are no doubt genuine in their philosophical beliefs).

    The problem though is NOT as Bill would phrase it. The problem is that followIng Bill’s desired philosophy forces that philosophy on others in contradiction to actual protections under the law for either AND both protections of expression and utility.

    Bill’s mantra equates to a hyper application of the worst of Communism and Fascism, obliterating the initial work of others.

    In the patent world, it would be like providing that ANY improvement (at all) to any item still under patent protection REMOVES and and all such protection, and labeling this theft as ‘promoting innovation.’

    While I certainly “get” that there are differences in the realms of protecting expression and protecting utility, the underlying “gist” of the “but software wants to be free” philosophy remains at its core the negation of any sense of property right to what went before because of mere ‘want’ and ‘convenience’ to use those items of what went before with NO paying.

    Instead, and in other words, this THEFT is passed off as ‘professionalism.’

    Of course, Big Tech (and it’s philosophical rags such as TechDirt, SlashDot and the like) are ALL FOR fostering this type of Lemming belief system, as they are direct beneficiaries of such ‘professionalism,’ and would MUCH rather compete on other factors such as size and established presence.

  • [Avatar for angry dude]
    angry dude
    February 19, 2020 06:36 am

    Kyle E. Mitchell @8

    “I believed the topic of programmer switching cost relevant because I see avoiding that switching cost as a substantial motivation of Google’s choice for Android”


    Programmer’s convenience comes last to google’s ability to make money via monopolizing markets

    All “free” goodies from google e.g. “free” gmail or “free” Android API come at a price – to you as loss of privacy (each and every word in your gmail is scanned and analyzed by google) or to someone else like Oracle in this case

    So Google IS EVIL trillion dollar multinational corporation without allegiance to the USA and should be broken up into pieces
    same with Apple and Amazon
    The sooner it happens the better for all of us
    But unfortunately it won’t happen any time soon
    because “Pecunia non olet”

  • [Avatar for Bill]
    February 19, 2020 03:19 am

    Disclaimer: I’m a software engineer. I have worked with Oracle products, multiple cloud offerings, and mostly work with FOSS. Here’s the problem: non-software engineers don’t actually know what an API is. It’s NOT the actual code underneath the hood (Google actually changed some parts from Java’s default in part to improve performance on ARM processors in a smart phone). An API is what the developers decided to name that code, and the input parameters they decided that code would take to turn it into output. It’s NOT an eight lane highway. It’s the word, “highway.” Copyrighting an API is like copyrighting a very useful word and then expecting people to pay you to use it. It’s not a sentence, paragraph, or book. These are the words of programming. Allowing them to be copyrighted will destroy American innovation because it will be extremely difficult to integrate solutions or to switch to a new solution when somebody (like Oracle) starts to provide less value per dollar and you want to switch to the latest, greatest, and most cost effective (AWS, Azure, GCP). Several companies will sink nearly overnight if this ruling goes into effect. Another way to describe an API is a contract. It tells one system what to expect when using it. When I’m speaking English, other English speakers know the words I’m using and the grammatical structure. If I started speaking Japanese, that all changes. Copyrighting an API would result in an explosion of new “languages” a programmer would have to learn. You know HTTPS? That could be copyrighted. No more browser security because no programmer would be able to know the 7 million new forms of sending internet packets that would crop up. OSes use CPU instruction sets, which are like APIs for hardware. If copyrighted, OSes would now require entirely different versions for every single CPU maker (they already do require different drivers and compatibility, but we’re talking changing the foundation of a house here). Software has decided that sharing these names is good, because it enables the same thing that human languages do: effective communication between large numbers of components (or people for human languages). Oracle poses a serious threat to everyone except Oracle and should be ashamed of what they’re doing. It’s downright dangerous and has implications as far reaching as national security. This is why some patent lawyers (pharma and chemistry come to mind) have PhDs and JDs. Knowing the law isn’t enough to know how to apply it to extremely technical fields. Case in point: how many lawyers would know that Vick’s nasal spray contains methamphetamine? Well, it’s L-methamphetamine, not D-methamphetamine, its controlled chiral cousin. If you don’t know what chirality is, you shouldn’t be able to handle, as a judge or lawyer, any suits involving chemistry. If you don’t know what an API is, you shouldn’t be able to handle Oracle v. Google.

    I am not taking a position on the other complaints from Oracle in this suit (I think there are two others, maybe more). Google may actually have violated Oracle’s IP. I’m exclusively commenting on why APIs being copyrightable is a horrible, dangerous idea.

  • [Avatar for Kyle E. Mitchell]
    Kyle E. Mitchell
    February 18, 2020 10:13 pm

    BP @ 6: I think we’d tend to agree about Google on larger points. But I’m not interested in Google v. Oracle either on account of Google or on account of Oracle. The industry is watching for the broader effect on copyright and software production. I trust the court will weigh the copyright issue, rather than the hearts and karmic balances of the parties overall.

