Side Show In Oracle, Google Patent Fight: Are API's Covered By Copyright?

from the we-may-find-out dept

Mon, Aug 22nd 2011 10:16pm — Mike Masnick

While the main event in the Oracle lawsuit against Google is around patents, there is an interesting copyright sideshow as well, focused on the question of whether or not Oracle can copyright an API. A couple weeks ago, Google filed for summary judgment on this, noting (among other things) that Sun's own CTO had flat out declared that "internet specifications are not protectable under copyright," prior to Oracle buying out Sun. The other key claim is that even if Google copied the API, the accusation has them copying 12 files out of over 50,000, which would qualify for de minimis copying, which is a common defense against minimal copyright infringement.

Oracle has now responded and is arguing that copyright for APIs is perfectly reasonable, claiming that the APIs "contain many original and creative elements." Just as Google quotes a former Sun CTO, Oracle (somewhat snarkily) quotes a current Google employee (and former Sun employee) in noting, "API design is an art, not a science."

As Groklaw notes in the above link, Google probably won't win on the motion for summary judgment on this issue, even if it has a better chance at trial:

Although we don't buy all of Oracle's arguments (most importantly, we don't believe much of what they assert is copyright protected subject matter is, in fact, protected by copyright, such as API's), Oracle has probably done enough in its response to put the issue of copyright infringement before a jury. Of course, the court still needs to rule on Google's summary judgment motion.

Indeed. I have trouble seeing how APIs can be covered by copyright. Oracle's key argument beyond that misleading quote is that creating a good API is "difficult." Difficulty alone does not determine if something is copyrightable, of course. Either way, allowing for copyright claims on APIs seems like a good way to create a lot more problems for important (legal!) things like reverse engineering. Once again, it seems like stupid intellectual property laws may get in the way of important methods for innovation.

Thank you for reading this Techdirt post. With so many things competing for everyone’s attention these days, we really appreciate you giving us your time. We work hard every day to put quality content out there for our community.

Techdirt is one of the few remaining truly independent media outlets. We do not have a giant corporation behind us, and we rely heavily on our community to support us, in an age when advertisers are increasingly uninterested in sponsoring small, independent sites — especially a site like ours that is unwilling to pull punches in its reporting and analysis.

While other websites have resorted to paywalls, registration requirements, and increasingly annoying/intrusive advertising, we have always kept Techdirt open and available to anyone. But in order to continue doing so, we need your support. We offer a variety of ways for our readers to support us, from direct donations to special subscriptions and cool merchandise — and every little bit helps. Thank you.

–The Techdirt Team

Filed Under: apis, copyright
Companies: google, oracle, sun

49 Comments

If you liked this post, you may also be interested in...

Reader Comments

Subscribe: RSS

View by: Time | Thread

Anonymous Coward, 22 Aug 2011 @ 10:19pm

The idea that APIs are covered by copyright reminds me of the arguments decades ago that software interfaces were copyrightable. Visicalc or something like that? Didn't that argument end up going nowhere?
[ link to this | view in thread ]
Anonymous Coward, 22 Aug 2011 @ 10:37pm

Eh, taken out of context
It probably should have been: "Good API design is an art, not a science", which would have immediately disqualified Java anyway ;)

(Java fans: I keed!)
[ link to this | view in thread ]
Lawrence D'Oliveiro, 22 Aug 2011 @ 10:52pm

�Functional� Elements?
Isn�t there something about elements that are needed to reproduce the �function� of something not being copyrightable? That is, you can copyright the code implementing the API, but copyrighting the API would prevent anybody else writing their own code to replicate the same behaviour, which would be an overreach of copyright law.
[ link to this | view in thread ]
Anonymous Coward, 22 Aug 2011 @ 11:07pm

Dear Oracle,

I have written a wrapper class around java that helps me detect idiots across all industries and unlike you r implementation of java, it does not offer garbage collection.

Sincerely yours
An android developer
[ link to this | view in thread ]
Ed C., 22 Aug 2011 @ 11:10pm

How about TOC?
Technically, an API is akin to a table of contents--both only convey the ideas of the content, not the actual expression of the content itself. For instance, you could give two authors the same TOC for a book on the Civil War. Even though both would cover the exact same topics, the actual contents would differ quite a bit. You'd get the same if you gave two programers the same API--two different sets of code that ultimately perform the same function. (Note that neither even have to be in the same language!)

