Linked by Thom Holwerda on Wed 2nd May 2012 22:32 UTC, submitted by PLan
Legal "The European Court of Justice ruled on Wednesday that application programming interfaces and other functional characteristics of computer software are not eligible for copyright protection. Users have the right to examine computer software in order to clone its functionality - and vendors cannot override these user rights with a license agreement, the court said." Bravo. A landmark ruling, for sure. If the US courts decide in favour of Oracle in the Google-Oracle case, Europe would instantly become an even friendlier place for technology companies.
Permalink for comment 516862
To read all comments associated with this story, please click here.
RE[7]: NOT APIs
by galvanash on Thu 3rd May 2012 06:26 UTC in reply to "RE[6]: NOT APIs"
galvanash
Member since:
2006-01-25

This clearly describes a header file which contains no code as being non-copyrightable, but perhaps the article was misleading.


I think it just boils down to how you acquire the header file to be honest. Let's say you write a header file for a library containing no functional inline code, nothing but interface declarations and constants. You then slap copyright on it.

If I copy the file and use it I would be violating the copyright. However, since the contents of the file are simply non expressive facts (and you cannot copyright facts) you could argue that copyright doesn't apply. This has always been the case, this ruling doesn't really change anything.

Or you could, instead of copying the code, simply instrument the library directly to discover its interface and then write a header that works the same way. The result would be the same, a functionally identical header, but clearly in the 2nd case you did not "copy" anything.

I think the EU ruling primarily applies to this scenario - i.e. recreating an interface (API) without having actually seen the copyrighted work (the implementation). It isn't new ground - reverse engineering was always legal (if done right) - it just strengthens existing law with another precedent. I think the only thing "new" is the part about not allowing licenses to bar reverse engineering.

However I'm finding the "ideas & principals" argument to be somewhat contradictory with the conclusions.

Strongly typed languages like java store the entire class definition in binary form (java reverse compilers can recreate the source files almost line for line). Thus, the API is fully expressed in the binary. The binary expression of a compatible function/class prototype will very likely need to be identical to the binary expression of the original function/class prototype. So would the court still claim that an API is not copyrightable under those circumstances even when the language mandates the expression of such an API to be recorded the same way each time it is expressed as a binary?


The ruling states you are allowed to instrument and observe a piece of software in order to determine its functional workings, and that those aspects of the work are not covered under copyright. If you did so and documented its interface and behavior and then created an independent implementation you would not be violating copyright, because the copyright doesn't cover the interface or behavior.

Reply Parent Score: 3