Thanks Mark for this very extensive and thourough analysis of the MS
Sender-ID license issue. I actually saved your email as a Document Worth
Keeping [TM] ;-)
We might want to add that furthermore, this license fails the simple IANAL
test, which means that it prevents individuals or businesses to easily
figure out what they can or cannot or might or might not do or may do
today but possibly not tomorrow with this algorithm, without recourse to
expensive legal analysis by attorneys.
Anyway, I believe your analysis clearly demonstrates that Sender-ID is
doomed as becoming a standard that could be widely accepted and
implemented, and in such a situation, if the MARID WG were to define it as
a standard anyway, it would be a major failure in its primary goal and
role.
I had personally stopped emailing to this ML until the "last call" as it
was obvious to me that all this was leading nowhere.
Mark Shewmaker a écrit :
My non-lawyerly conclusion:
===========================
This patent license is incompatible with the Open Source Definition, the
Free Software Definition, the Debian Free Software Guidelines, and the
GPL and LGPL licenses specifically.
Which you clearly demonstrate.
What this means:
================
The above incompatibility will severely restrict implementation and
deployment of the standard, which is a problem that can only be solved
by:
1. fixing the license to address all its problems (described in
detail below),
2. assuming that the nebulous IPR claims are not valid and thus
ignoring the license requirements completely, or
3. dropping the encumbered IPR from the standard.
We've passed the deadline to fix the license, so (1) is out, and (2) and
(3) are the only potential choices left.
I believe the general consensus here is that the encumbered IPR at issue
is the (obvious) PRA algorithm.
Now even though I personally can't see how the PRA algorithm could
possibly be considered a non-obvious idea, I'm guessing that enough
people and companies will be unwilling to risk their livelyhood on that
stance, which means that fear of lawsuits will probably keep (2) from
being a viable option.
That leaves (3), dropping the algorithm completely.
Dropping that algorithm effectively means going back to Classic-SPF.
I believe that is what must be done.
I fully share this point of view.
Cheers.
--
Michel Bouissou <michel(_at_)bouissou(_dot_)net> OpenPGP ID 0xDDE8AC6E