I propose the following as a way out of the current RTCWeb MTI issue.
* The MTI codec is motion JPEG.
* From 2018 on, the MPEG2 codec is also MTI.
The rationale for this is as follows:
* There is no decision process that can arrive at a choice between H.264
and AV8.
Parties that have already purchased access to H.264 are not going to
purchase access for a second codec simply to enable other parties to gain
access to a codec for free. Using AV8 without purchasing access to the
necessary patents would open them up to a major litigation risk and
potential damages awards in the billions of dollars and/or having their
products subject to injunctions without any benefit to themselves.
* The proposed vote is manifestly unfair and subject to legal challenge in
a jurisdiction of the plaintiff's choice.
Taking a vote in the IETF is not going to settle the matter and would only
expose the IETF itself to the risk of litigation. It would not surprise me
if one or more parties applied for a court injunction to prevent a vote.
The eligibility criteria for participating in the vote were only revealed
after the ability to become eligible had passed. Whatever the justification
the ADs and WG chairs might imagine for this approach it is not one that
the courts are likely to be impressed with. With several billion dollars in
potential damages at risk and a certainty of tens of millions of dollars in
legal fees fighting additional IP claims, litigation against the IETF is
certain.
* The WG does not have the right to make decisions for the IETF
The criteria for adopting standards is IETF consensus, not WG consensus.
The proposed mechanism excludes the majority of the IETF membership from
the decision which will inevitably be appealed. So the ultimate decision
body inside the IETF is the IAB.
* The proposed MTI codecs are widely implemented and require minimal or no
additional implementation effort in a Web application.
* The proposed MTI codecs are good enough to permit applications to
interoperate.
* Development of an alternative codec is impractical since any effort would
take much longer than the expiry of the remaining MPEG2 patents and the
resulting specification would be vulnerable to IP claims regardless of the
effort taken to avoid encumbered techniques. There is simply too much
encumbered art. Prior art is only a defense against damages, it does not
protect against the costs of defending a case.
The only codec work which might be of value is to look at the patent pool
and see if a subset of the MPEG2 standard could be defined with a shorter
encumbrance period. I note that one of the longest patent expiries is for a
tagging scheme providing for backward compatibility with MPEG1. But I don't
read patents unless I am being paid to do so and I doubt anyone else does
for fun.