Fedora Freedom and linux-libre

[Les Mikesell]

David Woodhouse wrote:
> 
> You're equivocating.
> 
> The _important_ difference, in this context, is how it's distributed. 

No, it's all about the content.

> The GPL clearly states that there can exist sections of which which
> _are_ independent and separate works in themselves --  but when you
> distribute those _same_ sections as part of a whole which is a work
> based on the GPL'd Program...

And you'll note that there are no technical details about putting things 
in the same box or on the same media or in the same file that define an 
aggregation as a 'whole'.

> The _distribution_ of stuff together is what makes the difference, even
> when some of that 'stuff' would be considered to be a completely
> independent and separate work, when distributed separately.

No, the distribution packaging or method is really unrelated to the 
content.  For example, the FSF considers it a GPL infringement to ship 
non-GPL'd content (even in source form) that an end user must link to a 
GPL'd library to function.  Personally I disagree with that and would be 
shocked to see a court uphold it, but you never know...

> You obviously disagree, but you haven't really explained why.
> 
> Maybe Les' mail is relevant here -- he seems to think that we should
> argue based on what we _want_ to be true, rather than what the evidence
> actually indicates?

I didn't say you 'should' do that.  I just don't see the point of 
arguing for something you don't want.

> Do you believe that copyright law _prevents_ the GPL from making
> requirements about those separate works, in such a way that still lets
> you distribute the GPL'd work without complying with the licence?

The license is the only thing that lets you distribute at all.  I don't 
think anyone argues with that.

> Or do you believe that the GPL does not actually impose the requirements
> it seems to impose in §2?

Aggregation is permitted.

> Perhaps you believe that _all_ forms of
> aggregation can be labelled "mere aggregation on a volume of a storage
> or distribution medium" and thus that the whole of those three
> paragraphs in the licence are just a big no-op?

Do you see something in there that defines how aggregation can and can't 
be done?  At one level, on unix-like systems a disk, a filesystem, and 
and archive are really all files.  So you are just quibbling over 
irrelevant file formatting details until you look at the content.

> Can we submit a
> non-GPL'd driver as a .o file, call it 'mere aggregation' and argue that
> it's not a GPL violation?

Linus was once widely quoted as saying that that it was not a violation 
even if he won't stand behind that statement today.  I probably would 
never have used it for any work had binary drivers been prohibited from 
the start.  Note the exception to the stock GPL in regard to the use of 
interfaces in the Linux license.

> Or is it another example of Les' "we argue what we want to believe,
> regardless of the facts"?

The facts are that aggregation is permitted and the techniques of 
aggregation aren't enumerated to some limited set.  If it were otherwise 
the current aggregation wouldn't exist.

-- 
   Les Mikesell
     lesmikesell at gmail.com