Oracle vs Google: Copyright Claims Must Remain

swandives writes "More in the Oracle/Google patent infringement saga. Oracle says no court has ever found that APIs for software like Java are ineligible for copyright protection. The claims were made in its objection to Google's request that the court make a summary judgment on Oracle's copyright allegations. In early August, Google asked the judge to rule that Google doesn't infringe Oracle copyright in its implementation of Android. In an objection to that request, Oracle asked the judge to let the charge go to trial. Earlier, Judge Alsup denied Google's attempt to get a potentially damaging e-mail redacted. Looks like this one could take a while."

Comments

Aaaaand, enter full bastardry.

[unity100]

What does this mean for the future of Java users and developers ? will we be oracle's bitches ?

i would like to let anyone know that i wont, if it comes to that.

Re:Aaaaand, enter full bastardry.

[Anonymous Coward]

Well, Java was a nice language while it lasted.

Re:Aaaaand, enter full bastardry.

[mrchaotica]

This is the copyright law equivalent to giving WOPR the launch codes. If a court rules that a company can own an API, then everybody's software becomes infringing!

Re:Aaaaand, enter full bastardry.

[nedlohs]

No, everybody who implements a compatible library or creates there own/ships header files would be infringing.

Re:Aaaaand, enter full bastardry.

[node 3]

This is the copyright law equivalent to giving WOPR the launch codes. If a court rules that a company can own an API, then everybody's software becomes infringing!

No, just those without proper licenses, like Google.

Re:Aaaaand, enter full bastardry.

[jc42]

If a court rules that a company can own an API, then everybody's software becomes infringing!

This isn't a new concern. Back in the 1980s and 1990s, I worked (as a "consultant") on a number of projects at Digital. One of the discussions that came up occasionally was why DEC's unix systems were all based on BSD, and not Sys/V. It was well-known that DEC had Sys/V running on their hardware internally, but for some reason they didn't want to sell it.

The explanation that came up every time was that the Digital lawyers had nixed the use of Sys/V and other AT&T code for the same reason that we're discussing now: The run-time libraries all contained AT&T copyright notices in every routine, so if you linked to those libraries, your binaries would contain AT&T copyright notices. This included libc, so pretty much all binaries produced on Sys/V contained lots of AT&T copyright notices. There was a very good chance that AT&T would have a legal claim on any software that contained those copyright notices.

The lawyers apparently did point out that the status of these copyright claims in binaries was a legal "gray area" that had never been properly tested in the courts. Their professional legal advice was to let someone else be the sucker^H^H^H^H^H^Hguinea pig who paid the legal fees to fight AT&T on the issue. Until that was decided, using AT&T binary libraries was legally too risky, and since the BSD libraries were not such a legal threat, DEC should stick with BSD, which did the job just fine.

Disclaimer: I never personally talked to any of these purported DEC lawyers to verify this story. But it was widely believed by all the DEC insiders that I talked to. I'd imagine that the same sort of discussions must be going on inside a lot of current companies with respect to java. I'd also guess that a lot of companies lawyers are advising that their clients minimize the use of java until the courts sort out the legal issues, just to be on the safe side. Why risk your company's profits on a language that may be legally incumbered in ways that are unknowable today, when there are similar languages (python, perl, etc.) that the geeks say are just as good and are legally safe to use..

(Yeah, I know I'm risking a language flame war by that last comment. Hopefully the mods here will mark them OT. ;-)

Re:Aaaaand, enter full bastardry.

[Anonymous Coward]

No, just those without proper licenses, like Google.

Except most of the software you use daily uses APIs (because they are published) without a proper license (because none was asked) and would be infringing if that is made precedent.

Re:Aaaaand, enter full bastardry.

[davester666]

No. It wasn't.

Re:Aaaaand, enter full bastardry.

[node 3]

No, just those without proper licenses, like Google.

Except most of the software you use daily uses APIs (because they are published) without a proper license (because none was asked) and would be infringing if that is made precedent.

No, this isn't about using (i.e., writing software that makes Windows system calls, for example) an API, but which duplicates an API. WINE could possibly be affected, but Photoshop would not.

Re:Aaaaand, enter full bastardry.

[pugugly]

More to the point it removes the entire point of white-box reverse engineering. Remember when Compaq reverse-engineered the IBM bios chip, creating the first IBM-compatibles? Q-Dos was recreated from the published manuals of CP/M.

If Oracle thinks they want to win this fight this way, they haven't thought how many principles of simple databases they have implemented that would fall under this concept.

Pug

API?

[Microlith]

Why the hell should an API, the computer equivalent of a phone number, qualify for copyright protection?

Implementation behind those, yes. The actual API itself? Well, I guess it's a great end-run against 3rd party reimplementations...

Re:API?

[Anonymous Coward]

Have you ever written an API? Ever admired a well-written one, or hated a poorly-written one? Guess not, or you wouldn't ask that question. It's obviously "the work of an author".

Re:API?

[Anonymous Coward]

Why the hell should an API, the computer equivalent of a phone number, qualify for copyright protection?

Implementation behind those, yes. The actual API itself? Well, I guess it's a great end-run against 3rd party reimplementations...

Because an API is like a patentable 3.5mm jack for software, and not like a phone number.

Re:API?

[JonySuede]

There are books on good api design, if you read one of them you will see that there is a lot of creative works involved into the design of a pleasant*1 to use yet generic API, even more so if you want to stay backward compatible. I am pretty sure that a well designed API qualify as a copyrighted worked.

1- Please ignore the Date class as they were young and did not know what persistent horror they unleashed upon the world.

Re:API?

[ZombieBraintrust]

Java took the C++ api in part. Oracle is arguing that they themselves are commiting copyright infringement.

Re:API?

[empiricistrob]

Uhm... an API is much more than a phone number. I think equating those two is ridiculous. An API is a huge set of interfaces with specific names and behaviors. A better analogy would be a phone book -- but one that you created from scratch and wasn't based on any public information.

The real point, in my opinion, is whether Oracle *protected* their copyright to the API. It seems to me that Sun/Oracle treated the API more-or-less like it was public domain. I never accepted an agreement that stated I only had permission to read the API docs for use with an approved JVM. This is sort of like publishing a short story on the internet, it goes viral and is used by tens of thousands of people (reposted on blogs, etc), and then ten years later someone puts it a book and the original author claims that it's copyright infringement. Typically with IP law you have to protect your IP or you loose your rights.

Re:API?

[guybrush3pwood]

I believe you're not thinking this properly. An API is a piece of software, just as any other. Given the current state of affairs, why shoulnd't it be subject of copyright? Perhaps you're thinking "But an API is only an Java interface, no implementation!" I say: so? Even if it an API was so simple (it's not; it includes input/ouput parameters, exceptions, a usage definition, documentation, and so on), why would the simplicity of something prevent it from being subject of copyright? Hint: comparisons with one-click-purchase do not apply.

Furthermore, an API is not the equivalent of a phone number. An accurate analogy is this: an API is to a piece of software what the keyboard is to the phone.

Re:API?

[Anonymous Coward]

Have you ever created a phone book? Obviously somebody worked really hard to create it. It's not eligible for copyright either.

Re:API?

[shutdown -p now]

I have designed APIs - it's still a part of my daily job. Yes, it's not something that's easy to do, but I wouldn't say that it's particularly creative, especially for non-novel things (which 99% of Java class library is).

That said, IMO, interoperability here should trump copyright in any case. Copyrighting APIs has a significant negative effect on everyone in the market, because it encourages lock-in, and thereby hinders fair competition.

Re:API?

[jedidiah]

Yes. Yes. Yes.

You are an idiot that clearly has never done any of these.

Otherwise you would know how absurd the concept of copyright on an API really is.

Re:API?

[jedidiah]

...and this here is why copyrights on APIs are a really stupid idea.

The people that came up with this gem must be "code illiterate" for not realizing the broader implications of their idea.

Re:API?

[SomePgmr]

I'm not sure if this is true, but I've read that publishers seed phone books with fictitious information so they can copyright the work and identify future infringement.

Re:API?

[Microlith]

An API is a huge set of interfaces with specific names and behaviors

And you can copyright the implementation of those APIs, and the shiny books that document them. But the APIs themselves cannot sanely be copyrighted (being arbitrary identifiers with specific parameters and expected outputs) otherwise reverse engineering of ALL kinds (WINE, etc.) would be illegal.

Re:API?

[JonySuede]

I find the STL and the Java standard class library to be quite different.
Ignoring the alien syntax, Java took more from Smalltalk than they did from C++....

To use a book about car analogy, it like reading a book about combustion engine in English and another book about the same subject written by a different author in German.

Re:API?

[Desler]

You are actually wrong. Yes, information, such as a phone number or address, is not eligible for copyright protection but that does not mean the phone book as a whole work cannot be copyrighted. You have incorrectly understood the ruling on such things. This is why I can create my own phone book with information copied straight out of, for example, the Yellow Pages but at the same time I can not take a Yellow Pages phone book, scribble out their name, put my own on it and claim it as my own work as the work as a whole can be copyrighted if it meets certain statutory requirements.

Re:API?

[msauve]

It's more like claiming copyright on the individual words one used to write a creative story.