It seems doubtful that an author could claim a copyright on webpage with nothing but the TOC of a book, so I don't see how an API would any different at all.
[ link to this | view in thread ]
Anonymous Coward, 22 Aug 2011 @ 11:13pm

Re: �Functional� Elements?
I agree. If you copyright the API, that means that nobody can implement the same functions.

Copyright is supposed to be for CREATIVE works. API's are not the user interface, where you might have some degree of creativity. They're purely functional. Functional stuff should be covered by patents, if anything.
[ link to this | view in thread ]
cc (profile), 22 Aug 2011 @ 11:13pm

This brings back memories of the SCO lawsuits. They were claiming copyright over POSIX interfaces that were derived from ye olde Unix, weren't they? They thought they owned things like identifier names and other basic API stuff...
[ link to this | view in thread ]
Anonymous Coward, 23 Aug 2011 @ 12:14am

Oracle is the antithesis of innovation.

The only reason they bought Sun was because they were fresh out of anyone who could actually innovate and needed to get a new set of patents and products that they could bully competitors with. It's far cheaper to just buy small, good companies with fantastic R&D sectors to merge with your giant, shitty company with almost no R&D and a massive legal department.
[ link to this | view in thread ]
Doug, 23 Aug 2011 @ 12:40am

API design is creative
API design certainly is creative and requires a high level of talent. It is both an art and a science, just as good architecture or good art requires both artistic and scientific components.

However, not everything that is creative or difficult is legally entitled to copyright protection.
[ link to this | view in thread ]
A Guy, 23 Aug 2011 @ 12:41am

The API probably is copyrightable. However, all Google needs to do to get around it is rewrite the offending files in such a way that they are not exact copies but offer the same functionality.
[ link to this | view in thread ]
Anonymous Coward, 23 Aug 2011 @ 1:15am

I know what my next username will be.
DarkOracle
[ link to this | view in thread ]
Richard (profile), 23 Aug 2011 @ 1:25am

Re:
The API probably is copyrightable. However, all Google needs to do to get around it is rewrite the offending files in such a way that they are not exact copies but offer the same functionality.

The functionality IS the API - if the API is copyrightable then you cannot reproduce the functionality without violating copyright.
[ link to this | view in thread ]
A Guy, 23 Aug 2011 @ 2:29am

Re: Re:
Sure you can. Add a description in comments. Rewrite descriptions in your own words. Any change that is sufficiently transformative should get around the copyright problem.

It's not a patent, so you can change form the without changing the functionality.

The only way it's a copyright violation, is if they took oracle's source and copied it directly. (imho)
[ link to this | view in thread ]
A Guy, 23 Aug 2011 @ 3:07am

Re: Re:
I'll address the table of contents hypothetical to illustrate my point.

Original TOC

Chapter 1 How to kill competition with IP laws

Chapter 2 The importance of organizing a PR campaign to silence and marginalize freetards and the freetarded

Non-infringing copy of the TOC with added commentary

Chapter 1 How to kill competition with IP laws for profit

commentary: Here, the author argues that just because something has been done before, no one else should ever be able to capitalize on it again. The author ignores the negative impacts to the economy and dismisses any scientifically validated criticisms with hyperbolic and emotionally charged rhetoric designed to confuse the reader.

Chapter 2 The importance of organizing a PR campaign to silence and marginalize freetards and the freetarded

commentary: Here, the author implicitly gives credence to the critics stance by refusing to engage in an intellectually honest debate. While these techniques may appeal to some of the basest elements of the debate, it will mostly serve to polarize participants and ensure that no compromise is possible. Further, these tactics will serve to show economists and intellectuals whom influence policy that there is a dearth of evidence to backup your arguments and that they would therefore make poor public policy.
[ link to this | view in thread ]
Lawrence D'Oliveiro, 23 Aug 2011 @ 3:34am

Re: Add a description in comments. Rewrite descriptions in your own words.
Those are �documentation�, not �API�.
[ link to this | view in thread ]
Scooters (profile), 23 Aug 2011 @ 3:42am

WTF?
Google's own APIs are covered under copyright! Just because the company doesn't charge for using them doesn't mean there isn't a copyright applied.

This is just getting stupid.
[ link to this | view in thread ]
Richard (profile), 23 Aug 2011 @ 3:47am

Re: Re: Re:
Not a good example. The point about an API is that it has functionality.

