Re: RESENDING - Incremental Deployment of IPv6-only Wi-Fi for IETF Meetings

[Leif Johansson <leifj@xxxxxx>]

Skickat från min iPhone

29 juli 2017 kl. 17:08 skrev Ted Lemon <mellon@xxxxxxxxx>:

Now that is political.

Sorry, Leif.   The deal is that the way you figure out whether something is working is to try it.   If nobody is willing to try it, there's no way to be sure that it's working.   The stack itself works fine, and we think the ops part works too, but in order to know that it works, we have to try it.

I was merely reacting to what I felt was a lopsided analogy between v6 (which has had ample time to get tested outside the IETF) and TLS 1.3. The latter is including *lots* of work on implementation and real-world testing in the WG workstream.

Why is this such an upsetting idea?   Why is it necessary to make someone a villain here?

It isn't. I see no villains.

On Sat, Jul 29, 2017 at 10:47 AM, Leif Johansson <leifj@xxxxxx> wrote:

Skickat från min iPhone

29 juli 2017 kl. 16:01 skrev Ted Lemon <mellon@xxxxxxxxx>:

On Jul 29, 2017, at 8:53 AM, Eric Rescorla <ekr@xxxxxxxx> wrote:

Of course this is true, but it's equally true that the people who want to use v6 only can use a separate network. It's not clear why those of us who just want to get work done need to be inconvenienced in service of a political point.

It's not a political point, Eric.

The point is that we claim that we have produced something that will Just Work for the average user.   If you really think you can't "get work done" on an IPv6-only network with working and functional transition tech, we have a problem.

So one of two things is true: you are right that such a network would not allow IETF attendees to "get work done," or you are wrong.   It would be good to know which of these is actually true.   This is the point of making it the default: so that you have to opt out rather than opting in.   That way we can see how many people opt out, and maybe get some feedback as to why.

Honestly, it's really frustrating to me how prevalent this idea is that getting things to work is someone else's problem.   You're the last person I would have expected to make this case.   Do you think that we should still be running TLS 1.1, and that trying to use TLS 1.3 when it becomes available would be a "political" choice?   That's essentially what you are arguing here, just in a different part of the stack.

Or maybe that part of the stack spent less time making sure stuff worked before declaring mission accomplished.