Computer languages should not be copyrightable, because copyright requires that the work be fixed in a tangible medium of expression. Just as individual words can be taken from one work and rearranged to create a wholly separate, noninfringing one, so too should computer language grammar, syntax, keywords, etc. be available for others to reuse.

If Oracle claims copyright over individual Java language elements, regardless of how they are arranged, is their next step to claim they own the copyright to any program written in Java?

Now, this is a bit different than the trouble Microsoft got in when they used their embrace, extend, and exterminate strategy with Java. As I understand that, MS had a license, and used code from Sun's implementation of the Java language, it was not a clean room implementation based on the formal language definition (APIs).

Re:API?

[JonySuede]

I agree with you on:

interoperability here should trump copyright in any case. Copyrighting APIs has a significant negative effect on everyone in the market, because it encourages lock-in, and thereby hinders fair competition.

But that should be written in the laws if it is not already written.

Re:API?

[Anonymous Coward]

Unfortunately absurdity is not the relevant standard.

Re:API?

[Anonymous Coward]

Please don't confuse copyrights and patents. Despite the stupid term "intellectual property", these are very distinct things.

Re:API?

[Animats]

Why the hell should an API, the computer equivalent of a phone number, qualify for copyright protection?

An API is not the "equivalent of a phone number". Not even close. It's clearly a creative work. The question is whether it is too abstract to be protected by copyright. "Copyright protection is not available for ideas, program logic, algorithms, systems, methods, concepts, or layouts.", says the Copyright Office. It might be patentable, if sufficiently original, but that's not an issue here.

Copyright in fictional characters has been tried, occasionally with success. That's one of the broadest forms of copyright protection. Trademarking the name of the character provides more protection, which is why we see Darth Vader(tm) markings. That doesn't apply to an API.

Re:API?

[Anonymous Coward]

They're not even talking about the part of the API that manages interactions, they're talking about interface files... Nothing more. How the hell do you copyright an interface file?

Re:API?

[Desler]

For some good information you need to read this, this and this. Once you read these then you will understand why you are quite wrong. It's the same reason why one cannot copyright a recipe, but your own expression of that recipe can be copyrighted. The same thing applies to phone books.

Re:API?

[vux984]

It's obviously "the work of an author".

So is a mathemtatical proof.
But they aren't copyrightable either.

Re:API?

[3vi1]

You totally missed the point. You're talking about copyright on the code implementing the API - while the real topic here is whether or not the API calls (function names) themselves can be copywritten.

And of course the answer is 'no', because that prevents any and all compatible implementations. In fact, you'd be in violation simply for writing a program that called the API - since you have to use the function names in the calling program.

Oracle's lawyers know nothing about programming, apparently. If things worked like they're trying to say they do, Microsoft could sue anyone that made software for Windows because at some point you used a header that included the Windows API function names.

Re:API?

[JAlexoi]

Ofcourse they don't. If they do succeed, then a crapload of their own software will be considered derivative from LGPL! Oh and I would love to see them defend their own hypocrisy...

Re:API?

[DamnStupidElf]

An API is just a list of function prototypes and the type definitions they depend on. The only reason function names are required is that so far there isn't a general semantic langauge that specifies (simply, and with clarity) the preconditions and postconditions of the functions, so the linker needs a way to bind calls to the appropriate functions instead of just pattern matching based on formal semantics.

If "type_1 function_name_1(type_2, type_3, type_4, type_5) ..." is your idea of copyrightable content, then I have some very boring novels to sell you.

Re:API?

[angel'o'sphere]

In germany, and I believe in all EU (europe) APIs and "database layout" are explicitly exempt from "copyright".

Looking at projects like "GNU claspath" the logical assumption is that this is also true for the USA.

An API is like the "mains power" supply in a house. The connector can't be under "copyright" or no one would be able to connect his electric power consuming device with a reasonable price to the grid.

angel'o'sphere

Re:API?

[sexconker]

Please don't confuse copyrights and patents. Despite the stupid term "intellectual property", these are very distinct things.

Please don't intentionally confuse people by insisting copyrights and patents are fundamentally different and offering no explanation as to how or why.
Despite the different terms and mediums (physical implementation vs written work), they are essentially the same fucking thing.

A copyright gives you the right to produce a copy of a written work.
A patent gives you the right to produce an implementation of a design.
Both things essentially say "This is mine, you can't steal it without paying me.".

Re:API?

[Anonymous Coward]

A proof is copyrightable. A proof is deeply the same thing as a comuter program. There are thousands of ways to write the same proof, and the idea of the proof may not be copyrightable. But any partiuclar statement of it is.

Re:API?

[JonySuede]

you have sane laws... especially the law on beer purity ;)

here, in Canada, i am not so sure anymore, and in the USA they have fucked up laws!

Re:API?

[larry bagina]

Smalltalk by way of Objective C and OpenStep. OpenStep was a collaboration between NeXT and SUN. The early java libraries looked suspiciously familiar.

Re:API?

[Anonymous Coward]

A copyright gives you the right to produce a copy of a written work.
A patent gives you the right to produce an implementation of a design.

Exactly, so the copyright held by Oracle has nothing to do with Google implementing a system that is API-compatible with Java, unless they copied the code itself.

Re:API?

[vux984]

A proof is copyrightable. A proof is deeply the same thing as a comuter program. There are thousands of ways to write the same proof, and the idea of the proof may not be copyrightable. But any partiuclar statement of it is.

Precisely like an API.

You can copyright your expression of the proof, but not the proof itself or even the idea used to make the proof.*1

This is how APIs should be treated. A particular expression or documention of one is clearly copyrightable, but the actual function signatures not so much. And anyone can write there own implementation or documentation of an API without violating copyright, just as anyone can write their own mathematical proof without violating copyright.

*1 -- On a bit of tangent:

A software patent attempts to patent the idea expressed by a computer program, but a computer program, as you noted, is "deeply the same thing" as a mathematical proof which is categorically "not patentable". This is why software patents are infuriating and contentious to me.

Re:API?

[Anonymous Coward]

they are essentially the same fucking thing

No, not really. Both are a form of intellectual property protection, but a copyright protects the expression of an idea in a fixed medium, and gives the creator exclusive rights to make copies and/or distribute the work. Patents protect a design or process, and not only does the creator have the exclusive right to copy and distribute the patented design/process, he has the exclusive right to even *use* it.

Re:API?

[Slashdot Assistant]

Has there been any patent or copyright case in the United States (or similar western nation) in which ownership was legally weakened because the clearly identified author delayed in enforcing their rights? The only problem I'm aware of is that if a word did go viral then the author will have to prove that they did indeed author the work in question. That could be tricky if all they have as evidence is the time stamp of a file saved on their hard drive.

Re:API?

[SharpFang]

If the book has the same layout of chapters, follows with the same progression of descriptions paragraph by paragraph, it is a plagiarism and copyright infringement. Neither translation to a different language nor rewording of paragraphs of the book while retaining the same information in the same (non-obvious) order are releasing you from the burden of copyright.

 

Re:API?

[shugah]

APIs are a 2-way street. You publish an API and distribute header files as part of an SDK, you make it easy for third party application developers to develop applications for your platform. It also makes it rather straight forward for another platform developer to emulate your API so that those same 3rd party apps, written to your API can be run on their platform. In the case of Java, Sun/Oracle also released and distributed the API and the whole JVM source code under a GPL license. So this made emulation of Java Class Libraries trivial. To now claim copyright infringement is absurd.
Again - this is not new. When Tim Paterson wrote 86-DOS (MS-DOS) he did it with an API programming manual for the yet to be released Digital Research CP/M-86 on his desk and basically emulated the system calls (which is one reason for some of the arcane crap that was in DOS for years afterwards).

Re:API?

[SharpFang]

You can also copyright header files that define these APIs. And the files would "enjoy" the same protection as books and such - mere renaming of everything, shuffling the order around or changing comments, without actually breaking the gist - the underlying concept - will be recognized as plagiarism. As long as your file does, every single thing the original file does, you're clearly in the black.

This is bad. Oracle has a solid point here.
Of course asserting copyright on API is both extremely stupid and extremely evil. APIs fundamental purpose is to be open.

Reverse-engineering is different here. You take an observed behavior and write code that seems to do the same thing. You don't really know if it does the same thing. If it does, good for you, but you create it from scratch and are never completely sure if you missed anything. Not a copy but a replacement. You got a lock, old and worn, and you make a key that fits that lock. The key will be unique and possibly quite different from the original - some stuck tumbler will be missed, some worn one will have a different depth, another will barely fit within tolerances.

OTOH, API must conform to specs. You get the specs of the lock and the key, and you create a key that follows the specs exactly. It will be an identical copy of the original one. The fact not one is a copy of the other but they both are derived strictly and exactly from the same abstract data is moot here.

Re:API?

[Anonymous Coward]

The identifiers aren't arbitrary to people, only to compilers. They tell the story of the API. They are human readable, and their selection and combination into the "nouns" and "verbs" of the API are the expressive part of API design.

If you take an API and replace all the identifiers with arbitrary symbols, then I can't see how you are infringing. You are taking the idea, and completely changin the expression.