Sure you can describe the functionality in different ways - and copyright would apply to each of those expressions as separate documents - but what ORACLE are claiming is that the functionality itself is copyrightable.

If this were true then it would be impossible to create a compatible program without violating copyright. (and incidentally s/w patents woulkd be unnecessary because copyright would do the same job!
[ link to this | view in thread ]
Richard (profile), 23 Aug 2011 @ 4:04am

Re: WTF?
Google's own APIs are covered under copyright!

The code is - the API isn't.

The point is this - an implementation of an APUI consists of five things:

1) The pattern of available function calls, their parameter lists etc.

2) The names of the functions.

3) The functionality of the functions (as an abstract concept).

4) The documentation describing the functionality of the functions.

5) The actual code.

Clearly 4 and 5 are covered by copyright, which can easily be avoided by a third party by re-writing rather than copying.

The case here centres around whether any of 1-3 can be covered by copyright.

If 2 is covered then it makes life very confusing (imagine if the manual in every different make of car had to have a different name for the brake pedal) but life could carry on.

if 1 or 3 are covered then you cannot create a compatible system without infringeing copyright - and that is a major headache.
[ link to this | view in thread ]
abc gum, 23 Aug 2011 @ 5:01am

Re:
Lets not forget header files and comments.
[ link to this | view in thread ]
CommonSense (profile), 23 Aug 2011 @ 5:12am

An excuse...
"API design is an art, not a science."

That's just an excuse that geeks use when they can't figure something out right on the first time. What they should really say, is that there's more than one way, and there are varying degrees of correctness, but that would make them sound less smart or less valuable.
[ link to this | view in thread ]
A Guy, 23 Aug 2011 @ 5:15am

Re: Re: Re: Re:
The point of a copyright is to protect a form, not a function. If the source code takes a different form, then it is not a copyright violation. It doesn't matter if the compiler throws out the differences or not.

All copyrighted works have a function. That doesn't mean copyright protects functionality.

I believe you are conflating copyright protections and patent protections.
[ link to this | view in thread ]
A Guy, 23 Aug 2011 @ 5:18am

Re: Re: Add a description in comments. Rewrite descriptions in your own words.
I'll try to be clearer. I believe their source code qualifies for copyright protection. Documentation is part of the source code, not the functional elements of the api.
[ link to this | view in thread ]
DannyB (profile), 23 Aug 2011 @ 6:37am

Creating a phone directory is difficult
But that doesn't make it copyrightable.

A court found that a pure compilation of facts is not copyrightable.

Therefore, difficulty alone is not enough to qualify for copyrightable.
[ link to this | view in thread ]
DannyB (profile), 23 Aug 2011 @ 6:39am

Even if, even if, even if an API is copyrightable
Oracle (Sun) put those API's under an open source license.

Therefore, Google has a license to use those allegedly copyrightable API's.
[ link to this | view in thread ]
grumpy (profile), 23 Aug 2011 @ 8:04am

Re: Re: Re: Re:
The point about an API is that it has functionality.

No it doesn't. An API is really the statement "talk to me like this and I'll promise to respond like this" with a bit more details, nothing more. The functionality is in the code hiding behind the API and that is copyrightable. Copyrighting an API is like patenting a business model - insane. Not that that ever stopped anyone...
[ link to this | view in thread ]
grumpy (profile), 23 Aug 2011 @ 8:06am

Re: An excuse...
True. It's like saying "Math is hard, let's go shopping", only with data structures.
[ link to this | view in thread ]
Ed C., 23 Aug 2011 @ 8:30am

Re: Re: Re:
No, adding commentary makes it a synopses. A TOC is purely functional; it's only purpose is to list the chapter titles, and maybe section headings, for a book. Copyright explicitly does not cover "functional" pieces, regardless of whether it would otherwise be considered art. You can't copyright a chair or an apple pie because they are both fundamentally functional--beyond any other artistic value they may have.

