So translating header files for bindings is considered a derived work also? And also the GNU guys and the porting of unix sofware to linux, that can't be considered derived work. I guess it comes down to the definition of the word "expressive" like you say, but programming languages are so precise, I wouldn't have thought there is much room for expressiveness.
On Friday, 29 April 2016 10:22:33 UTC+8, Steven G. Johnson wrote: > > On Thursday, April 28, 2016 at 9:21:03 PM UTC-4, 2n wrote: >> >> If you re-write or translate something doesn't it necessarily mean that >> you are no longer using licensed code? that you have full copyrights to the >> new code? >> > > No, typically a translation would be considered a "derived work" of the > original work. For a derived work, the copyright is jointly held by both > you and the original author, i.e. you need permission from both to > distribute (which, in the free/open-source world, you typically get via a > license, in this case the LGPL). > > A re-write is a tricker issue. A "clean room" rewrite, in which the > algorithm is re-implemented by someone who hasn't looked at the original > code, but has only read the mathematical description of the algorithm, is > typically only copyrighted by the new author. However, if you've > rewritten after studying the original code, it may be difficult to avoid > copying "expressive" elements of the original code that are copyrighted. > > (Ultimately, there is no deterministic way to know whether something is a > derived work of something else -- it can only be determined on a > case-by-case basis in a court of law. In practice, people usually assume > that if you've studied someone's code and implement something similar then > it is probably a derived work.) >