But are you allowed to copy not only the API design (which is probably not subject to copyright), but all the names of classes, interfaces, types and exceptions?

If so are you allowed to take a literal copy of the API specification (e.g. header files), or must you re-write them?

The best argument for copyrightability is that particular identifiers used in the API constitute a protected expression of the idea of the API, and you're free to copy the idea and change the arbitrary identifiers to something else. Of course this result is pretty darn bad, as two expressions of the same API ideas would not be compatible.

Re:API?

[vranash]

And Germany still has those 'hacking tools' laws covering nmap, nessus and such, correct? I've been having a hard time (for about ten years now!) finding somewhere worth emigrating to that doesn't have some fscked laws that remove them from my list of potential candidates. Sad really, because every year seems to find the US falling further and further into insanity (as it has been since I was in High School, at least. 9/11 changed everything? No, Columbine changed everything.) Where's my pacifier? This IS a nanny state, isn't it? :D

Re:API?

[JonySuede]

here is the index to the stl by SGI :http://www.sgi.com/tech/stl/stl_index_cat.html
there is the same type of index in java : http://download.oracle.com/javase/7/docs/api/allclasses-noframe.html

in the light of the links I posted, my book analogy was flawed it should have been an English book about the combustion engine and a German one about car subsystems including some chapters on the combustion engine.

Re:API?

[wurp]

When the hell did "you'r e an idiot" posts start qualifying for an upvote of any kind?

Re:API?

[russotto]

Oracle's lawyers know nothing about programming, apparently. If things worked like they're trying to say they do, Microsoft could sue anyone that made software for Windows because at some point you used a header that included the Windows API function names.

That's the result they desire. And oddly enough, the FSF appears to take the same position on occasion.

Re:API?

[node 3]

A proof is copyrightable. A proof is deeply the same thing as a comuter program. There are thousands of ways to write the same proof, and the idea of the proof may not be copyrightable. But any partiuclar statement of it is.

Precisely like an API.

You can copyright your expression of the proof, but not the proof itself or even the idea used to make the proof.*1

This is how APIs should be treated. A particular expression or documention of one is clearly copyrightable, but the actual function signatures not so much. And anyone can write there own implementation or documentation of an API without violating copyright, just as anyone can write their own mathematical proof without violating copyright.

This is a variation on the whole "a digital file is just a number, and you can't copyright a number" rhetoric. A mathematical proof is a description of something that already exists. It's a discovery. An API is an invention. It's something that did not exist until someone created it, like a song, book, or movie.

Re:API?

[guybrush3pwood]

I was not addressing the Oracle/Google case, but the rather general case proposed by the GP. I think it's pretty clear that you can design, publish and copyright APIs, them being software artifacts. They posses no inherent condition that make this impossible, IMO... besides the "software is an algorithm and algorithms are math" argument, of course.

On the other hand, could we make the case that an API is not software?

Re:API?

[Doctor_Jest]

Not necessarily pay... it's up to the copyright holder. Copyright doesn't guarantee revenue, it guarantees (for a limited time....that's a laugh) control of a work. It's perfectly within the copyright holder's rights to give away or permit unlimited copying of said work. Patents have to be defended or they're lost as well. Patents last for a much shorter time, and are subject to prior art (if the Patent Office could be bothered...) There are similarities though... Both copyright and patent trolls infest our world like that pee smell in the back of city buses. :)

Re:API?

[Anonymous Coward]

an API is more than a function signature. there is an associated behavior. IMO, it is not copyrightable, i.e. work of art, but more akin to engineering and problem solving, and hence patentable.

Re:API?

[vux984]

This is a variation on the whole "a digital file is just a number, and you can't copyright a number" rhetoric.

No, not at all like that.

An API is an invention. It's something that did not exist until someone created it, like a song, book, or movie.

Like a book? So this little list below:

Foundation
Foundation and Empire
Second Foundation
Foundations Edge
Foundation and Earth

Is that an invention? Or simply a description of what exists?

A software library is a set of callable functions. How is a list of the function signatures (the "API") in that library somehow different?

Ah, but perhaps you'll tell me I'm putting the cart before the horse, the API was written first, and the library came after?!

So what? Suppose Asimov, had scribbled the following on a napkin back before he penned Foundation... (And yes set aside for a moment, that Foundation wasn't penned as a novel originally...it's not the point)

Foundation
Foundation and Empire
Second Foundation
Foundations Edge
Foundation and Earth

At that point, yes this list of titles was an invention, a work of fiction on its own, the books did not exist, he had yet to write them. Fast forward a few decades... the books are written. And the list, is now a description of what exists, and it would be absurd to argue that people wishing to enumerate his works should be forbidden from writing:

Foundation
Foundation and Empire
Second Foundation
Foundations Edge
Foundation and Earth

Simply because he had written this list on a napkin before he started.

How exactly is an API different?

Further, an API is by definition, the method by which other software interfaces with it.

Copyright law has specific exceptions that explicitly and specifically allow reverse engineering and decompiling just to figure out what the interface specification actual is -- in the event that its not readily available/documented so that the discovered interface could be used for interoperability.

It would absolutely absurd if after going through all that trouble to legally protect our ability to discover what the interface is ( (hmm "discover"... as is describe something that exists), to then prohibit us from writing it down or using it, when the express purpose of the section of law was to enable interoperability.

Re:API?

[Anonymous Coward]

> Why the hell should an API, the computer equivalent of a phone number, qualify for copyright protection?

If it does, Oracle better hope like hell that IBM doesn't have a bunch of copyrights covering the SQL language. Surely SQL would qualify as an API for database access, and I doubt that Oracle's implementation of the SQL "API" would be seen as sufficiently unique in a copyright context - surely it would be 90+% identical with IBM's implementation.

No, I'm not a lawyer, but thanks for asking...

Re:API?

[Anonymous Coward]

All Columbine, 9/11, and other similar incidents have done is to turn everyone into one big collective pussy. Everyone's afraid of everything now. I heard that when this earthquake hit in Virginia, some people's thoughts were "OMG it's 9/11 again".

Re:API?

[WorBlux]

An API is a language, a way for one computer construct to talk to another. Can you copyright the english language? French? Latin? Tolkein Elvish? Esperanto? You can copyright particular dictionaries, but not the entire language (even in the case of purely invented languages).

Re:API?

[WorBlux]

You cant' copyright the behavior of a computer function. It's not in a fixed medium. Patents may be on some behaviors, but not a copyright. A name is not copyright-able either. If you invent the word giblywonk to describe the sound a mediocre metal band makes while practicing drunk and stoned and I then trow ten tom cats and a howler monky in a metal garbage can and beat on the lid with a baseball bat, there's no reason I can't use the word giblywonk to describe that sound I just made, as for all practical purposes there is no difference. An API is just a language for describing functions.

Re:API?

[mike.mondy]

No, an API isn't like a phone number. Save phone numbers for the IP address analogies...

Queue car analogy.

An API is like the power socket or lighter socket on a car. They're all the same diameter, all provide 12 volts DC, etc no matter whose car you buy. This means anyone can create or sell a device that "hooks" into the car's power.

The "I" in API is "interface" of course -- An API is a standard for how two separate things talk to each other.

Oracle seems to want to say feel free to build something that plugs into our power socket, but don't you dare build something with a power socket like ours. Hopefully, copyright can't be twisted to allow that.

Re:API?

[Cyberllama]

It would be one thing if we were talking about the source code that made the API function, but we're not. We're just talking about the idea of the API itself which is something Sun put out there explicitly for people to use (admittedly they wanted you to use THEIR virtual machine--not one written by Google). It may not be what they had in mind, but it's not as if Android is competing against Sun's VM so there's no real harm here. At the end of the day it's just a money grab by a company that bought Sun and is now growing increasingly irrelevant.

Re:API?

[ajo_arctus]

Actually, it is. At least here in the UK - I haven't looked in to it in the US yet. We have something over here called 'database' rights, which protect your collection of otherwise public data. It's meant to protect the effort that goes in to collecting and collating data for a directory. It's a bit of a pain when you want to start building your own directory of E.g., business listings (which is why I looked in to it).

Re:API?

[ajo_arctus]

A better analogy would be a phone book

Wouldn't a better analogy be the phone system itself? The connector, the line voltages, the time between signals, etc. Surely they would be the API? The phone number is just a value you pass in to the API. The phone book is just a dictionary of those numbers.

Re:API?

[Ami Ganguli]

You can also copyright header files that define these APIs. And the files would "enjoy" the same protection as books and such - mere renaming of everything, shuffling the order around or changing comments, without actually breaking the gist - the underlying concept - will be recognized as plagiarism.

This is the crux of the issue. My understanding (and I guess also Google's understanding) of copyright law is that the copyright is only for the "creative" element of the work, and explicitly excludes things required for interoperability.

So if you take out all the comments, organize the remaining header in a completely non-creative way (ie. sorted by subject area, and then alphabetically), and remove any lingering "private" parts that aren't strictly part of the API, what remains should not be subject to copyright.

It may be true that this has never been tested in court, so I don't know were I got the idea that this is how things worked, but if I'm wrong then I've been under this delusion for at least ten years or so.

Re:API?

[Paradigma11]

...