You could argue that the chapter titles have creative value, but that value really only exist as part of the content of the book itself, not when taken out of context. The same is absolutely true for an API. The function headers are only useful, especially to a compiler, when they're included with the rest of the underlying code. In fact, an API that hasn't even been tested within an actual working system is of absolutely no functional value either!
[ link to this | view in thread ]
Ed C., 23 Aug 2011 @ 8:35am

Re: Creating a phone directory is difficult
Exactly! I'm not sure how that point slipped my mind.
[ link to this | view in thread ]
Ed C., 23 Aug 2011 @ 8:37am

Re: Re: An excuse...
It's more of an excuse for lawyers to use the work of others to extract value from someone else.
[ link to this | view in thread ]
Anonymous Coward, 23 Aug 2011 @ 10:18am

Re:
As noted in another post, though not in reply to this portion of the thread, I too thought of "But isn't an API just a set of facts, and aren't fact's not valid for copyright?"
[ link to this | view in thread ]
A Guy, 23 Aug 2011 @ 11:38am

Re: Re: Re: Re: Re:
That is the idea of an API. The idea of an API is not copyrightable. The code to a specific API probably is.
[ link to this | view in thread ]
A Guy, 23 Aug 2011 @ 11:44am

Re: Re: Re: Re:
So, adding commentary transformed it from a simple TOC into a synopsis?

If only there were some way to use that to deal with copyright law....

As for the rest of it, I would argue that almost any use of an API would fall squarely within fair use anyway, but that doesn't mean that you cannot copyright it anyway.
[ link to this | view in thread ]
A Guy, 23 Aug 2011 @ 12:20pm

Re: Re:
It's a set of instructions, just like all code.

I'm not arguing that Google did anything wrong. Even if it is copyrightable, I believe that using that portion of the code is probably fair use because it only contains a description of the functions/classes used. However, that doesn't take away one's ability to slap a copyright on it, or almost anything else for that matter.
[ link to this | view in thread ]
John Fenderson (profile), 23 Aug 2011 @ 12:44pm

Re: Re: Re: Add a description in comments. Rewrite descriptions in your own words.
Source code & documentation does indeed qualify for copyright protection. However, an API in neither code nor documentation. It is more like a protocol -- an agreement for how to communicate or interface with with another piece of software. You can copyright the description of the API (the docs), the specific implementation of the API (the code), however the API itself should not be (and I believe it is not) copyrightable.
[ link to this | view in thread ]
Lawrence D'Oliveiro, 23 Aug 2011 @ 2:36pm

Re: I believe their source code qualifies for copyright protection.
But �functional elements�, which have to be written in exactly the same way in order to reproduce the same function, do not.
[ link to this | view in thread ]
Ed C., 23 Aug 2011 @ 3:00pm

Re: Re: Re: Re: Re:
There's no need to evoke "fair use" when using something that can't even be copyrighted to begin with.
[ link to this | view in thread ]
Ed C., 23 Aug 2011 @ 3:02pm

Re: Re: Re: Re: Re: Re:
Well that explains it...you don't really even know what an API is.
[ link to this | view in thread ]
A Guy, 23 Aug 2011 @ 4:58pm

Re: Re: Re: Re: Re: Re: Re:
I know what an API is...

It's the portion of code that describes how to access a programs/devices functionality and usually includes a set of classes/functions, including inputs and outputs, and a description of what those classes/functions do.

I am not a programmer, but I have experience programming in various languages including C, C++, and Java.

I believe you can slap a copyright on almost anything, but that doesn't always provide the leverage many people here think it does. It's just my opinion, but I believe it has merit.
[ link to this | view in thread ]
A Guy, 24 Aug 2011 @ 1:28am

Re: Re: I believe their source code qualifies for copyright protection.
Really, we're arguing about a legal theory that (imho) lacks any real importance. I believe you can copyright the code and documentation, however fair use would severely limit your options to enforce it to almost being non-existent.

You believe that you cannot copyright it to begin with because it would make the legalities of programming slightly more complicated (though not much).

I can see your side of it... You believe that due to the informational nature of the code makes it noncopyrightable.

I believe that there is some creative value in properly naming your classes and variables and writing concise explanations of your work.
[ link to this | view in thread ]
Anonymous Coward, 24 Aug 2011 @ 3:06am

Oracle Foot Shoot
Oracle makes a lot of software which has APIs. If they are going to go legally nasty against anyone using "their" APIs, that would make people hesitant to use anything from Oracle, for fear of a lawsuit. That means Oracle gets to have design loses instead of design wins.

Java got to be popular because, among other things, it came from Sun, a company that people generally trusted. By threatening anybody over APIs, then Oracle is demonstrating that it cannot be trusted. Oracle is shooting itself in the foot.