    As for the threat that programmers feel from the idea of API copyrightability, or API copyrightability without a sufficiently mitigating fair use defense, it’s difficult for me to find good legal metaphors, despite combining both backgrounds. I think that’s partly because law practice—or at least the transactional law practice I know—largely skirts for itself the kinds of copyright issues that coders fear for their craft. There’s at least one infringement case of copyright in a contract form on record. There’s no reason copyright can’t apply to our work product. But for the most part drafters riff, crib, and flat-out copy with impunity.

    The closest comparison I’m left with is case law. Wexis publishes a lot of case law. Much of that case law falls outside copyright under 17 USC 105. But to cite that case law in most courts, you need pagination from official reporters, often commercial products. Commercial reporters come peppered with annotations, summaries, and other material covered by copyright, salting the whole. Whether you want commercial annotations or not, officially paginated case reports aren’t widely available for free, and in fact West sued Matthew Bender, a legal CDROM publisher, for trying its own hand. West lost, but in many ways is still winning, practically, because you can’t just copy reporters. You have to copy reporters and pick out every grain of salt.

    The details are different, but the concerns about API copyright run parallel. Programmers that object to API copyright in something like the Java standard library see Oracle’s move as using the relatively small, if admittedly creative value of the API design to salt the altogether more important value of those APIs as functional machinery and a standard.

    The history of the Sun-Google license negotiations that you cited might very well come in on the purpose-character prong of the fair use analysis. But the purpose and character of the use is as commercial as can be with or without those facts.

    If I had to bet on the case, I’d probably bet on Oracle. And if they give Google a reprieve, I’d be very surprised to see a new rule clear enough to afford repeatable relief to many other defendants. But even so, I don’t think a new rule or exception for predominantly functional works would push the Supremes off the deep end, to wreck copyright as a whole.

    From the industry side, I’m not aware of any sizable market in API designs independent from software as yet. I suspect more than a few deals have been done in the shadow of the ongoing litigation. It may become more common in the future. But as yet, APIs without accompanying implementations haven’t been in brisk trade, at least in the corners of the industry that keep me busy.

  • [Avatar for Kyle E. Mitchell]
    Kyle E. Mitchell
    February 18, 2020 09:39 pm

    “Angry Dude” @5: I am sorry that you are so angry. And I am sorry that ageism comes so readily to mind for you, perhaps due to personal experience. That top[ic certainly did _not_ come to my mind as I wrote my comment above.

    For what it’s worth, I am well into my thirties, but picked up Go, Rust, and the C99 additions to C89 in the past couple of years. I pay the switching cost of maintaining projects in Ruby, JavaScript, Go, or Rust code nearly every day, so it’s a kind of pain I know well.

    I believed the topic of programmer switching cost relevant because I see avoiding that switching cost as a substantial motivation of Google’s choice for Android.

  • [Avatar for Kyle E. Mitchell]
    Kyle E. Mitchell
    February 18, 2020 09:33 pm

    Anonymous New Law Grad @4: My argument was predicated on no such generalist aphorism. Your words, not mine. I rather intended to make things more concrete, by focusing on an area where the article seemed to mishandle relevant facts and circumstances.

    If I shared the one-sided formulation of the situation you expressed, I’d wonder why the Supreme Court granted cert at all. There’s a fascinating and important issue on the boundary of copyright here, which becomes apparent only in light of the facts. We have an answer to the question as it intersects the threshold copyrightability question. The court may or may not evolve fair use doctrine responsive to factors that analysis elides or downplays.

  • [Avatar for BP]
    February 18, 2020 08:57 pm

    Kyle E. Mitchell @1

    Ruling, para 34: ” The point of contention between the parties was Google’s refusal to make the implementation of its programs compatible with the Java virtual machine or interoperable with other Java programs. Because Sun/Oracle found that position to be anathema to the “write once, run anywhere” philosophy, it did not grant Google a license to use the Java API packages.”

    I can’t really figure out Kyle E. Mitchell @1’s point: That being said, a great many programmers, especially Java programmers, see Oracle’s claim standing for a much broader threat to their professional autonomy, not to mention the substantial intellectual investments they’ve made in becoming “Java programmers”. When _they_ reach for punch, they often hyperbolize that strong copyright in software APIs would proprietize and balkanze the very language of computer programming.

    Google stands as a threat to the autonomy of many. As DHH, developer of Ruby on Rails stated: it’s not a “search engine”, it’s an “ad engine”. Playing close to the edge with IP rights, that’s what many monopolies do. How’s search doing on respecting trademark rights? Hmm.

    Let’s hope the result is not to the detriment of others, not worth tearing down the copyright system just so a big monopoly that did not really “develop” something can win/have its way.

  • [Avatar for angry dude]
    angry dude
    February 18, 2020 04:49 pm

    Kyle E. Mitchell @1

    “Hope this is helpful”


    This is NOT helpful at all !

    You are either a programmer or not

    If you are a PROGRAMMER then you’ve probably programmed in a few languages, maybe a dozen…

    I remember myself writing programs in Fortran and even Algol, who cares about those now ?

    As you age it becomes more difficult to learn and memorize new computer languages and APIs , same with human languages btw

    But how is this relevant here in this discussion ?

    You learn and master languages, human or computer, at young age. Period.

