w3process
w3process copied to clipboard
w3c
Charter Approval
The director-free branch indicates in s 4.3 that the mechanism for approving a charter will now solely be Advisory Committee Review. 5.7.1 defines the success criteria as a W3C decision, which is:
A W3C decision is one where the Team has exercising the role of assessing consensus of the W3C Community after an Advisory Committee review.
Little guidelines are given to the Team, beyond that. Is this really appropriate?
First brought up in #316.
We've got a PR trying to cover this (and related topics) already. Hope you like it: https://github.com/w3c/w3process/pull/646
A W3C decision is determined by the Team on behalf of the W3C community by assessing the consensus of the W3C Community after an Advisory Committee review.
After the review period, the Team determines the appropriate W3C Decision [...]
(emphasis added)
This language is problematic; it puts authority to call consensus in the hands of the Team. While their hands are tied somewhat by the presence of a FO, this langauge gives them significant leeway in what's allowed if there is no FO:
If there is dissent (i.e., there were Formal Objections, at least some of which were sustained) or if there is not consensus because of insufficient support [...]
(emphasis added)
If the determination of sufficient support were able to be independently calculated without any judgement by the Team, that would be one thing; here, they have considerable discretion.
In other places this mechanism is specifically a 5% threshhold -- why is that not used here? It's odd that we're so specific elsewhere but not here.
W3C Decision MUST be one of: The proposal is returned for additional work, with a request to the initiator to improve the proposal [or] The proposal is rejected.
Again, there seems to be considerable discretion here.
The set of pull requests that are setting-up director-free already considerably reduce the amount of discretion given to the Team. Agreed, it's not reduced to zero, but we've experienced a number of cases where that discretion is warranted, so we're not introducing purely mechanical rules. We could spend more time to try and assign that responsibility to someone else, but I don't think that should stop us from making these improvements.
The primary goal here is not "let's revolutionize everything about how we work", but "let's stop pretending we don't know how to opperate without a director". I'm not opposed to considering further steps later, but I do want us to make steady progress.
@frivoal I don't know what to take away from that -- you're making general statements, not addressing the specific issue that I've raised.
@mnot https://github.com/w3c/w3process/pull/646 provides some guidelines to the Team. You said they were not enough and wanted the Team to have more hard-coded rules and less discretion. Florian responded that giving the Team discretion has been useful and is intentional. How is this not responding to your issue?
Also I do want to note that the Team often drafts up policy guidelines for itself, and I think that's a better way to handle this than hard-coding automatic rules such as a 5% threshold into the Process.
It has been suggested that it should be possible for the community to cause an WG closure vote to happen; I suggest that we should similarly make it possible to cause an AC vote to happen on a Charter if the team is declining to present it to the AC. This presumes a Charter text has been written and could be adopted, and the reasons that the team is declining may be practical, so this will take thought. But there should be a 'safety valve' here
The safety valve is already present: if no Charter is approved then when the current Charter expires the group will no longer be chartered. I don't believe anything more is needed.
It has been suggested that it should be possible for the community to cause an WG closure vote to happen
Could you provide a pointer please?
OK, thanks. It looks like there is not yet a summary of the most recent AB meeting, which I understand was earlier this week.
In lieu of anywhere else to register any reaction to the suggestion that the community could cause a WG to close ahead of its active Charter end point, it seems like an incredibly bad idea to me. The idea that a mob of folk uninterested in some particular WG's work get to close down the work in which others are actively participating seems like the cultural opposite of the W3C's historical approach.
The idea that a mob of folk uninterested in some particular WG's work get to close down the work in which others are actively participating seems like the cultural opposite of the W3C's historical approach.
Yeah, but the question is whether W3C's historical approach is fit for the purpose of stewarding the evolving web. While I am not convinced there is a pressing problem that having a way for the community to shut down a WG would solve, it seems like a useful thought experiment.
For example, what if W3C had, amidst the "web3" hype a year or so ago, recruited some of the more egregious Ponzi schemers as members and chartered WGs intended to bolster the credibility of things like proof-of-work blockchains, NFTs, or other stuff that now has an unsavory reputation. It might be appropriate to have a mechanism for the community to say "ooh, bad idea, let's pull the plug now rather than waiting for these WGs to fail." (Or waiting for those WGs to tarnish W3C's reputation with something that ends up as snark fodder on https://web3isgoinggreat.com !)
Still, I'm more interested in tightening the process and culture so currently popular but possibly unsound ideas don't get into chartered WGs until they have been incubated well enough to show that a real community of users and implementers are productively engaged. And not giving them the imprimatur of Recommendation until there are interoperable implementations that solve the problems the group was chartered to solve.
W3C's deliberate process, and possibly inadvertent refusal to jump on various Next Big Thing! bandwagons that went nowhere, may have saved it from needing to close down work that should never have been chartered.