This is a variation on the whole "a digital file is just a number, and you can't copyright a number" rhetoric. A mathematical proof is a description of something that already exists. It's a discovery. An API is an invention. It's something that did not exist until someone created it, like a song, book, or movie.

Only if you are a mathematical platonist http://en.wikipedia.org/wiki/Philosophy_of_mathematics#Platonism .

Re:API?

[Anonymous Coward]

Not if that API is considered part of the language, which in Java - it does (and this fact is promoted as an advantage of the Java language). By making it part of the language Sun relinquished any copyright claim on it.

Re:API?

[angel'o'sphere]

Yes unfortunately there are laws regarding "hacking tools", however I don't know which tools are covert by it. Very likely tools like nmap, that are installed by the vendors by default are excluded ;D
In fact those laws are in general ignored. I'm working right now as sysadmin ... in a bank, that is. I'm not responsible for security stuff, however I have a brought overview of the network ... and surely we use *everything* available to test, monitor, investigate our network.

Re:API?

[AmiMoJo]

That said, IMO, interoperability here should trump copyright in any case.

In the EU it does. Making a compatible interface is specifically allowed by law. I'm surprised there is no such provision in US law because it allows companies to monopolise and prevent others from competing with them.

Re:API?

[Anonymous Coward]

If APIs were just lists of facts rather than creative works then no APIs or parts of APIs would ever be deprecated.

Re:API?

[Anonymous Coward]

An API is just a list of function prototypes and the type definitions they depend on.

You've created a list, choosing to include some items and to leave some out. The list you have created is an original work and is protected by copyright. You now decide to publish this list calling it an API, so I guess that makes it lose its copyright protection...?!?

Re:API?

[Anonymous Coward]

I find the STL and the Java standard class library to be quite different.
Ignoring the alien syntax, Java took more from Smalltalk than they did from C++....

Look in java.util.* some time. Take a long hard look at things like java.util.Vector (compare against std::vector) and java.util.LinkedList (against std::list), pay special attention to the concept of an "Iterator" (java.util.Iterator). The function names line up pretty well by themselves, never mind the implementation similarity; it's little more than a rewrite in a different syntax.

Re:API?

[tbannist]

When they started being informative, insightful and underrated?

Re:API?

[Anonymous Coward]

You're obviously not speaking from experience, but if you really think so, try it. Copy some mathematical proofs, publish them with your name on them, and see where that gets you!

Re:API?

[vranash]

Good to know! That happened to be my biggest disincentive to looking at Germany. The other hurdle being to get my car legalized so I could take it with me, and then out to Nurburgring :)

Re:API?

[angel'o'sphere]

What kind of car do you have? I would asume a "standard" car only needs to visit "TÃoeF" to check its condition and then you register it to get the license plate. However if you have tinkered with it, you might have more work with the TÃoeF or if you have a british car with the steering wheel "on the wrong side" perhaps you can not register it (no idea ;D)

Re:API?

[Anonymous Coward]

Oracle allows anyone to create their own compatible Java implementation. Just download the language specification, VM specification and Javadocs for the libraries and get coding. Or if don't want to start from scratch, you can base your work on Oracle's implementation, which is available under the terms of GPL (and presumably other licences). An API being copyrighted or not doesn't affect this at all.

Re:API?

[idontgno]

That's a good insight, and the previous mention of phone numbers is also useful.

Feist Publications, Inc., v. Rural Telephone Service Co.: the actual phone numbers in a telephone book are not copyrightable, but are rather mere facts, and therefore can be copied into another phone book without infringement. Structurally, this smells like the whole API signature thing to me.

Along the same lines, a cookbook is copyrightable, as a creative assembly; the individual recipes aren't, if the "copy" of the recipe expresses the facts of the recipe (the ingredients, the steps) in a manner that is at least slightly different than the original.

If you google Idea-Expression Dichotomy, you can get a pretty good overview of the current thinking.

I think Oracle's chances aren't that good, other than the whole "it doesn't have to make sense or be consistent with precedent in other areas of copyright law" thing. (I.e., Oracle could win this just because they confused a judge thoroughly enough, or because the judge thinks computer programs are somehow fundamentally different than recipes and phone books, or Oracle's lawyers are better-looking, or whatever.)

You know, the sad situation where the logically wrong and socially bad thing wins and becomes the law of the land, simply because Justice is an imperfect process.

Re:API?

[idontgno]

No, but we can argue that not all software is sufficiently creative to merit copyright protection. Since APIs are both creative/expressive and functional, the question boils down to "does the non-copyrightable functional aspect outweigh the creative/expressive aspect"? If the answer to that is "yes", then bare API signatures aren't copyrightable.

And how to answer that question? I don't think there's any objective self-contained criteria. Since an API signature would work as well if you, for instance, renamed all the parameter names... and even the method name, assuming you have some kind of mapping mechanism... they're not explicitly creative (i.e., the method itself, and its parameter order, are pretty much functional). But the canned signature usually reproduces the original method and parameter names, which can be amazingly creative sometimes.

So... it may come down to social and commercial effects (that is, costs and benefits outside of the work itself, but coming from the precedent set.)

If you can copyright APIs, interoperability is dead. A proprietary software developer can forbid any kind of unlicensed interoperability because you have to copy the APIs to have interoperability. Farewell, most free software compatibility with proprietary protocols, libraries, interfaces, and data formats.

Whether that's good or bad depends on if you depend on lock-in to make your money, or not.

Re:API?

[guybrush3pwood]

Copyrighting an API would not kill interoperability, because when you use an API, you don't copy it; you just use it. In particular, using a Java API based on interfaces, requires having the classfiles, not the source files. Thus, using a Java API does not require "copying the API", in the sense of using for free somebody else's work. Now, if you build you own system, which has an API, and instead of coding your own interfaces, copy the source files someone else wrote... that is "copying the API", and you'd be effectively commiting what copyright tries to protect.

Another example: web services. The API is defined by a WSDL. If you download the WSDL of an Amazon web service, for instance, and then use it in your own on-line-bookstore system or whatever, then you're copying it, and quite possibly infringing a copyright. If you implement a caller of the web service, you are using the interface for interoperability purposes, without infringing any copyright.

Re:API?

[node 3]

A list of books is not an invention, the books themselves are the invention.

An API is part of the invention. It's more like an index of a book that is so extensive one can recreate the book from the index alone (which is what Oracle is complaining about here). Or it's like sheet music, or a movie storyboard, etc.

Copyright law has specific exceptions that explicitly and specifically allow reverse engineering and decompiling just to figure out what the interface specification actual is -- in the event that its not readily available/documented so that the discovered interface could be used for interoperability.

Copyright law has a *LOT* of ambiguity, you mean. Just because it says something vague in one part does not mean it trumps something specific in another part. "Fair use" is very difficult to define sufficiently objectively.

It would absolutely absurd if after going through all that trouble to legally protect our ability to discover what the interface is ( (hmm "discover"... as is describe something that exists), to then prohibit us from writing it down or using it, when the express purpose of the section of law was to enable interoperability.

The 'hmm "discover"... as is describe something that exists' does not mean what you think it means. This "something that exists" is a creative work. You can't "describe a creative work" in such detail as to duplicate it.

Re:API?

[node 3]

Except Google took Sun's (now Oracle's) "dictionary", and copied it. They even admit knowing violating Sun's IP in the process. First they tried to negotiate a deal, then, in an internal email, stated that their choice was to come up with something entirely different, or just take Sun's IP without their permission.

On the other hand, it's absolutely absurd to compare a wholly invented by a single entity API with languages whose roots go back thousands of years and are the cumulative result of the impact of millions, and even billions, of people.

Re:API?

[node 3]

Only if you are a mathematical platonist http://en.wikipedia.org/wiki/Philosophy_of_mathematics#Platonism .

This isn't a philosophical argument, it's one of legal definitions. No matter what your philosophy of the nature of mathematics are, the same legal definition applies to everyone. You've also left out vast alternative/related philosophies that hold the same pertinent feature of discovery instead of invention.

Re:API?

[SharpFang]

This would be true for a very straightforward, simplistic API: initialize(), a whole bunch of very simple tool functions, some kind of flush() and optionally end().

But if the API suggests use of some clever design pattern, if it operates on original abstract ideas, has some automagic features that cleverly work around the language's shortcomings, expands the language syntax (e.g. through clever use of templates), cleverly automates or streamlines tasks that would have to be called in sequence - generally, goes out of the way to be more than just a dumb connector - then it's definitely a creative work and mere reshuffling and cleanup won't change that.

The other post writes about cookbooks. Yes, as long as the recipe is simple, generic "add, stir, bake, flip, chop", it's all right to just slightly modify it. But if the recipe first describes gently spraying the dough with liquid nitrogen starting from the corners, then drilling holes in a precise hexagonal pattern, pouring a matching cover of liquid tin to form airtight seal and then placing it in induction oven - then plain replacing the pattern with pentagonal will not really void your copyright on the recipe...

Re:API?

[vux984]

An API is part of the invention.

Its a description of what the invention does. You can't patent that. You patent the invention itself. There is NO patent protection for "what it does" the patents protect the "how it does it".

  It's more like an index of a book that is so extensive one can recreate the book from the index alone (which is what Oracle is complaining about here).

