[License-review] Submitting CC0 for OSI approval

[Richard Fontana]

On Sat, Feb 18, 2012 at 10:29:58AM -0600, Christopher Allan Webber wrote:
> Richard Fontana <rfontana at redhat.com> writes:
> 
> > On Fri, Feb 17, 2012 at 10:03:37PM +0000, Thorsten Glaser wrote:
> >> Richard Fontana dixit:
> >> 
> >> >I am not sure whether there is precedent for an OSI-approved license
> >> >(at least, one that is, today, widely used) that explicitly reserves
> >> >all patent rights.
[...]

> Does it reserve patent rights, or does it declare that out of scope of
> the license?  

What do you see as the difference?

> The actual bit of the license text says:
> 
> > 4. Limitations and Disclaimers.
> > 
> >   No trademark or patent rights held by Affirmer are waived,
> >   abandoned, surrendered, licensed or otherwise affected by this
> >   document.
> 
> Let's consider that *most* OSI approved licenses actually effectively
> are the same, but without explicit statement.  Patent provisions are out
> of scope of MIT, BSD, and almost every non-apache non-*GPL license tool.

Maybe I am not sure what you mean by "out of scope".

What is the position of Creative Commons on the following question: To
what extent can a CC0 Affirmer sue a CC0 Affirmee for patent
infringement for the mere exercise of "Copyright and Related Rights"
waived/licensed under sections 2 and 3 of CC0? I read 4a as saying
that the CC0 Affirmee is on notice that he or she or it has no defense
based on CC0 itself to a subsequent patent infringement claim. That's
certainly true on the trademark side, and that's entirely
noncontroversial, because of the relative distance of trademark from
copyright. Many FLOSS licenses have clauses saying "this license
doesn't grant you any trademark rights", and I'd argue that those that
don't (like the GPL) do so by common-sense implication.

> What's being said here isn't "there's a patent license over here that
> you need to sign onto" but "this is a public domain document, it's about
> the waiver of copyright, so these other things don't apply".  And as a
> general purpose public domain tool, it makes sense that patent issues
> are not directly attached.

I'm not sure I understand that. It does mean *potentially* that
"there's a patent license over here that you need to sign onto". So if
OSD 7 does not cover the mere possibility of some later-in-time
additional license requirement by the same licensor or its successor,
that's fine. 

My main concern here is actually trying to understand how the OSI
makes its decisions, and, if the OSD is still relevant, what OSD 7
actually means.  It may be that the real reason why OSI should approve
CC0 is that evidence already suggests that CC0 is well-accepted as a
FLOSS license, as Andy Wilson has pointed out.

> To re-quote that statement of the OSD:
> 
> >   The rights attached to the program must apply to all to whom the
> >   program is redistributed without the need for execution of an
> >   additional license by those parties.
> 
> I think the license does do this... again, the purpose of this section
> is to say "these issues are out of scope".  They're out of scope of
> *most* free software licenses also, since they're not mentioned at all.

> So we should reframe this question:
> 
>   "Does explicitly stating that software patents are out of scope of a
>   software license (or public domain tool), as in CC0 1.0, create a
>   greater software patent threat than if it was never mentioned at all,
>   but still remained out of scope, as in MIT and BSD?  And if so, by
>   what mechanism does that happen?"

By possibly absolutely foreclosing an argument of (even perhaps very
narrow) implied patent licensing or the like. The mainstream members
of the MIT and BSD family do not do this.

I think your remarks and Mr. Cowan's have been helpful in clarifying
that analysis of the actual likelihood of a software patent threat (in
general, and from the actual provider of the software under the legal
instrument in question) ought to be relevant when considering whether
a submitted license should be approved as "Open Source" by the OSI,
and in understanding what the OSD means (assuming it's relevant).

I suspect, or at least would like to think, that if Creative Commons
were drafting CC0 today, knowing that CC0 would be enthusiastically
applied to software as well as to nonsoftware content, those two words
"or patent" might not be in the document, and something else might be
there in its place, or at least that inclusion of the "or patent"
language would be accompanied by some satisfying policy rationale
addressing the issue that I have raised.

- Richard