Oracle needs to think deeply about how legal threats affect other people's perceptions. Then it needs to pull its legal boys under control.
[ link to this | view in thread ]
Ed C., 24 Aug 2011 @ 3:42pm

Re: Re: Re: Re: Re: Re: Re: Re:
An API is really just a set of specifications for the interface of a target system. None of the original source code needs to be given for a programer to use it, and any code included from the original source, like function headers, couldn't run or do anything useful on its own anyway. No, in order to use any API, you still have to have your own code to give/receive any input/output for the target system. Also, any code included would have to be rewritten if you're using a different language anyway.

To put it another way--say you're a craftsman that's hired to build the doors for a new house. You don't need access to the house or any part of the original blueprints to do your job. No, all you need for your doors to "interface" with the house is the measurements for the door frames. The measurements themselves might even have to be rewritten to use them on your equipment anyway--thus none of the original information about the house was directly used in your product. Your arguing that because the blueprints are copyrighted, the purely descriptive measurements of the door frames must also be copyrighted.
[ link to this | view in thread ]
Ed C., 24 Aug 2011 @ 3:45pm

Re: Re: Re: Re: Re: Re: Re: Re: Re:
BTW, I have a degree in software engineering, and have probably used many languages over the years that you never even heard of.
[ link to this | view in thread ]
Ed C., 24 Aug 2011 @ 3:52pm

From what I've heard about Oracle, this sort of anti-competitive shenanigans is nothing new. The only reason they are still in business is legacy vendor lock-in, and possibly a bit of Stockholm Syndrome.
[ link to this | view in thread ]
ike, 25 Aug 2011 @ 1:17pm

The actual documentation and header files can be covered by Copyright, but not the API itself. An API is terms by which parties abide.
[ link to this | view in thread ]
Anonymous Coward, 26 Aug 2011 @ 9:49am

Would standards be possible w/o API copyrights?
Most open standards involve a standards development organization defining a set of protocols that can be used by any licensee to accomplish some function. It is important to recognize that these protocols are essentially so-called API's. Some of these happen internally within a CPU, and some happen over the network. But the technical effect is the same.

Of course there are many many SDOs. Many of them have a free-to-use license. However, one thing an SDO depends upon is enforceability of the standard. If you choose to 'fork' or unilaterally change the standard, you loose your license rights. The goal of course is interoperability.

What is wrong with this? Why do people assume copyright of APIs is proprietary and necessarily bad? It is actually what creates software markets and allows multiple organizations to write their own implementations and compete openly. All this openness happens on a foundation of copyrights and patents.
[ link to this | view in thread ]
Anonymous Coward, 26 Aug 2011 @ 9:56am

Re: Re: Re: Re: Re: Re: Re: Re: Re:
This is a classic example of the difference between API and implementation.

The architects plans represent the API the outline of the house. It is definitely copyrightable.

The plans exploit industry standards, such as door sizes, etc. Each of those standards are usually copyrighted by organizations and made available for general use.

The house itself is an implementation by the contractor. The specific techniques and practices he uses are either again industry standards (protected by copyright by some organization) or his own proprietary techniques.

The key here is that some copyrighted items are intentionally made available for "open" use, and some are not. But in general everything has some element of copyright in it!
[ link to this | view in thread ]
Ed C., 26 Aug 2011 @ 12:07pm

Re: Re: Re: Re: Re: Re: Re: Re: Re: Re:
Not even close. The "blueprints" for software architecture usually includes an outline for the API, but it also includes the internal structures of the system, which are almost never exposed to the outside though the API. You're literally confusing the measurements for door frames with an entire house!

Unlike houses, a system's API is custom built for that system. It may use common data types like null-terminated strings, which are not copyrighted, the actual design is usually specific to that system. (Think of trying to fit a Chevy carburator on a Mustang.) However, as I said before, an API can be described such that none of the copyrighted information has to be copied. For instance, the BIOS for the first PC compatible computers written by reverse engineering the API from the original IBM PC. Even though IBM's code was protected by copyright, the judge ruled against IBM's claim the API was also copyrighted.

Also, the measurements of a door cannot be copyrighted. Functional items such as doors and purely factual information such as measurements are prohibited from copyright.
[ link to this | view in thread ]
Ed C., 26 Aug 2011 @ 12:11pm

Re: Re: Re: Re: Re: Re: Re: Re: Re: Re: Re:
I meant "Unlike door frames...".
[ link to this | view in thread ]
marc, 9 May 2012 @ 5:34pm

patent
it's always interesting to see two big companies get into a patent fight, just leave the small guys out of it
[ link to this | view in thread ]