Any implementation of the software based on the API will be completely different from another one. They will have the same interface, and they will accomplish the same things, but they will not be the same.

Or it's like sheet music, or a movie storyboard, etc.

You cannot patent book indexes, sheet music or story boards either. Not every idea is patentable.

This "something that exists" is a creative work. You can't "describe a creative work" in such detail as to duplicate it.

Duplicating its functionality is not duplicating it.

In unversity, I was given an assignment to take X inputs and produce Y outputs. That is essentially an API. Everyone turned in a different program.

APIs do not describe creative works in nearly enough detail as to duplicate them.

Re:API?

[Anonymous Coward]

You've also left out vast alternative/related philosophies that hold the same pertinent feature of discovery instead of invention.

Only because Platonism is 2,500 years old - everything more recent is still under copyright.

Re:API?

[vranash]

Nah just a basic left hand drive car. And I already had read up some on the rules regarding modified vehicles and conditional use permits (IE you can have a street legal car with slicks but it's only legal to drive in dry conditions.). And actually compared to the euro spec model, mine was downtuned for the US market (Euro model was specced for a higher top speed on the Autobahn and related European roads, while the US model was specced for better low end torque.)

Re:API?

[Anonymous Coward]

The copyright standard is pretty low. FOSS drove a case to CAFC that established the breach of license claim.
Do you know what the copyright was? A collection of items
That is the base line in the 9th circuit. It sure looks to me that an API is much more copyrightable than a collection of items.

Mitel, Inc. v. Iqtel

[ZombieBraintrust]

Mitel, Inc. v. Iqtel, Inc. 10th Circ., 1997 In this case, the court found for the , saying that (1) menu structure of a computer program was an uncopyrightable process under 102(b), that (2) there was a merger between its idea and the limited number of ways in which it could be expressed, and that (3) the work lacked originality. It lacked originality because the command codes comprised a method of achieving a certain result based on call controller’s functions, the carrier’s technical demands, and the telephone customer’s choice.

Re:Mitel, Inc. v. Iqtel

[Anonymous Coward]

Wrong. Menu structure is not a software API (Application Programming Interface).

Re:Mitel, Inc. v. Iqtel

[bongey]

Just dove into the US Copyright Office.
IMPO an API is combination of a name and architecture copyright.
"No. Names are not protected by copyright law."http://www.copyright.gov/help/faq/faq-protect.html#band
Copyright law also also forbids architectural copyrights for
"structures other than buildings, such as bridges, cloverleafs, dams, walkways, tents, recreational vehicles, mobile homes, and boat "

Re:Mitel, Inc. v. Iqtel

[ZombieBraintrust]

Trademark could protect keywords if you appended your mark to them. For example org.apache.struts2.components; has the apache mark in it.

Re:Mitel, Inc. v. Iqtel

[ZombieBraintrust]

The above menu structure was command line. So yeah it was an Application programming interface. It was more of an API than java is.

Re:Mitel, Inc. v. Iqtel

[nbossett]

I would hope that courts wouldn't accept a claim that interoperability defenses are void if an API function call (at the compiled executable level you need to link to) happens to be named after a particular vendor or requires you to pass the brand name of a particular vendor into every call and that only people they license to use their name are allowed to use it in that context.

Hi Florian Mueller!

[Anonymous Coward]

FYI, you forgot to log in.

Re:Hi Florian Mueller!

[Anonymous Coward]

Screw you, I'm too busy holding my ankles for Larry to take the time to login.

Why does Oracle care?

[Anonymous Coward]

Can anyone explain why Oracle care so much about this? Are they in a proxy-battle for someone? They are after all a database vendor, right?

Re:Why does Oracle care?

[ZombieBraintrust]

Than plan to get money out of Google. They want a share of Googles mobile add revenue. Android pretty much destroyed their mobile Java app buisness.

Re:Why does Oracle care?

[Anonymous Coward]

Wouldn't that be the former Sun's mobile Java app business?

Re:Why does Oracle care?

[Lisias]

I beg your pardon, but...

The J2ME market destroyed itself. Confusing, locked down, artificially and arbitrarily limited. It was projected and implemented to meet carriers expectations and manufacturer's interest, not costumers'.

If Android didn't came in, another one would did it. Hell, perhaps even Windows Mobile Phone would be a good thing. =)

Re:Why does Oracle care?

[Anonymous Coward]

I work in a Java shop, do almost all of my development in Java and have written a few J2ME apps. I wouldn't wish that on my worst enemy. I would give up the ghost and move to .NET or something else before I'd get a job writing J2ME again. It was a nightmare. People think Android has fragmentation issues? It's a godsend compared to J2ME where you couldn't get consistent APIs, you'd have Java 1.2... plus a little bit of Java 1.3 randomly mixed in and that's what you were coding to.

Re:Why does Oracle care?

[Lisias]

Add to that emulators that doesn't correctly meet targets behavior and you're basically... SCREWED UP.

You never know if your program will, in fact, works on the target device until trying on it.

And since *no* J2ME devices (at least, to my acknowledge) allows device debugging or any kind of KVM's error monitoring by the independent developer (where's the manufacturer's developers have access to that), you have to use any resources you can imagine (from Ouijas to Black Magic) to try to figure out what the hell is going on.

Last year, one program of mine crashed on a (at that time) prototype from recently famoused manufacturer of (cheap) mobiles. That God damned program works on every other mobile we had in hands, except that fscking one (obviously, our chop was hired to build a feature program to be embedded exactly in that fscking model!). After almost a month of emailing and incompetence accusations, I managed (using social tricks that almost smelled as deception and traffic of influence) to discover the KVM's manufacturer by asking for an KVM's error report (basically, the stack dump of the killer exception) directly to one of the manufacturer's firmware developers (I don't know what my boss did to manage that, and I will never ask).

With the KVM's manufacturer's name, I found the products's bug tracking where a very known bug with LWUIT was reported, and a workaround published. 30 minutes after, I got my application "fixed", tested and shipped.

But what really seriously pissed me off is that the KVM's institutional page had all the mobile models that uses it - so why in Devil's name this information was omitted by the manufacture's support team was a mystery to my until we heard rumors that our costumer had asked the manufacturer's development team for the application first, but didn't agreed with the conditions. ;-)

I have, simply, not a single drop of sympathy or pity for any J2ME related business. They're going down by their own hands.

Re:Why does Oracle care?

[shugah]

So are you saying that rather than:

FTFA, Edward Screven, Oracle's Chief Corporate Architect: "Java, you know, is, in my mind, pretty well locked out of the smartphone market because of Android..."

This should read:

FTFA, Edward Screven, Oracle's Chief Corporate Architect: "Java, you know, is, in my mind, pretty well locked out of the smartphone market because of our own incompentance..."

Re:Why does Oracle care?

[jonwil]

Having played with development using J2ME, the only thing I found that is worse than J2ME is DoJa (which is basically a Japanese version of J2ME that happened to land on a few phones here and there in the west)

Re:Why does Oracle care?

[Lisias]

If you aren't part of the solution, then there is good money to be made prolonging the problem

I could not said it better. :-)

API Copyright?

[Kylon99]

If APIs are found to be protected under copyright, won't this mean something like Mono C# violates copyright as well?

And what about a DirectX emulator for Linux for example?

This doesn't sound right. These 3rd party APIs are, to me at least, the software equivalent of reverse engineering. Figuring out how the original works by providing emulation. This should be protected behavior or else it will be easier for companies to inadvertently gain new monopolies...

Re:API Copyright?

[ZombieBraintrust]

What about Java itself. Seem pretty similer to C to me.

Re:API Copyright?

[gnasher719]

If APIs are found to be protected under copyright, won't this mean something like Mono C# violates copyright as well?

The court didn't say "APIs are protected by copyright". The court said "there is no rule that APIs are never protected by copyright, therefore Google's request for summary judgement is rejected. The court will have to examine Oracle's APIs and then make a decision whether that particular API is copyright protected or not".

It will depend on the amount of creativity that went into writing the API. Now writing a good API is hard work and greatly appreciated, but that doesn't mean the author should be creative. It's more like a good API is not creative at all. On the other hand, a badly written API can be very creative and would then deserve copyright protection.

Re:API Copyright?

[Kylon99]

The court didn't say "APIs are protected by copyright".

This I am well aware of, but it seems that there is a chance to decide that APIs are copyrightable. Regardless of how 'creative' APIs are, I believe that as the point of inter-operability, they need to be non-copyrightable in order to allow people to reverse engineer. I can see how the courts cannot offer a summary judgement, but at the same time I don't have much faith in the U.S. court system to always come up with the right decision...

By the way. and this is just a side point, but there are 'creative' works out there that are non-copyrightable, such as culinary recipes.
http://www.copyright.gov/fls/fl122.html

We may need a lawyer to speak up on this, but it seems that the creative work would be the food itself, and the mere listing of its ingredients aren't copyrightable. Maybe the API definition will fall under this, being merely a list of functions and not the actual code itself.

Re:API Copyright?

[ZombieBraintrust]

I don't think the court has said anything yet. The above is Oracles argument against summary judgement. It is not the courts statement.

Re:API Copyright?

