On 21/10/2015 00:52 am, Ted Hardie wrote:
I have just commented in each of the two drafts calls for adoption,
noting that I support adopting each of them. I did that in part because
I agree with Mirja's note that forcing the wg to make a decision has not
worked previously and in part because I see evidence of collaboration
between the teams and approaches (the format changes in tcpcrypt are
one example and the reviews of tcp-tls another).
Collaboration can be useful, but the question is whether it improves
anything? Not really, marginally, but it's not worth the cost. Both
protocols can be made to work.
I understand Steven's concerns about working to create two long term
solutions, but I believe that much of the issue so far has been because
there has been a fundamental assumption that they are competitors. I
believe that Mirja's note casts them as collaborators, and I believe
that moving forward on that basis is a better bet than having one "win"
now. We don't need to decide on the final output of that collaboration
right now, but I think we do need to move to that basis as soon as possible.
Just my opinion, of course,
To me it comes down to simple game theory.
1. would we get a better result by having two competing or even
cooperative proposals?
No. Both proposals have the technical chops to do a reasonable job.
Both can be made to work.
2. Would we get a slower result by having two proposals?
Yes. There is no doubt at all that we will expend energy and
potentially enter into deadly embrace if we have two proposals. We've
already lost a year.
Even if the two proposals enter a convergence path, it will still be
lost time, probably another year, more likely 2-3. We also have a
serious risk of entering deadly embrace.
Draw it out as a 2 x 2 matrix. In which square do you want to be?
iang
_______________________________________________
Tcpinc mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/tcpinc