    You wanna start some age discrimination discussion here ?
    I highly do not recommend it

    P.S. The ramp up time for a young capable programmer to completely master new computer language is between few weeks and no more than 3 months
    Unless its some very short-term project the points you bring are IRRELEVANT

  • [Avatar for New Lawschool Grad]
    New Lawschool Grad
    February 18, 2020 03:27 pm

    Kyle, your entire line of thought is predicated on the insidious and faulty idea that life is best when it is “fast, good AND cheap”. As the old saying goes, you can only ever have 2 out of 3 of those things.

    In Google’s case, they released the Android platform to the (developer) market and said, “Look – we did this Fast (spent no time developing our own APIs), it is Good (i.e. we are leveraging proven and familiar APIs), and it is Cheap (i.e. we paid nothing for these proven APIs)”!

    Only in reality – it was not cheap at all. Google simply stuck Oracle with the bill – and now Oracle has paid dearly.

  • [Avatar for anonymous]
    February 18, 2020 03:15 pm

    Google needed to get Android out quickly, so it cheated and copied code. It felt it had no time to innovate, so Google plagiarized. Then Google generated over $42B on Android. Efficient Infringement, indeed.

    It is no wonder Amazon, Apple and others have adopted willful infringement as part of their business model.

    Is there anyone in Congress willing to step up against this unfair competition?

  • [Avatar for angry dude]
    angry dude
    February 18, 2020 01:42 pm

    in other news… bezos is saving the planet.. no less

  • [Avatar for Kyle E. Mitchell]
    Kyle E. Mitchell
    February 18, 2020 01:27 pm

    I’d encourage the author to revisit the first paragraph of “Making Google Hum”. It is _not_ the case that every lawyer (or judge) ruling on the API copyright issue has to understand all the technical aspects as engineers do. But I fear the picture painted sacrifices a lot of veracity for a little punch.

    I don’t recall the suit implicating Google’s search engine or other online services. Rather, the suit concerns Google’s use of certain Java application programming interfaces in its Android mobile phone software. The two relate, but on a much higher plane than copyright and fair use. Availability of Android phones, and Android integration with Google services, helps increase use of those services. But the implication that Google’s crown jewel, search, was had by infringement of Java APIs, misleads.

    The cliche metaphors for APIs properly emphasize usability, interoperability, and standardization. Familiar APIs work akin to standardized plugs and sockets, or the standard layout of a computer keyboard, or at a stretch, to a spoken language. Programming languages help us instruct computers to perform mundane tasks, like alphabetizing a list of names, and to combine those mundane tasks into less mundane solutions, like Android, Google Search, or the software that runs this blog. To be complete, programming languages must offer both a grammar for computer instructions and a vocabulary of common tasks. The Google-Oracle suit concerns the structure of the vocabulary Java provides.

    In one programming language, telling the computer to alphabetize a list of names might look like `sort(names, ALPHA)`. In another, it might look more like `names.alphabetize()`. The concept—and effect—is the same, differing only form of expression.

    Programmers of all stripes spend a lot of time looking up such trivial details, or dealing with errors when they misremember instructions. But the more familiar one becomes with a particular programming language and “standard library”, or vocabulary, and its APIs, the less time one has to spend looking at references. The more “fluent” programming becomes.

    There are a _lot_ of fluent Java programmers out there. Google apparently sought to leverage that fact by making Java the lingua franca of its Android system. In that sense, Google very much drafted off of creative effort, title to which came into Oracle’s hands by acquisition. The project of designing the APIs for a language like Java is essential, but by no means easy to do well. Oracle has rightly emphasized the effort and expertise applied to the Java APIs’ design.

    That being said, a great many programmers, especially Java programmers, see Oracle’s claim standing for a much broader threat to their professional autonomy, not to mention the substantial intellectual investments they’ve made in becoming “Java programmers”. When _they_ reach for punch, they often hyperbolize that strong copyright in software APIs would proprietize and balkanze the very language of computer programming.

    Those curious might like to investigate the parallel case of Apple, as a contrast. The lingua franca of iOS computer programming has long been a language proprietary to Apple: initially Objective C, and now a newer language called Swift. That’s well and clean, from a certain IP point of view. But also, in a fairly clear way, suboptimal. Developers “fluent” in Apple programming can’t readily apply their fluency in other domains. They are, to the extent, rather beholden to Apple. To escape those boundaries, they must learn or refresh a different programming language.

    Hope this is helpful.

Varsity Sponsors

Junior Varsity Sponsors

Deprecated: Constant FILTER_SANITIZE_STRING is deprecated in /www/ipwatchdogcom_574/public/wp-content/plugins/the-events-calendar/common/src/functions/utils.php on line 1130

Deprecated: strpos(): Passing null to parameter #1 ($haystack) of type string is deprecated in /www/ipwatchdogcom_574/public/wp-content/plugins/the-events-calendar/common/src/Tribe/Repository.php on line 2129

From IPWatchdog

Deprecated: str_replace(): Passing null to parameter #3 ($subject) of type array|string is deprecated in /www/ipwatchdogcom_574/public/wp-content/plugins/jetpack/modules/photon-cdn.php on line 80