[WorBlux]

That doesn't make any sense to me. If I have perfect pitch and reverse engineered a song and performed it, such and act wouldn't get me around copyright issues. In fact I would be in a better position if the composer had freely given my a copy of his scoresheet instead. API's are really more like a list of word making up the language a computer program "speaks".

No court has ever found ?

[Anonymous Coward]

Its just that courts don't do abstraction - look at the wiki for the Borland vs. Lotus case -

"The court also considered the impact of their decision on users of software. If menu hierarchies were copyrightable, users would be required to learn how to perform the same operation in a different way for every program, which the court finds "absurd." Additionally, all macros would have to be re-written for each different program, which places an undue burden on users.[4]"

A reasonable ruling in this case would be similar: (using cut and paste similar to the way I code)
"The court also considered the impact of their decision on users of software. If software APIs were copyrightable, developers would be required to learn how to program the same operation in a different way for every language (a new one of which would be required for every company), which the court finds "absurd." Additionally, all software would have to be re-written for each different program, which places an undue burden on developers.[4]"

See http://en.wikipedia.org/wiki/Lotus_Dev._Corp._v._Borland_Int%27l,_Inc.

Goodbye Java

[Anonymous Coward]

"write once run nowhere"
html5 bring it on!

doing it wrong

[Anonymous Coward]

An API is a huge set of interfaces with specific names and behaviors.

A good API is nimble, lightweight, and efficient. If you think it's huge, you're doing it wrong.

Re:doing it wrong

[sexconker]

An API is a huge set of interfaces with specific names and behaviors.

A good API is nimble, lightweight, and efficient. If you think it's huge, you're doing it wrong.

Wrong. A good API is:

Functional
Covering
Fast

Whether or not the API is "lightweight" is more a testament to whether or not your application/service is trivial and pointless.
Arguing for lightweight over functionality and coverage is like arguing for a keyboard with fewer keys. Who needs that fucking Q anyway?

Re:doing it wrong

[WorBlux]

Arguing for lightweight over functionality and coverage is like arguing for a keyboard with fewer keys

Chording keyboards allow typing speeds of up to 300 words per minute.

Re:doing it wrong

[teg]

Whether or not the API is "lightweight" is more a testament to whether or not your application/service is trivial and pointless. Arguing for lightweight over functionality and coverage is like arguing for a keyboard with fewer keys.

Huge overdesigned APIs like CORBA begs to differ.

Re:Google stole Java, and everyone knows it

[Lisias]

FUD.

Microsoft's problem was that they contaminated the API with proprietary extensions and still claimed it to be JAVA Standard compatible. Nothing more, nothing less.

Google gone another way. They re-implemented the whole thing (see Dalvik).

Oracle is pursuing a new way of litigating probably because Motorola's acquisition by Google armed them with a fearsome portfolio of patents - almost a insurance of mutual destruction.

Go

[Anonymous Coward]

I've been following this case a tad, and I must say that I can't see why hasn't Google just ditched Java and fully adopt Go.

Is it because of a) Go's maturity or b) Go is not suitable for smartphones and the like or c) it would be pretty expensive at this point to revert stuff already in production, etc.

I don't know, but I am most likely missing something here and I guess it would be cool to use Go on a smartphone the way we use Java.

Re:Go

[Anonymous Coward]

I don't know, but I am most likely missing something here and I guess it would be cool to use Go on a smartphone the way we use Java.

Google Go is a much worse language than Java actually. It's error handling is really weak, encouraging programmers to ignore errors, not document them, or handle them when they happen despite not knowing what the caller wants to do with them. Meanwhile it's automatic implementation of interfaces makes dispatch ridiculously slow (~80 cycles to call a Google Go interface method vs ~1.2 cycles in Java). Module versioning is basically non-existent so you have to compile in all the modules your program depends on.

On top of that Google Go has a syntax that looks like crap. Code that properly handles errors is about twice as verbose as C, and Google Go syntax is supposedly lighter. The whole language is shit and if Google decides to use it for Android they'd be competing with hands tied behind their back.

Oracle should know better

[bogaboga]

Oracle says no court has ever found that APIs for software like Java are ineligible for copyright protection.

That's not the issue here. No body ever said APIs for software like Java are ineligible for copyright protection.

There are precedents in matters of law. This could be well be one of them. Oracle should know better.

I even have concerns about Oracle's confusing statements about Java. Java can mean the language, the VM or the libraries associated with it or even the combination of everything I have mentioned.

When Oracle says, "Google Chooses To Base Its Android Platform On Java Without Taking A License", one wonders what they are really talking about. I certainly need no license to use the Java language. Do I?

Re:Oracle should know better

[wierd_w]

If api headers and primatives are "copyrighted", and "derivative works" are prosecuted under copyright law (how do you prove your implementation of Free() is not a derivative work of the oracle java implementation?) Then making use of the language (you DO have a valid license to read and replicate the API calls, right?) could well require draconian contract law to undertake.

What I would see come out of this is a creative commons based comminuty api set (probably based on other languages) released by the FSF, to combat the "if you want to program, pay up bitch" shitfest that would come about as big companies claim ownership of C, Java, .net, python and pals.

The sad part is I wouldn't put it past companies like Apple to say dumb things like "if you want to develop for our platforms, you have to use our languages and apis, and so also need to enter a contract to license our oh so special IP." On top of the "pay us 30% off the gross of your appstore sales" shenannigans.

If they could get away with it, I don't see why they wouldn't enforce it.

On the flipside, since linux kernel apis have already effectively been granted gratis use by Linus and the linux foundation, such moves might spell utterly desasterous for the companies that try it.

Embrace Extend Extinguish

[aralin]

Google did to Java on Phones exactly what we criticized Microsoft for through all those years of Slashdot existence. They took the API, they partially implemented them, then made their own incompatibilities, then took over the market with their incompatible implementation. Do No Evil? I don't think so.

This is the exact same type of case as Netscape vs IE. And they will probably get away with it. Sun wanted a very modest license for Java and Google certainly could afford to pay, but they are comprised of so many hackers that think everything should be open source and free and if they can re-implement it, then there is no harm done. But clearly that is not the case as we plainly saw when Microsoft destroyed Netscape and took over the market. The same is happening now with Google, Palm is gone, now WebOS is gone too, Symbian is as good as dead, Blackberry OS will soon follow and Windows on Mobile is laughable and if it wouldn't be for Apple, then Android would already be the only game in town. There used to be a whole set of Java based phones and Java apps on a variety of mobile OSes and they are all gone thanks to this shameful E.E.E. tactics.

I don't understand how can we still cheer Google even though they are clearly in the wrong here.

Re:Embrace Extend Extinguish

[gknoy]

It's slightly different. Google didn't market it as Google Java, and in fact took pains to say that "the syntax looks like Java but it is compiled to run on the Dalvik VM".

Re:Embrace Extend Extinguish

[JAlexoi]

Do even know what the issue with Sun vs Microsoft was? Misrepresentation and inappropriate use of Java trademark. Microsoft called their version of Java - Java. That was the main issue, although their Java continues existence in the form of J++.

Do No Evil? I don't think so.

FYI: Google's moto is "Don't be evil" not "Do no evil".

Re:Embrace Extend Extinguish

[Billly Gates]

How?

Google used a free VM and an api from Apache. They were all freely available. Now Oracle is suing them for using a product that they do not even own?? Of course they are going to create their own API to access parts of the phone. Every programmer writes one to do things in their program and I do not see how what Google did was different.

Now if they made a Java like langauge and called it Java but was purposedly incompatible to force users to use Google's java that would be different. That is not the case

Re:Embrace Extend Extinguish

[aralin]

Nice to see that so many moderators still think that down vote is a disagree button. :)

Re:Embrace Extend Extinguish

[Briareos]

FYI: Google's moto is "Don't be evil" not "Do no evil".

Actually Google's "moto" is "rola"...

Re:Embrace Extend Extinguish

[owlstead]

"...although their Java continues existence in the form of J++."

For some meanings of "continues existence". Basically, if you ever thought that would keep seeing the light of day, then you're now truly borked.

Re:Embrace Extend Extinguish

[Anonymous Coward]

Google doesn't provide a compiler. Their tool takes a finished Java application and transforms it so that it can be run on Android.

Re:Embrace Extend Extinguish

[shugah]

Are you sure about this? I don't think this is true.

Google provides an Eclipse plug-in that includes a Dalvik bytecode compiler. This is the crux of how Google got around Sun/Oracle's IP restrictions on Java for mobile platforms. I'm not an Android developer, but from what I can tell, the Google Android tool chain looks like:

[Eclipse Android SDK] -> [Java Source Code] -> [Google's fork of the open source Apache Harmony Java SE class libraries] -> [Eclipse Dalvik bytecode compiler] -> [Dalvik bytecode] -> [Dalvik JIT compiler] -> [Dalvik VM] -> [Android platform native code]

The stuff in bold is entirely Google's invention, the stuff not in bold is open sourced. Oracle is basically making 2 claims:

- A copyright claim, that Google's Apache Harmony based Java SE libraries include Sun/Oracle copyrighted content. I doubt that this will succeed because the "copied" code consists of header files containing little more than function prototypes and definitions that are required to implement or even write to the API. The Java SE API is itself pretty much a copy of AT&T's C/C++ API.

