On Tue, Apr 27, 2021 at 12:36 PM Keith Moore <moore@xxxxxxxxxxxxxxxxxxxx> wrote:
On 4/27/21 12:17 PM, Carsten Bormann wrote:
> I think Keith is recollecting some experience with WGs where the chairs used the WG document process to run the show without much regard to the WG’s opinions.
Yes, I've seen that happen. But I've also seen something subtly
different happen, which is to ask the WG for "consensus" before there's
really been time for participants to have informed opinions, and then
(sometimes) treat that "consensus" as if it were set in stone for the
lifetime of the WG.
I have seen that happen many times without any draft being involved.
"We must have a DPRIV solution in 12 months or it won't be any use, so your ideas can't be considered" - 2014.
"It's too early to consider that" ... "And now it is far too late, the WG has gone in a different direction"
One of the reasons I make damn sure to flag that type of move is precisely because I have seen that type of move being tried in the past. And the people who engage in underhand tactics are usually the same people engaging in whispering campaigns in the bar.
I am still pretty angry about the 'DNS Directorate' and the people wearing 'black helicopter' hats to ridicule those of us pointing out that it's not a consensus process when a WG chair can refer the decisions of the WG to a pocket cabal that takes a year to make up its mind and then decides for the chair.
And the folk who were eager participants in that were being manipulated from start to finish by the very organization they thought they were defeating by keeping DNSSEC pure. I am pretty sure they did the same on IPSEC as well.
If you want to stop a crypto spec going through, all you need to do is to pick one feature that makes deployment impractical and persuade enough people that it is absolutely necessary. So IPSEC didn't work with NAT and was undeployable except as a worse VPN than SSH and DNSSEC deployment was stalled for almost a decade.
Perhaps what we need is not to change the boilerplate in drafts but have a clearer description of what a consensus process is and introduce moves that cause data to be logged in the tracker. So that a WG chair can't say 'we will discuss that later', they have to schedule it as an action that has to be cleared. I would also like the IESG to have a standing rule that any WG predicated on the idea there is too little time to discuss the alternative approaches not be chartered in the first place. If there is too little time for consensus, this is not the place to have the discussion.
