Fedora Desktop future- RedHat moves

[Les Mikesell]

Alan Cox wrote:
>>> That is the heart of the question - are you just using some generic
>>> interface (and US caselaw is mostly against API copyright) or are you
>>> creating a new work which incorporates an existing work.
>> The strange part is that this concept can change after the fact or by 
>> the existence of code you didn't know about. I think the legal system 
>> will have a problem with it.
> 
> I never said that it would change. Whether one work is incorporating
> another doens't change because two copies of the "other" exist. The
> question is much more about whether it is generic.

But that _is_ the FSF claim and the crux of the RIPEM case.  Writing an 
interface-compatible non-GPL'd library (fgmp) as a potential alternative 
made the whole claim that the separate work was 'derived' go away, even 
though the alternative wasn't really ever used.  The case did change 
because of the existence of the other copy gave clear proof that the 
choice to link to the GPL'd version was the end user's choice and the 
calling code couldn't possibly be considered a derived work of one or 
the other.  Of course it was never proven the other way, but that was 
the FSF position which, as usual compared to less restricted code, just 
caused everyone extra trouble.

> Suppose you write a piece of software which is split into a collection of
> modules. At what point do I need your permission to dynamically link
> another program against them.

Never, if you obtained them with the right to use your copy.

> Clearly I do if I just put all of your
> modules together again.

Not if you have the right to use them.

 > Fairly clearly I do if I put most of your program
> back together and tweak a few bits (eg if you write a game and I link all
> of your game back together with a single module change say to let me turn
> the music on or off). At the other end if I use your library
> implementation of something like printf (say you had written your own),
> it would appear I do not, as it is a generic interface.

I believe the right to modify your copy of something has been 
established in at least some cases.  And is explicitly given on GPL code.

> There is a point at which I am taking your creative work and modifying
> it, and a point at which I am just using generic interfaces you provided.
> The challenge is to work out which is which, and even more importantly to
> evolve a set of meaningful tests to make that decision.

Agreed - but the FSF makes assumptions about it.  Personally, I think 
the only purpose of a library is to provide use of its interfaces and 
functions and there is nothing you can do outside of it that constitutes 
a copyright violation.  It would be intriguing to see someone try to 
prove otherwise in a case where the library's own license specifies that 
there are no usage restrictions.

>>> Just as with books - am I creating a book referencing another work or a
>>> book that incorporates chunks of another book, and that depends on
>>> context and isn't entirely clear - see the current Harry Potter lawsuit.
>> Book examples aren't very similar. 
> 
> It seems to be where many lawyers look to understand the equivalences in
> copyright law where there is caselaw, along with things like record
> compilations.

But none of those start with the ability of anyone to obtain their 
licensed copy freely, so there is no equivalence.  You'd have to start 
with the premise that everyone who could possible be the target of the 
claimed derived work's infringing copy already has their own licensed 
copy of the original.  It would be more like providing a custom case for 
your own copy of a record than a compilation that includes the same content.

-- 
   Les Mikesell
    lesmikesell at gmail.com