- A patent infringement claim, that Google's Dalvik VM infringes upon Sun/Oracle's "invention" of the Java virtual machine. The problem here is that Sun didn't invent virtual machine technology. IBM's CP-40 VM-CMS dates back to the late 60's. Alan Kay and Dan Ingalls' (PARC) Smalltalk-80 was an object oriented programing language that compiled to bytecode which in turn is run through a JIT compiler and or interpreted by a virtual machine on each host platform. Baan, the German ERP vendor's bshell was a VM that ran the Baan 4GL language and dates to the late 70's. The "inventions" that Oracle is asserting with respect to the VM technology in its patents are common programming techniques as ubiquitous as the Unix fork() implementation.

Wait a minute:

[bogaboga]

Google did to Java on Phones exactly what we criticized Microsoft for through all those years of Slashdot existence.

Not exactly: -

First, Google made the code open source.
Second, Google never proclaimed JavaVM compatibility.

That's a far cry from Microsoft's behaviors.

They took the API, they partially implemented them, then made their own incompatibilities, then took over the market with their incompatible implementation.

While I agree with this, it is not illegal. After all, everything was open source according to SUN.

Next please.

Re:Wait a minute:

[aralin]

And again the magic word open source as if it would solve all the problems and bring world peace.

The fact remains that Davlik does not run Java programs properly, it is intentionally incompatible. The very reason why Java license only allows full implementations of the entire API, so this fragmentation would not happen. Open sourcing it does not change it one bit, since there is nothing that can be done about it.

And actually it is illegal. Same argument as we always use with GPL applies. You don't have a copyright unless granted by the Java license, if you break the license you don't have any rights to the copyrighted material.

That is the essence of the case.

Re:Wait a minute:

[bogaboga]

The fact remains that Davlik does not run Java programs properly, it is intentionally incompatible.

Does it have to? Not at all. Dalvik is NOT Java. Not in the least, and Google has never claimed it is. Get the distinction?

The very reason why Java license only allows full implementations of the entire API, so this fragmentation would not happen.

So that means that if I do not want 'full implementation' of Java, I do not need to get a license...or do I? Please tell me. I agree with your assertion in the last sentence about fragmentation though.

Authors of C# never called it Java at all, though it borrows many bits and pieces from Java. Have a look for yourself.

Tell me: Should Oracle sue C# authors over these class names as well?

Re:Wait a minute:

[aralin]

From developer.android.com the first sentence of Application Fundamentals: "Android applications are written in the Java programming language." So much for Google never claiming ...

Second you only get a license for Java if you do a full implementation. Otherwise you don't get one. And if you still go ahead you are in violation of Sun' IP.

C# is completely different case. If Google came with their own language instead of using Java, even if it was similar and used similar syntax or class names, I don't think there would be any legal action even though it would violate some of the patents probably.

But this blatant case of attack on Java while trying to piggyback on the infrastructure, tools and libraries developed for Java at the same time is shameful and we should call Google on it instead of trying to defend them. That is just my opinion though.

Re:Wait a minute:

[bogaboga]

..."Android applications are written in the Java programming language."...

This is not in dispute at all. They could also be written in C, C++, Python and a number of other languages. There is no legal problem in writing an app in the Java language (read "using Java syntax and semantics").

Second you only get a license for Java if you do a full implementation. Otherwise you don't get one. And if you still go ahead you are in violation of Sun' IP.

Not exactly, this is probably what you meant: -

You only get a license for Java (meaning the VM), if you do a full implementation of the Java VM. Otherwise you don't get one. And if you still go ahead you are in violation of Sun' IP.

Google does not implement the Java VM. Google only uses it to produce bytecode which it then 'translates' to Dalvik bytecode.

I repeat: Google does not implement the Java VM and nothing binds Google to implementing the Java VM afer using the Java language initially, which as mentioned earlier, is legal.

Re:Wait a minute:

[aralin]

No, that is not what I meant and thank you for not puting words in my mouth. This is what both you and Google are mistaken about. That Java is just the VM. But it's also the libraries and APIs and other parts and there is where Google is in violation of IP. But even in implementing Davlik, although it is not a Java VM, they are in violation of a bunch of patents.

But again, as I said, if they did their own language with a Davlik VM and maybe even very similar syntax to Java and their own APIs and Libraries, there wouldn't likely be a lawsuit. But they made everyone thing they can write Java apps and run them as Andriod apps and then only implemeneted the API / Libraries partially, thus fragmenting the language and then run out with the substandart implementation and through Android market share pushed out real Java out of the market and took over Java developers and as Oracle claims: "Sucked the air out of the market."

All that was illegal, it was exactly embrace, extend and extinguish tactics. Deplorable.

Re:Wait a minute:

[shugah]

I don't see how using a particular language syntax ties one to using a particular compiler and class libraries. If I write a program is C++ I can use gcc/glibc/libc or MS Visual C++/MFC, or VxWorks, gcc and WindRiver's versions of glibc, or Borland/Embarcadero C++Builder, or ... Writing a program using Java syntax doesn't mean you have to use Sun's entire tool chain.

Re:Wait a minute:

[aralin]

The more you talk the more it is obvious you don't even understand Java. So here it is in nice big slides so even you can comprehend: http://jcp.org/aboutJava/communityprocess/ec-public/materials/2011-02-15/Compatibility.pdf

Re:Wait a minute:

[bogaboga]

I do not know how best to say this...

First, from the PDF document you link to:

IP rights are granted only to those who create compatible implementations.

==>This is a non starter because Google has never claimed that it implements Java. It only employs the Java language together with its associated compiler. These 2 things are legal.

Google does not need to implement compatible implementations of Java. In fact, it does not. It is not necessary, though it would be good to, ( I hear you).

One does not legally need a license to implement a spec. In the US of A, specs cannot be copyrighted because specifications, binary interfaces of a library are regarded as statements of fact, rather than a creative works. Copyright does not apply to statements of fact. Period. This is Law.

There is one important thing for you to consider, thus:
 
"One of the fundamental tenets of copyright law is that protection extends only to the author's original expression and not to the ideas embodied in that expression. 17 U.S.C. Sec. 102(b) codifies this tenet by denying copyright to ideas, concepts and principles.

So Google or even you, can give a function the same name as one that is found in Java or any other language. You may then use your own logic to code to the spec. With Oracle's logic, they will come after you for patent and copyright infringement.

Good luck with that!

Trial by battle!

[kakyoin01]

Personally, I am of the belief that this should be settled via some intense Quake 3 tourney play. Just claim it's a match over the JScrollPane class, or something.

Re:Trial by battle!

[Anonymous Coward]

See, you're on to something good, but you screw it up by choosing a video game.

Trial by Battle is an EXCELLENT idea! All the RIAA/MPAA/BSA execs, together with their lawyers (at least those that didn't repent), on one side of the battlefield, armed with sharpened pencils and using attache cases as shields, the pirates on the other end, armed with whatever IP-infringing armaments and defenses they can come up with. Then let the battle commence.

Re:Trial by battle!

[wierd_w]

Excellent.

Putting a 1/4 inch wide groove concentric to the spindle hole on the underside of a compact disc would do wonders to stabilize it for long distance flight. Ideally on the outer most periphery of the disc surface for maximum stabilization effect. (This is basically how frisbees stay oriented.)

Loading those up with pirated beatles and rolling stones titles, then launching them using a custom built robot at *IAA lawyers and corporate execs could be quite satisfactory.

Does anyone know just how fast you can spin these discs before they fly apart from centrifugal forces?

Additionally, since coopting university resources is a time honored tradition among consumate media pirates, would it be unfair to modulate a torrent seed stream over a 1MW+ femtosecond pulse laser from the physics department? I mean, they need to be able to overcome "interference" over the "transmission medium", and one of the best ways to do that is to increase the broadcast power output....

What about Oracle vs AT&T?

[Dcnjoe60]

If Oracle is proposing that the APIs are copyright protected, then what about their use of the C APIs that they publish with their Oracle database? Would that not be an infringement of AT&T's copyrights? Likewise, Oracle software uses and provides hooks for various Windows APIs. Does that mean Oracle owes royalties to Microsoft?

Android stifles the creation of an oPhone

[Anonymous Coward]

FTFA, Edward Screven, Oracle's Chief Corporate Architect:

"Java, you know, is, in my mind, pretty well locked out of the smartphone market because of Android..."

should read:

"Oracle has never tried to get into the smartphone market before, and we should automatically be as successful as other companies who have already been in the market for years."

Usual misleading headline

[Pop69]

Oracle have asked the court to rule that the copyright claim should remain.

It's a standard reply brief, nothing to see here until the court make a ruling.

and python prospers

[superwiz]

No wonder I see more and more requests for Python instead of Java. Oracle's policy of feuding with everybody including their friends is making Python a more and more attractive alternative.

