Google wins over Oracle

[latepaul]

Originally Posted by pwalker8

The obvious solution is that software is not copyrightable. As a programmer, I know that every piece of software is built on top of someone else's software. That's the way programming works, you use common patterns that have been used by everyone for years. It's pretty obvious that copyright was intended for individual works of art. The idea that software was copyrightable was simply a lawyer's construction in an attempt to create a barrier against other people entering their market.

I'm also a programmer. I think software should be copyrightable but algorithms ("common patterns") should not. They should be protected by patents if they contain significant amount of ingenuity, but the bar should be higher than it is at present sadly. Patenting "one-click" buy should not be possible.

[fjtorres]

Originally Posted by pdurrant

I don't really care what happens to the two principles in this case. But if use of APIs require a licence because of copyright, far more people will be affected.

All those PostScript clones will stop. Wine will be no more. there's an amazing amount of software built on the free re-implementations of APIs.

Patent protection would be worse.

Copyrighted material can be legally reverse-engineered via clean-room code development. Patented material, it's not that easy; you can't replicate the same process with different code. Instead you have to achieve the same result via a different process. (Think: Type 1 vs TrueType.)

[latepaul]

Originally Posted by HarryT

In all honesty I can see no reason why an API would not be protected by copyright. Is the design of an API not a creative process? If it is (and it certainly seems to me that it is) then it should automatically be protected by copyright under the Berne Convention.

Does anyone know what the arguments are to the contrary?

I don't know the legal arguments but from a point of principle I don't they should be. APIs are the connecting tissue of different pieces of software that interact. If it becomes impossible to implement a third party API due to copyright restriction then that discourages rather than encourages creativity - which is the intent behind copyright.

[ApK]

Originally Posted by latepaul

If it becomes impossible to implement a third party API due to copyright restriction then that discourages rather than encourages creativity - which is the intent behind copyright.

Well, the purpose of copyright is to encourage that sort of thing BY MEANS OF SEVERELY LIMITING IT FOR A SHORT TIME. So it wouldn't theoretically go against the intent of copyright. Encourage the development of clever and useful APIs by means of ensuring a period of exclusivity, then afterwards it's availble to benefit all. Yay.

But even aside from the fact that copyright has been corrupted into an indefinite profit protection scheme in violation of it's intent and so doesn't even serve it's intended function for traditional material, I still don't think it should apply to APIs because their nature and use is too different to be served well by copyright as it exists (even if it wasn't corrupted).

[theducks]

Patents are for processes (how)
Copyrights are for the embodiment (the specific code)

To issue a copyright for the idea would muddle the murky line even further


Patents were supposed to be for non-obvious processes.
I frequently see patents issued for what was a Manual Process, being patented (or getting a New) because a Computer is now being used or we have a Menu (Lotus vs Borland) at the bottom for control (That one was overturned, but it severely damaged the company)

[pdurrant]

Originally Posted by fjtorres

Patent protection would be worse.

Copyrighted material can be legally reverse-engineered via clean-room code development. Patented material, it's not that easy; you can't replicate the same process with different code. Instead you have to achieve the same result via a different process. (Think: Type 1 vs TrueType.)

There's no way to clean-room reverse engineer an API. There is no code in an API. It's just a specification of the functions provided and the parameters to those functions. You use the API when you're doing clean-room reverse engineering. Allowing copyright over APIs prevents any kind of emulation of software.

[fjtorres]

Originally Posted by pdurrant

There's no way to clean-room reverse engineer an API. There is no code in an API. It's just a specification of the functions provided and the parameters to those functions. You use the API when you're doing clean-room reverse engineering. Allowing copyright over APIs prevents any kind of emulation of software.

Personally, I agree that APIs can't be reverse engineered.
But that was Google's first line of defense in this case. Until Oracle proved they hadn't "clean roomed" it.

That said, I don't think there is any law of man or nature that says anybody is entitled to copy somebody else's creation. Copyright, patent, fair use, are all creations of the legal system. If the system allows it, fine. If it doesn't, then pay up or do without. Don't like it? Talk to the IdiotPoliticians or take it to court.

[fjtorres]

More general: Too many people arguing against IP rights "on principle" arepeople who never created anything worth copyrighting or patenting.

The typical curve on the matter is exemplified by the chinese position: for years, the Chinese government did nothing to protect foreign IP. Once their companies started developing original (and valuable) IP, suddenly China got real interested in protecting IP rights at home and abroad.

The same happened with the Manhattan publishing Houses: they had no respect for foreign copyright when their's had little to no foreign value. Once that changed they became copyright zealots to the extreme.

