Re: W3C patent license policy question

On 08/02/2011 05:39 PM, Ian Jacobs wrote:
> What does "end up in W3C" mean precisely? Does it mean that a Working Group would start with the IETF text as a contribution?

I'm not familiar with the structure of RTC-Web, but it seems more likely
that the IETF Opus text will end up as a normative reference in a W3C
recommendation.  From the FAQ, I understand that the W3C's policies are
much less strict in this case, and our license is unlikely to be an issue.

>> How can we make sure that our license would be acceptable at the W3C?
> 
> Good question, but I think we need to understand the expectation more before finding the best solution. 

Maybe this question is about how contributors actually license patents in
the W3C.  For example the FAQ
(http://www.w3.org/2003/12/22-pp-faq.html#ownlicense) says:

"Organizations ... may not make individual modifications to or departures
from the licensing requirements".  "each entity ... makes a choice between
two straightforward options: 'we agree' or 'we don't agree' to the terms
of the Patent Policy".

but then also says

"Organizations offering licenses meeting the W3C Royalty-Free requirements
may point to the specific contact information and entity-specific license
terms"

The first part suggests to me that the W3C does not recognize any
entity-specific license terms, and the second says that they _may_ or even
_must_ offer their own licenses (that meet the requirements).

Here's a concrete example: W3C appears to specify that the reciprocal
grant and termination conditions only apply to "essential" patent claims,
and provides some guidelines for determining whether a claim is essential.
 We have been unable to find a satisfactory definition for "essential",
and so our license says:

"If you ... file a Claim for patent infringement against ... an
Implementation ... provided that a Reference Implementation also infringes
the patents asserted in the Claim, then any patent rights granted to you
under this License shall automatically terminate retroactively"

We choose this wording to avoid lawsuits in cases where a non-infringing
alternative exists but is impractical or undesirable.  Arguably, we are
merely operationalizing the word "essential".  (WebM does something similar.)

So what would happen if a W3C member were to offer this as their license
in a working group?  I see 4 options:

1.  W3C doesn't care what their license says, because they have already
automatically offered a license under the W3C Patent Policy (i.e. the
policy is an automatic license).

2.  W3C sees this as an acceptable license.

3.  W3C sees this as an unacceptable license, rejects it, and requires the
member to provide an acceptable license
3a.  ... because the only acceptable license is a copy of the text of the
Patent Policy.

4.  W3C takes no position on whether a member has offered a compliant license.

I'd like to know which of these options is most likely ... especially if
it's 3, so that we don't end up recommending a license that puts people in
conflict with the W3C.

>> P.S. I think there's a bug in the patent policy text.
...
> Your comment associates 5.6 #6 with Summary #3 but those aren't intended to be associated. 
> 
> In light of the above pairings, does it still seem like a bug?

Nope, you're right.  The W3C draws a distinction here between "conditioned
on a grant of a reciprocal RF license" and "may be suspended ... when
licensor is sued".  I find this confusing and possibly redundant, but
that's par for the course in licensing.