OK, thanks. It looks like there is not yet a summary of the most recent AB meeting, which I understand was earlier this week.
In lieu of anywhere else to register any reaction to the suggestion that the community could cause a WG to close ahead of its active Charter end point, it seems like an incredibly bad idea to me. The idea that a mob of folk uninterested in some particular WG's work get to close down the work in which others are actively participating seems like the cultural opposite of the W3C's historical approach.
Um, Nigel, that's a strawman; you're attacking something that (as far as I know) no-one has said. Please can you also avoid words like "mob of folk uninterested", it comes across as pejorative?
One scenario that concerns us is when there is, from the community point of view, a really problematic WG that most think should be closed, but there are one or more members of the team that like it, and those team members prevent team consensus to close, so even though most in the team and most in the community want it closed, the team doesn't propose it for closure, and since there's no "decision", it's not clear a formal objection can be filed, or an appeal. It's a little too strong for preserving status quo, perhaps.
It seems that there should be a way for it to be brought to the AC for debate and decision. That's all.
that (as far as I know) no-one has said
@dwsinger it was you who said it had been said, in https://github.com/w3c/w3process/issues/651#issuecomment-1333968467 - I accept that I built on that, but it didn't seem like a big step.
It seems that there should be a way for it to be brought to the AC for debate and decision. That's all.
The implication is that between a WG being chartered and that Charter expiring, usually within 2 years, something significant changed that made the WG seem like a bad idea. I'd contend that the first step is to address whatever it is that seems bad about the WG first, before attacking the WG's existence itself.
The second step is, whenever the WG tries to transition a Rec track document associated with whatever bad thing is going on, submit appropriate review feedback and objections. Give the WG a chance to resolve the problems first, and if they can't, they're essentially prevented from publishing anyway.
I'm arguing that we have sufficient safety valves already, and I'd be interested to know if there are scenarios where that is not the case.
that (as far as I know) no-one has said
@dwsinger it was you who said it had been said, in #651 (comment) - I accept that I built on that, but it didn't seem like a big step.
I think converting "It has been suggested that it should be possible for the community to cause an WG closure vote to happen" into "a mob of folk uninterested in some particular WG's work get to close down the work in which others are actively participating" was a very big step indeed.
Is it a big step? It seems like a logical inference to me. Help me see it a different way.
Existence of WG implies some group of people are participating, which implies support. If they weren't, there'd be nothing to worry about. (ok, non-participation for too long possibly is the one good reason to close a WG down, but since doing nothing will have the same impact, it would be pointless busy-work)
Community wanting to close a WG down: "Community" implies "those people not participating", which in almost all cases will outnumber those who are participating. At that point, you have one big group of people trying to close down the work of a smaller group of people. In most walks of life, when you're trying to mind your own business and a big group of people actively tries to stop you, that feels like a mob. I don't mean this is in a pejorative way: it's a reflection of the proposed scenario.
I'd have to agree with David that it was a big step, though I can see how you got there.
In short, given that the ultimate authority here would be the AC, I don't see your concern - if a larger group of AC members believe work should shut down, over a smaller group who believe it should continue, that sounds like a lack of consensus by ANY measure - and the charter would not have been approved if it came up again.
Leaving to one side the worry that the AC might change its mind about whether a WG should exist or not within a period of (usually) 2 years, I agree, that what's being described here is a lack of consensus.
In this situation the baseline state is "WG exists with a Charter" and the proposal is "close the WG" and there is a lack of consensus for the proposal. So the proposal is rejected, and the WG continues, right?
Let's see. "mob" is a pejorative term, implying people who are uninformed and out with pitchforks, which you further emphasize with "uninterested".
Take the possibility that a WG is straying into territory that could bring the W3C moral or even legal problems. "We should simply stop this work before the situation gets worse" could be a very well-informed and considered opinion. Other examples include work that is contrary to our values (e.g. intended to make collection of personal data easier and more widespread).
And yes, we should be able to force a debate in the membership about the work, expose the issues and considerations, and make an informed decision as a community.
Again, we already have mechanisms for dealing with both bad behaviour and technical reports that could bring the W3C into bad repute, including the CEPC, wide review, horizontal review and ultimately AC review.
I'm struggling to see how a WG could manoeuvre itself past all of those checks into a situation where the only feasible resolution would be to close it down ahead of its Charter expiry.
Perhaps we should be exploring where the openings are in the process and working methods of W3C that might allow a "bad actor" to, say, publish something bringing W3C into moral or legal problems, before creating a new sanction.
I think we're seriously off track here. We already have the possibility that the Director can close a group prematurely. We're moving to a member-led consortium, and it seems imprudent to leave this power solely in the hands of the team (who are not the founding Director).