It's all a function of whose ox is getting gored and who is getting a free ride.
In Oracle vs Google? No clean hands, no heroes.

And, again: the story isn't over. This can still blowup in everybody's faces.
Both sides have deep pockets and despise each other.

[rcentros]

Originally Posted by fjtorres

That said, I don't think there is any law of man or nature that says anybody is entitled to copy somebody else's creation.

So you're saying a phone book (a list of names) should be subject to copyright? That's essentially what an API is. It's not a creative process to compile a list of names or an API. It's a mechanical process.

[pdurrant]

Originally Posted by rcentros

So you're saying a phone book (a list of names) should be subject to copyright? That's essentially what an API is. It's not a creative process to compile a list of names or an API. It's a mechanical process.

To be fair, defining an API is creative. There are innumerable ways that one could be created for any system. Doing it well makes writing software using it easier.

But I still think that applying literary copyright to it doesn't make sense.

[ShellShock]

The API is often the most creative part of software, because it can be the most difficult to get right. The API is the most publically visible part of your code, and once it is released into the wild it can be very difficult to fix it later or make substantial changes, without breaking a whole heap of dependencies. So you better make damn sure you put a lot of your creativity into the API to start with, and try and get it right first time. Thus I think we should be able to copyright APIs. But I am glad Google won this latest battle.

[pwalker8]

Writing the phone book can be a creative process according to some patent/copyright lawyers.

The US Constitution says that one of the enumerated powers of Congress is

"To promote the Progress of Science and useful Arts, by securing for limited Times to Authors and Inventors the exclusive Right to their respective Writings and Discoveries."

An API is neither a discovery nor a writing. Of course, over the years, the meaning of this particular phrase has been stretched beyond all recognition by various lawyers looking to see exactly what they could get away with (answer, a whole lot), but the initial meanings of the words are pretty clear.

You can't patent the concept of a 6 mm bolt nor can you file a patent that covers all wrenches that would fit a 6 mm bolt, though plenty of lawyers have tried. That is the equivalent of copywriting an API. Software simply isn't an invention nor is a work of art, any more than building a specific bridge is an invention or a work of art. They didn't try to copyright or patent the Brooklyn Bridge.

[crane3]

Originally Posted by fjtorres

More general: Too many people arguing against IP rights "on principle" arepeople who never created anything worth copyrighting or patenting.

The typical curve on the matter is exemplified by the chinese position: for years, the Chinese government did nothing to protect foreign IP. Once their companies started developing original (and valuable) IP, suddenly China got real interested in protecting IP rights at home and abroad.

The same happened with the Manhattan publishing Houses: they had no respect for foreign copyright when their's had little to no foreign value. Once that changed they became copyright zealots to the extreme.

It's all a function of whose ox is getting gored and who is getting a free ride.
In Oracle vs Google? No clean hands, no heroes.

And, again: the story isn't over. This can still blowup in everybody's faces.
Both sides have deep pockets and despise each other.

The copyright doesn't apply to book titles which is part of the book; witness the many book titles that are the same. So why isn't book titles part of the copyright if the book is supposed to be "original"?

Any software that is copyrighted should have to prove that it is original & that no one has come up with it or used something the same way in the whole world. How many programmers solved an internal business problem & told the boss that the method should be copyrighted? How many programmers on the planet earth ever thought of even having any part of their code copyrighted when working on their job as per specs?

The copyrights & patents is only given on who applied for the copyright or patent first. Apple should not have gotten free with the mouse just because they used a single button instead of 2 buttons. Wonder if the auto companies have patented the gear shift patterns......Apple disregarded the trademarks of another country's company with the iphone & got away with it as the USA has more guns than Brazil.

[rkomar]

Originally Posted by pdurrant

But I still think that applying literary copyright to it doesn't make sense.

Yes. An API is usually implemented as header files, which are meant to be embedded into other people's code (i.e. to be copied). Making them illegal to copy by default seems to go against the whole point of publishing the APIs.

[ApK]

Originally Posted by rkomar

Yes. An API is usually implemented as header files, which are meant to be embedded into other people's code (i.e. to be copied). Making them illegal to copy by default seems to go against the whole point of publishing the APIs.

That's not a relevant comparison.
The question of whether implementing an identical API should be allowed is, I think, orthogonal to how other systems interface with the API.

As an aside, I'd argue with the "usually" aspect of header files. There are plenty of APIs where the concept of a header file simply wouldn't apply, that have no source component at all, no SDK, barely any documentation....