For Approval: CeCILL

[Wilson, Andrew]

Stéphane Dalmas wrote:
>> 
>> This language, and the equivalent language in 5.3.1 about distribution of
>> unmodified Software, appears to obligate Licensees to provide source code
>> to the world at large, not just to those who received a distribution of the
>> binary from the Licensee, for the entire term of the license.  This is much
>> broader in scope and potentially much longer in duration
>> than the equivalent obligation to provide source code under GPL.
>
> To me, it says that if A does some derivative work that it distributes in
> binary form only, and B receives it, B can request the source code. It does
> not say anything about a C that would not have received the derivative work
> in binary form from A (that C would not be a "future licensee", future
> meaning here licensee of A). Of course, C would be able to receive the source
> code from B if B ever requests it from A. The GPL may be somehow broader as it
> says that A should give a written offer to B to give the source code to any
> third party. Which may sound a bit strange as the offer is made to B but it is
> valid for a third party, and not public. You are of course correct that there
> is no explicit time limitation in CeCILL (compared to 3 years in the GPL).

The "future licensee" terminology is confusing and unnecessary.  One is either
a Licensee (having accepted a copy of the Software under the terms of the
license) or one is not.  If you are a Licensee, then the party from whom you
received the Software either should have given you the source code, or offered
to give you the source code.  This section can be rewritten to convey this
concept in a much more direct fashion.

Please consider placing a time limitation on the obligation to provide sources.
This is a point of pragmatic interest for those of us in large companies
who move people and projects around constantly, and for whom the three year
obligation in GPL is already a significant compliance issue.  An obligation
to provide source in perpetuity is commercially unacceptable, and of dubious
value to the community.  How many downloads would there be today of the
GNU tools circa 1991???

>> This obligation not to "directly or indirectly infringe" any IP of the
>> Holder and/or Contributors is extremely broad, and seemingly not limited
>> to IP as embodied in the Software as received by the Licensee.  
>
> The IPR on the Software is what you would call "copyright" and nothing else
> (as for French law software are not patentable and protected by "droit
> d'auteur").

If "IPR" in this context really means "copyrights in the Software as originally
received by the Licensee," then why not just say so?

Andy Wilson
Intel Open Source Technology Center