I suspect we should fork this into two issues, separate from this general one: is there a mechanism for bringing a charter to AC vote, if the team does not bring it to vote? Is there a mechanism for the members to vote on closing a WG before charter expiry, if the team does not bring it to a vote?
(Another example might be a WG that is dysfunctional, such that we end up with only one organization in it, chairing it, and driving it, because everyone else gave up and left).
it seems imprudent to leave this power solely in the hands of the team (who are not the founding Director).
+1 to moving the decision away from the team alone.
... but also that the criteria for terminating an activity should be extremely high: complete lack of participation/progress or deliverables that will necessarily fall outside the scope of the W3C.
The mere fact that a majority of members is not interested in (or do not "like") an activity should not be caused to terminate it if it is in scope of the organization and it is marking progress. Open standards organizations have a bad track record at predicting which standards will be ultimately successful.
The mere fact that a majority of members is not interested in (or do not "like") an activity should not be caused to terminate it.
Indeed it should not, but all is good because nobody has suggested majority. The text of the process for closing a WG is an AC Review, and these are very much not majority votes. We're not suggesting adding a alternate majority vote process, but the ability for someone other than the Director/Team to initiate an AC Review.
We already have the possibility that the Director can close a group prematurely.
Yes, in very specific prescribed conditions, rather than by an AC Review or such-like, and not due to some sense of moral or legal danger. And there's an opportunity for an AC Appeal. I would trust the team to recognise if those conditions are met.
We already have the possibility that the Director can close a group prematurely.
Yes, in very specific prescribed conditions, rather than by an AC Review or such-like, and not due to some sense of moral or legal danger. And there's an opportunity for an AC Appeal. I would trust the team to recognise if those conditions are met.
That's the problem. There is not an opportunity for AC Appeal when the Director (or whoever replaces the Director in this function) doesn't do anything, because there is no announced decision to hang it on. And the team works (at least today) by consensus, and consensus favors the status quo, and so one influential team member could cause stasis.
I continue to think that the AC should be able to opine, in a formal way, on closing a working group, if it wishes, even if the team seems unable or unwilling to act.
I continue to think that the AC should be able to opine, in a formal way, on closing a working group, if it wishes, even if the team seems unable or unwilling to act.
That would move the problem of "one member causing stasis" from the Team into the AC, without particularly solving anything, wouldn't it?
I'm being swayed by the argument that we need a mechanism to close groups, without a Director to decide, and without consensus by the Team. The current process notes the Director my close a group when "There are insufficient resources to produce chartered deliverables or to maintain the group, according to priorities established within W3C." I can more easily imagine a situation where a group is marginally useful and making some progress toward a low-priority goal that W3C no longer has the resources to pursue, than a situation where a WG needs to be shut down before it causes serious harm. But in that insufficient resources scenario, It's easy to imagine a Team member blocking consensus that "their" group's mission is low priority.
I don't have a proposal for exactly who should be able to shut down groups. The CEO, accountable to the Board and subject to a potential Appeal to the AC maybe? I can't really imagine a scenario where there would be AC consensus to shut down a WG without the CEO's agreement, but I wouldn't object to putting some AC-initiated mechanism in the Process.
[Ahh, I see we've wandered into #653 which has a resolution -- the AB or TAG as well as the Team/CEO can propose closing groups. SGTM]
I can more easily imagine a situation where a group is marginally useful and making some progress toward a low-priority goal that W3C no longer has the resources to pursue
This is exactly the sort of comment that makes me react strongly against the idea that the AC should have any powers to initiate a WG shut-down at all. Who is defining that utility is "marginal" or that the goal has low-priority, or that W3C does not have the resources, given a situation where the AC already approved the group's Charter? Only people who are either uninterested in the group's goal or would like to pursue a counter-interest outside W3C.
I see this as representing a push towards a less plural, narrower and frankly poorer view of what W3C should be doing than the view that I value: W3C should be able to take a holistic view of all the ways in which people interact using the web, and the technologies that are needed to support them.
If there's going to be an AC-initiated mechanism to begin the process of shutting down a group ahead of its Charter end date, then the potential reasons for allowing that mechanism to be triggered need to be very tightly tied down, much as they are for the Director today.
Perhaps the best middle ground here is to say that the AC should have a way to hold the Team to account, so if someone in the AC perceives that the shut-down criteria are met, but the Team is not taking appropriate action, then at that point they should have an escalation mechanism.
W3C should be able to take a holistic view of all the ways in which people interact using the web, and the technologies that are needed to support them.
We're probably wandering far from the actual issue here, but I see it quite differently: Lots of the ways people are using the web are somewhere between sleazy and criminal; W3C REALLY MUST NOT enable such people in the name of holism / pluralism / decentralization / whatever.