Clarification of who is empowered to advance stages, and when/where

[keithamus]

In the joint WHATWG/CSSWG/OpenUI-CG meeting we discussed whatwg/html#12273. A request to advance this to stage-1 was proposed. We opted to do this, but spent a long time discussing if this was the right venue to advance stages, even though chairs, editors, and at least two implementers were present.

Minutes around here. According to my IRC logs we spent around 10 minutes discussing the stage process. So around 1/6th of the meeting.

I'd like us to clarify (codify?) in the stages document to explicitly provide answers to these questions:

• Who needs to present in order to advance a stage? What is the minimum viable number of people present in order to advance a stage? Editors? Implementers?
• What level of formality is required? Can we only advance in a meeting? Let's say folks in a hallway or non-minuted meeting discussed this an advanced?
• What is the right forum? Is a meeting required? If so is it the WHATNOT meeting?

We should have confidence in the staging process. My worry is that we do not, and spend more time than necessary debating the procedure of staging.

[cwilso]

To parrot some of what I said in the OpenUI/CSSWG/WHATWG meeting that this came up during, I agree these questions should probably have a little more formalization now. We initially created the Stages policy to see if it was a useful tool; I think it absolutely has become one.

Some of these questions are covered in the policy, but with a lot of looseness because we didn't want to formalize something we hadn't tested out. For example, the policy says

Advancing to a new stage requires support for the decision by at least two implementers (via their browser engine representatives), and there should not be any strong implementer objections (per the working mode). This support can also be gathered in any public manner, e.g., GitHub issue comments, triage meetings, etc.

...but doesn't say exactly how that will be managed. In practice, it's been managed by the WHATNOT triage meeting chair (i.e. me). IMO, I think that's a reasonable pattern; Editors can probably make this determination too, but especially for moving to stage 1, I think it's better to make it as easy as possible.

[annevk]

I was a little surprised by this happening outside of WHATNOT. I think some additional structure would probably help and I think it would also be good to move away from making decisions during the meeting and instead have tentative decisions that can be contested within some time period as currently we are somewhat in violation of this rule:

Any decisions made during in-person meetings are to be considered non-binding (tentative and subject to appeal from those not present).

[cwilso]

I don't think we are in violation of that rule, btw - the only result of a decision made in a meeting is to add/remove labels on issues, which can certainly be appealed.