Re:and python prospers

[Anonymous Coward]

You shouldn't advertise Python then. If it gets too popular then the vultures will start looking at what patents *Python* infringes ;)

Re:and python prospers

[Anonymous Coward]

Good point. I am heavily invested in the use of Java and don't use Python, mostly because of the lack of time to invest in relearning and recoding. Yet, the tact taken by Oracle of trying to nickle and dime an IP monopoly out of every bit and word implemented by and in Java has rapidly stunted its potential as a more universal language. This is most clearly seen in their handling of the Java Community Process, which is now run more like a private prison operation rather than a community. Consequently, I have begun to look more seriously at other platforms and languages to build more open and universally available code.

Ironically, Oracle might well do better financially in the long run if they loose their case, since then the language could be more widely used, even if the entire revenue stream does not wind up in Larry Elison's pocket. Their current strategy seems destined to make Java so closed that it is rapidly shrinking its universal potential. If they win this API is copyrightable argument, it would likely make writing anything in Java subject to potential litigation and open to uncertainty over just how much of a programmer's success will be subject to greedy palms at Oracle Corporation. Too bad Google didn't simply buy Sun when they had a chance. Talk about bad business decisions.

Re:and python prospers

[Anonymous Coward]

As a follow up to my response below, I went to the Python site and looked at the link for Java Programmers considering Python

http://wiki.python.org/moin/MovingToPythonFromOtherLanguages

However, there seems to be a problem with this link.

Anyone have an alternate source for this missive.

Re:Google stole Java, and everyone knows it

[shugah]

The Microsoft - Sun Java case was a contract / licensing dispute not a Patent / Copyright infringement suit. From what I recall, Microsoft licensed Java, implemented their own, polluted run time environment and JVM while continuing to call it Java, which violated the terms of their Java license. Sun won an injunction against MS and eventually settled out of court. MS and then phased out Java in favor of .NET, which from an IP perspective is not all that different than Java, Dalvik or Mono, etc.

In Google's case, they are not propagating a polluted JVM under the Java name in violation of a license. Rather they have created a cross compiler for the Java language that generates Dalvik byte code which is dynamically linked to the Dalvik (Apache Harmony) class libraries and a JIT compiler and Dalvik JVM that execute the Dalvik bytecode on the mobile device. To be clear - Oracle is not asserting that Dalvik is a faulty implementation of Java. Rather they are asserting that the programming constructs, algorithms and "inventions" used in Dalvik were invented by Sun and protected by software patents. Oracle is ALSO making certain copyright claims against Google, but I don't see these progressing very far as they are asserting copyright infringement primarily on test interfaces and APIs.

The Dalvik model is really no different from any other VM based platform. Mono - is an open source implementation of the .NET virtual machine. Grasshopper is a Visual Studio plug-in and compiler that produces Java bytecode so that you can run .NET applications on a Linux/J2EE application server. Parot is a cross platform VM that was originally developed for Perl and Python, but now supports compiler frontends for a variety of languages including Java. IKVM is an implementation of Java over Mono for .NET. Sun at one point was going to develop a JVM for iOS, but didn't go ahead with it. The combinations of IDE, bytecode compiler / converter, class library API emulation or Run Time Environment, Virtual Machine and JIT compliler are endless: C#/Visual basic/CLI on a Dalvik VM, Objective C on .NET, Java on iOS, Perl/Python on Mono/.NET, etc. etc.

A license, yes, but what license?

[robbak]

I strongly doubt that Sun would have given google the kind of open, transferable license that Google needed. Sun always tried to keep strict control over what you could do with"Java". Any license they could get would have left google with constant headache.
When the only thing they would have got for such a license would have been the name Java, why would they?

-1 Incorrect

[maroberts]

Alsup did not deny Googles request - he said it was up to the magistrate judge (who he has handed off discovery issues to) to make a decision. Google and Oracle are still fighting like cats and dogs over the issue.

And the copyright brief is a brief, not a ruling, so this issue is still yet to be decided.

Maybe Slashdot contributors should leave legal reporting to the professionals...

Corruptible

[tepples]

True, it should be. But the U.S. Congress has shown itself to be somewhat more corruptible by for-profit special interests than the courts. Copyright in particular suffers from the entertainment industry's control of the means of reaching the electorate. If we don't want to (ab)use the courts to help clarify the statutes that Congress has enacted into law, I guess we need to fix Congress first.

Isn't it just like a user interface?

[rastoboy29]

I seem to recall something about user interfaces not being copyrightable a la the old Excel/Lotus 1-2-3 case?

Re:Isn't it just like a user interface?

[plankrwf]

Apple begs to differ.
(Allthough I completely agree with you).

A trademark cannot be used as an ersatz copyright

[tepples]

A trademark cannot be used as an ersatz copyright. Sega v. Accolade; Dastar v. Fox.

As for the use of namespaces within org.apache: The name "org.apache.struts2.components" just refers to "components of Struts2 published by ASF", and referring to something is an acceptable use of a trademark. I'd imagine that the use of the Apache mark in an imported namespace doesn't confuse people as to what specific part of a particular program is endorsed by ASF.

• If you say " package org.apache.struts2.whatever" you're saying "this is part of an ASF product" which may mislead readers of the code.
• But if you say " import org.apache.struts2.whatever" you're saying "this is a complementary good to Struts2 and no more endorsed by ASF than any other complementary good".

What about Sony vs. Bleem?

[rollingcalf]

Didn't the court rule that it was legal for Bleem to replicate the Playstation APIs?

Sour Grapes

[jklovanc]

Here is a quote from the article;
'"This case is not about Google creating a compatible platform. It is about Google picking and choosing some Java APIs, but not others, knowing it would create an incompatible platform," it wrote.

Also, Oracle noted that Google itself essentially asserts copyright on its own software. "Notably, Google requires its OEMs to maintain the full set of Android APIs -- including the 37 APIs it copied from Oracle to prevent fragmentation of the Android platform," it wrote. Android's license is similar to Java's, Oracle said.'

So Android implemented the parts of Java that were relevant to mobile devices and ignored the parts that were not. They did this to decrease code size and not as a ploy to create an incompatible platform. They then went on and added APIs that were not covered by Java but were needed for mobile devices. Every Java application does this. If this is an issue they everyone should stop making Java based APIs.

The requirement for OEMS to "maintain the full set of Android APIs" is a condition on being able to call themselves Android. They are not calling it Java and therefore do not need to implement the entire Java API.

Android is not trying to replace Java as a language like Microsoft tried to do. It is just using some APIs as a starting point for a mobile based platform. Oracle is just pissed that Google turned their language into a mobile platform before they did..

Copyright does not protect facts...

[Roogna]

"Copyright does not protect facts, ideas, systems, or methods of operation, although it may protect the way these things are expressed." straight from http://www.copyright.gov/help/faq/faq-general.html

Now I'm not a lawyer, but it was always my understanding that copyright doesn't protect APIs for that reason. An API is really a statement of facts. Now a specific header -might- be covered by copyright and a specific source file (implementation) can definitely be covered by copyright. The API itself though is really a statement of fact to the compiler (And the human). Like a phone book, you can copyright layout and the specifics of a phone book, but you can't actually copyright that "xxx-xxx-xxxx" calls "john doe". Because that is a simply a statement of fact. This is why WINE can even exist. As their implementation of the Win32 APIs is -not- a implementation copy of the Win32 source, it is a implementation that follows the same "truth".

So:
extern double sqrt( double );

Can't really be copyrighted, because it is a literal statement of facts. The implementation may be copyrighted. (in this case the math is ancient, but say a hand tuned assembler might be copyrighted, though a clean room implementation of the api obviously doesn't intrude upon that copyright)

And the java api:

static double sqrt(double a)

isn't violating anyone's copyright by declaring the identical facts for the appropriate language.

Anyway, obviously the courts will have to sort that out. And what the fuck do I know. But if APIs suddenly become copyrightable, then welcome to a world of hurt for everyone who writes software, including Oracle themselves.

French Ariane 5

[G3ckoG33k]

Word! "I have heard Java sux" many times.

Still, the French Ariane 5 engineers did NOT use Java (http://en.wikipedia.org/wiki/Ariane_5_Flight_501#Arithmetic_Overflow).

So, in conclusion, you can crash a rocket without using Java. Impressive!

Oracle wants a war ? Fine

[Anonymous Coward]

I think Google should :
      * un-reference all Oracle related pages from Google Search.
      * Move from Java to another language like Scala or C++ ( hmm, maybe costly , but that's the law of war). Scala would be a good candidate (perhaps)
                      [ moving from Java to Scala could be automated ]
                      http://www.scala-lang.org/
      * start writing (if not already done) a good database software better than Oracle's db and cheaper.
      * hire me for my good ideas (that's for fun)
JPB

 

Scala

[Anonymous Coward]

I'm replying to myself to precise my thoughts about Scala :
        It could be feasible to retarget the scala compiler to Dalvik op codes, so the object file would be executable on Dalvik without much hassle.
        Some Scala classes which rely on Java classes would have to be rewritten.
        By doing so, the Java bytecode wouldn't even be used at all.
        Of course Google could retarget javac to produce Dalvik code, but in this case they'd still rely on Java and Oracle would complain, again and again.

        We all know that Google has probably the best minds in IT: Ken Thompson, James Gosling , Rob Pike , and many other talented persons, so this
        task would not be huge.

        Legalese: Scala is BSD licensed, that's very very good.

JPB