[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Bug#176267: ITP: mplayer -- Mplayer is a full-featured audioand video player for UN*X like systems

On Sun, 2003-01-26 at 18:24, Glenn Maynard wrote:
> I don't believe that A's license has any bearing on the determination of
> whether or not B is a derived work of A.  It seems clear that a binary
> linking A and B together is a derived work of both (therefore the GPL
> can place restrictions on it), but it's less clear (to me) whether B's
> source, which only references A (in the form of API calls), is a derived
> work of A.

I'd say no. For example, on some theoretical island, you are typing up
code, and through some astronomical chance, you write code which happens
to conform to the API of library foo, which is licensed under the GPL.

Once you've returned to civilization, you distribute the code under an
"all-rights-reserved" license. This shouldn't be a violation of the
license of foo, right?

Well, how's it different if you read source code or documentation? As
long as you don't copy code directly, but instead apply what you've
read, this should be straightforward - you aren't violating foo's
license. The mere existence of calls to an API whose implementation is
licensed in a certain way shouldn't have anything to do with licensing.

As well, IIRC, the reason that binaries compiled against a given library
are considered derived works is that bits of the library are actually
included in the binary executable itself, even with dynamic linking.

This discussion is getting more and more d-legal-ish, btw.

Reply to: