w3process
w3process copied to clipboard
w3c
Ensuring charters fit the mission and values of W3C
Turning @mnot's https://github.com/w3c/w3process/issues/328#issuecomment-1210093948 into a standalone issue: In a W3C without an engaged Director, how do we ensure that chartered work is consistent with W3C's mission and values?
Underlying problem: Initially W3C had a Director to make judgments on how to maintain the conceptual integrity of the w3c's standards system, but now that role is essentially vacant. So we need some mechanism for the community to evaluate such matters at charter time, and encourage any concerns to be explicitly raised and authoritatively resolved during the approval process.
Symptoms: W3C's recent public controversies tend to happen when a spec gets to Proposed Recommendation and only then are are fundamental questions surfaced about its appropriateness for W3C's mission, culture, implicit values, etc. The obvious example is EME https://en.wikipedia.org/wiki/Encrypted_Media_Extensions#Criticism . We can certainly expect future controversies based on privacy and sustainability concerns. Once W3C is launched as a non-profit corporation without even nominal supervision from academic institutions or Director applying an explicit vision and implicit values to approve work at various maturity stages, it will be even more important to ensure such issues are resolved early. Rejecting a spec worked on for years for fundamental philosophical reasons will seem like a betrayal of trust and drive members away. Conversely, not ensuring that W3C's Recommendations align with a coherent mission and set of values will drive away those who engage in W3C hoping to close gaps between the web's "full potential" and current (sometimes ugly) reality.
Possible solutions:
-
Tweak the current / proposed Director-free process to ensure charters get horizontal review to surface issues and the AC is informed of these results when voting on a charter. That seems to be the intent of #328, but as @mnot notes that is "sweeping some signficant problems under the carpet."
-
@mnot's suggestion: Make the TAG (or possibly TAG+AB) "responsible for new charters (much as in the IETF)" . But this will add to the already significant burdens of the TAG, and it's not clear that the current TAG members want the role.
-
My previous suggestion in #457 and embedded is other issues for a Technical Director(ate) that would be explicitly hired/elected/appointed to be responsible for issues at the intersection of architecture and strategy such as the problem here. But that proposal has gotten little support, and raises the obvious question of who exactly W3C could entrust with such responsibility and how to find a qualified independent person W3C could afford to hire. Plus this would require engagement of the Board of Directors who will have their hands full restructuring the current organization.
Perhaps there are other possibilities, or an efficient / workable way of combining some of these options to address the problem.
Also, perhaps this issue must be deferred until there is an authoritative and operational statement of what W3C's mission and values are. This has been an AB "priority" for some time now, and perhaps will get some cycles once a BoD takes over the job of advising and supervising the team on "business" matters. But I suspect discussions of the process for ensuring the values are respected in charters can proceed in parallel with discussions of the content of the values.
It requires AC approval to charter new work. If some stakeholders believe that a charter does not fit the mission and values of W3C, they would object (this has happened in the past). That objection would be resolved as any other objection (e.g. Council). It is not obvious that we need anything more process-wise to address this issue.
The guide documents horizontal reviews ofcharters: https://www.w3.org/Guide/process/charter.html
+1 to Jeff. This is a (major) reason for AC review. Whether the AC can see what comment HR groups, TAG, AB, and (for charters offered by other than the Team) the Team might have, is a separate issue.
So you are comfortable replacing the Director's expert, principled strategic guidance with a political process?
When I was an AC Rep, I was once told by the team that I had "exceeded my quota of formal objections" by voting against too many charters. AC Reps have NO incentive to run a rigorous charter review by the experts in their company, and are actively discouraged from objecting unless a a charter is so toxic it would "drive W3C over a cliff."
I respectfully disagree that this is a sufficient substitute for having someone or some small group empowered to ensure that charters advance W3C's mission, would create standards that maintain the conceptual integrity of the web, and are consistent with explicit values.
If the AC is only arbiter of whether charters fit W3C's mission and values, then drop the whole "formal objection" notion and just make it a (supermajority?) vote. Don't put on AC reps the burden of having to publicly justify and defend a NO vote (and endure social media attacks from promoters), just let them exercise their judgment and tally up the collective decision.
Mike, in a member-led consortium, for that is what we are and formally becoming, it's the members' responsibility. We can't duck it, and we shouldn't delegate it either. We elect the TAG and AB to look at these, and if it takes a formal objection, and a community debate, and a formal decision from the Council, that's good. Forcing us to measure questions against values, and not delegating that wrestling, is good for the community.
The Revising W3C Process CG just discussed Ensuring charters fit the mission and values of W3C.
The full IRC log of that discussion
<fantasai> Subtopic: Ensuring charters fit the mission and values of W3C<fantasai> plh: I'm not sure we can do a lot in this iteration of the Process
<wseltzer> +1 to defer
<fantasai> ... at minimum I expect us to defer for now
<plh> ack plh
<plh> ack dsinger
<Zakim> dsinger, you wanted to suggest that we say that individual council members MAY report on their own participation (but not others) and that absence should be recorded
<plh> ack jeff
<florian> q+
<Zakim> jeff, you wanted to comment on #620
<fantasai> jeff: I provided some logic in a post last night
<fantasai> ... which if you want could be used to close the issue
<fantasai> ... I think it's important that charters fit the mission and values of W3C
<fantasai> ... I believe the process that we have is sufficient to take care of that
<fantasai> ... and I think trying to go through some mega-definition of the mission and values and create a new specific process under that
<fantasai> ... would result in a very long process text
<fantasai> ... which would not serve a lot of purpose
<fantasai> ... so I won't argue strongly for that, it's just my input
<fantasai> github: https://github.com/w3c/w3process/issues/620
<plh> ack florian
<dsinger> +1 to Jeff. We have HAD charters where the AC has had a lively discussion over whether something is in scope, fits mission and values
<fantasai> florian: I partly agree with Jeff, but only partly
<fantasai> ... I agree that we fundamentally have the right framework
<fantasai> ... and that the AC approving a charter is the thing that ensures it serves the mission of W3C
<fantasai> ... but as mentioned in other issues, what happens prior to AC review needs improvement
<jeff> q+ to agree with Florian
<fantasai> ... Horizontal review, and review from AB and TAG, there's a number of areas where we could improve the pre-WBS review of charters
<fantasai> ... and the info that goes into that review
<fantasai> ... and that will strengthen the fundamental framework that the AC ultimately decides the fitness of the charter
<fantasai> ... whether we close this now or later, I'm unsure
<plh> ack jeff
<Zakim> jeff, you wanted to agree with Florian
<plh> q+
<fantasai> jeff: I agree with Florian
<fantasai> ... we have other issues raised against specific things like TAG review, and much cleaner to deal with specific issues like "TAG shoudl review or not" rather than intergalactic issues like this one
<dsinger> (I also agree that TAG and AB should be given opportunity to comment, and the usual HR groups, and for charters offered by other than Team, the team too, and that all such comments should be evident in AC review)
<plh> ack plh
<plh> https://www.w3.org/Guide/process/charter.html
<fantasai> plh: Process for horizontal review of charters is in /Guide right now
<fantasai> ... we have horizontal Team people who are in charge of ensuring it gets done, but not how it gets done
<fantasai> ... e.g. idk if Yves goes to the TAG
<fantasai> ... We definitely don't have anyone from AB, and if we want to put AB in the loop could do that
<plh> ack fantasai
<Zakim> fantasai, you wanted to +1 jeff and florian
<fantasai> ... Maybe should improve /Guide and link to that part of the /Guide from the Process
<plh> fantasai: agreed with Jeff and Florian.
<florian> fantasai: I agree with Jeff and Florian, this is something to work on in specific issues, and there's nothing to do in this issue
<fantasai> i/fantasai/scribenick: florian
<fantasai> scribenick: fantasai
<fantasai> plh: So I'll put Propose to Close flag on it
<florian> q+
<fantasai> plh: wseltzer, anything you want to look at from /Guide perspective?
<fantasai> wseltzer: I always welcome suggestions to improve the /Guide
<fantasai> florian: I think this meta-issue is not useful
<fantasai> ... but the individual issues are
<wseltzer> q+
<fantasai> ... and see that there is appetite for additional improvements from the Membership
<plh> ack florian
<fantasai> ... some in /Guide and some in /Process
<plh> ack wseltzer
<fantasai> wseltzer: Then I should move to quickly update the /Guide
<fantasai> plh: I did put Propose to Close on the issue, and will assign the issue to wseltzer
<fantasai> florian: Do it in /Guide or Process is one question
<fantasai> ... another question is do we do it with one large issue that talks about everything, or in individual issues for specific improvements
<fantasai> ... I prefer the latter
<fantasai> ... not prepared to say today that all the improvements belong in /Guide
<fantasai> jeff: Could close by saying that those issue that should be Process issues, people should raise individual issues and raise TAG thing as an example
<fantasai> ... and what goes into /Guide raise against /Guide
<fantasai> plh: so we should treat as a more granular level
<fantasai> plh: perhaps Florian we should sit down and split up the issue?
<fantasai> florian: we have some already
<dsinger> q+ to wonder why we are not proposing to close this?
<fantasai> plh: if conclusion is we already have issue and don't need to open more
<dsinger> q-
<fantasai> florian: I think we should propose to close this, and work on the individual issues, and if we're not happy with the result we can work on it some more
<fantasai> dsinger: Specific comment is that AC is supposed to review for scope mission and values
<fantasai> ... that's one of the reasons for AC review
<fantasai> plh: [missed]
<fantasai> github: https://github.com/w3c/w3process/issues/620
<fantasai> plh: Ok I'll comment that we will close in 4 weeks, and ppl should open new issues if they want something specific addressed
<fantasai> Topic Any Other Business
<fantasai> s/Topic/Topic:/
<fantasai> plh: need to check if got AB feedback on AB-feedback issues
<fantasai> plh: In the Agenda I split issues into Needs Proposed PR
<fantasai> ... have a bunch of such issues
<jeff> q+ to comment on #497
<fantasai> ... so starting in October I will try to assign those to make sure we have ppl who will take responsibility for writing a PR for those things
<fantasai> ... and we do have a bunch of remaining DF issues
<fantasai> florian: I think this is going to get easier soon, because we've solved quite a few related PRs
<fantasai> ... both for issues that need PRs and those that don't
<fantasai> ... we need to do a round of triage, and for half the issues we've already solved
<fantasai> ... many are overlapping, or stale
<fantasai> ... so we will need to do that kind of triage soon
<fantasai> plh: Feel free to add proposed to Close for anything already resolved
<plh> ack jeff
<Zakim> jeff, you wanted to comment on #497
<fantasai> jeff: I had raised 497 anonymous formal objections before dsinger raised 618 ability to defer decisions
<fantasai> ... I think we struggled with handling anonymous objections
<fantasai> ... given we have deferral available
<florian> q?
<florian> q+
<fantasai> ... wanted to get reactions to [missed]
<plh> ack florian
<fantasai> florian: I think this is thought-provoking and interesting, but not ovious that it's right or wrong
<fantasai> ... but for Council to be able to decide to delegate a decision, they need to know what it's about
<fantasai> ... so you still need to give some summary of what the objection iss
<dsinger> q+ to agree with Jeff and Florian
<fantasai> ... Council might decide to handle the decision with the Team as proxy
<fantasai> jeff: Not saying all anonymous FOs need to be deferred
<fantasai> ... in some cases maybe okay to proceed with anonymity and full council
<fantasai> ... but it's an interesting tool to have, ability to delegate
<plh> ack dsinger
<Zakim> dsinger, you wanted to agree with Jeff and Florian
<fantasai> dsinger: I find myself agreeing with both of use
<fantasai> ... it's a good use case to mention for deferral
<fantasai> ... e.g. Council realizies it can't process this and respect anonymity, and delegates to some unit
<fantasai> ... but in many cases can address question without compromising anonymity
<fantasai> ... So to 618, do we add possible reason to defer as "can't process this while respecting anonymity"
<fantasai> plh: okay, I'll put a comment on 618 and then propose to close 497
<fantasai> ... and ppl can check
<plh> q?
<plh> fantasai: we'll do triage of issues after TPAC
<fantasai> dsinger: if anyone really wants to join, we'll be here in SF, but it'll just be a working session
<fantasai> plh: if you need me more than welcome to ask to join remotely
<fantasai> plh: next meeting is Thursday after TPAC
<fantasai> ... we should process any feedback from TPAC
<fantasai> Meeting closed.
<fantasai> s/Subtopic: 623 Recusal/Topic: 623 Recusal/
<fantasai> i/florian: In the middle of/Subtopic: Process Text vs Guide
<fantasai> i/florian: going down the list/Subtopic: Strong Pre-existing Opinion/
<fantasai> i/florian: Next question in/Subtopic: Contesting Reasons to Dismiss/
@dwsinger I agree with the "members responsibility" point. But the FO process is an anachronism unless there is some Director (Directorate, Technical Director, whatever) to appeal to. The Council punted the most recent "hard" question about DID to the Founding Director (so I hear), and that's not actually in the Process yet. So I for one have little confidence that the Council mechanism will work for hard strategic issues.
So given that there's no interest in a Technical Director / Directorate, the way forward seems to be to accept that these are member decisions. But don't bias the process against NO votes by forcing them to be "formal objections" to the institutional memory of a Director that must meet a high conceptual bar ("run off a cliff") and force the NO voter to fight an almost certain-to-lose public battle to overturn the supposed :"consensus" .
Just make it a secret ballot vote by the AC to approve a charter or not.
Just make it a secret ballot vote by the AC to approve a charter or not.
I don't think that a secret ballot vote by the AC is an effective means to determine if a charter fits the mission and values of W3C.
The current process creates a MASSIVE bias in favor of approving charters. It takes zero thought and effort to vote yes (and not be publicly accountable if the charter creates problems down the road), but a willingness to go public (and be ready for abuse on social media) , craft a compelling argument, and defend it repeatedly in order to formally object.
@dwsinger there are many tools in the governance toolbox -- resorting to straight voting by the AC is only one option (and it has some serious downsides, as explained). Many successful systems use bodies that are both accountable to voters and invested in finding the right long-term solution (rather than just the most popular/expedient one).
I was once told by the team that I had "exceeded my quota of formal objections" by voting against too many charters
I find that outrageous. If you filed frivolous objections, they should be dismissed because they're frivolous, even if you only filled one of them. Otherwise, they should be considered seriously, not matter how many you have.
The Council punted the most recent "hard" question about DID to the Founding Director (so I hear)
No, it did not.
On January 28th, the AB said it was happy to take it, but that given the rest of its pipeline, it expected a delay of a few weeks, and said that if the Team did not find that acceptable, they had the ability to do it themselves under the current Process[1]. The team decided to do it themselves, as they felt it was urgent and could not wait. [2] (The Team/Director rendered their decision on June 30 [3].)
[1] https://lists.w3.org/Archives/Member/w3c-ac-members/2022JanMar/0021.html [2] https://lists.w3.org/Archives/Member/w3c-ac-members/2022JanMar/0024.html [3] https://lists.w3.org/Archives/Public/public-did-wg/2022Jun/0002.html
Although I personally was comfortable with closing this issue - seeing that several people do not agree - I think we need to leave it open.
"What is in scope for W3C" is a pretty common bone of serious contention. HTML, Verifiable Credentials, Payments, Linked Data / Semantic Web have all been argued over long and hard in relatively recent memory. Meanwhile it seems topics that don't involve a major shift in the work of a "heavyweight" W3C member are often uncontroversial, however tangential they appear, if there are a reasonable number of interested members who want to work on them.
It seems reasonable in this organisation, better than many at dealing with controversies and strongly entrenched positions, to start from a presumption that what a "sufficient" proportion of members think is in scope is reasonable work to do, rather than inviting people to take as a mission, suggesting that it probably isn't. In other words, I think what @michaelchampion describes as
zero thought and effort to vote yes (and not be publicly accountable if the charter creates problems down the road), but a willingness to go public (and be ready for abuse on social media) , craft a compelling argument, and defend it repeatedly in order to formally object.
seems reasonable to me. In part because I think it's somewhat inaccurate and somewhat hyperbole. Also in part because I believe that a large number of the "philosophical objections" I have seen over the last couple of decades and more are just ill-disguised efforts to tilt the market in one direction or another or to save someone from having to rework something.
The AC review process is far from perfect, and we are not as efficient as if everything were decided by one very active leader with total control over direction. But even TimBL was never that. While I favour W3C have a formal technical director on staff, i don't think that would change the vital role that AC review - and the subsequent discussion - plays in dealing with these questions when they arise.
a large number of the "philosophical objections" I have seen over the last couple of decades and more are just ill-disguised efforts to tilt the market in one direction or another or to save someone from having to rework something
That's a pretty offensive and sweeping claim. Asking for you to substantiate it would likely make the situation worse, as well, as you'd have to name names, so I won't ask for that either. Please try to imagine that honest people may sometimes reasonably disagree and we do not need to make such accusations, and should not. Thank you.
Building on https://github.com/w3c/w3process/issues/620#issuecomment-1240121831, it is critical, post-Director, to clearly state in the W3C Process what the mission, values and scope of the W3C are, so that folks have a basis for their objection (and their input in general) -- otherwise it will be just one opinion against another.
As a reminder: "The W3C Process Document describes the organizational structure of the W3C and processes, responsibilities and functions that enable W3C to accomplish its mission." The Process isn't the place to define the mission, values, and scope of the W3C. Instead, it should point to where those things are defined. And the mission/scope/goals should come from some combination of AB/TAG/BoD, within their respective scopes and past decisions.
I see #651 and #653 as a better to discuss the issue of how charters are established. The team filters the request for new charters (based on mission/scope/values)in coordination with AB/TAG, and makes decisions that can be objected.
The Process isn't the place to define the mission, values, and scope of the W3C. Instead, it should point to where those things are defined. And the mission/scope/goals should come from some combination of AB/TAG/BoD, within their respective scopes and past decisions.
Fine with me as long as:
(a) text that defines the mission, values, and scope of the W3C is removed from Process (b) new documents are created and referenced from the Process
In the meantime, the Process contains text that defines the mission, values, and scope of the W3C
In the meantime, the Process contains text that defines the mission, values, and scope of the W3C
can you point me to an example of such text in the current or draft Process? Except for the first sentence of the abstract, I can't find any statement on mission/values/scope. The Process already links to https://www.w3.org/Consortium/mission
See screenshot above from https://www.w3.org/2021/Process-20211102/
To be sure, my objective is to make sure that there is a stable and complete set of documents that establish the rules to which members are subject to.
The AC review process is far from perfect, and we are not as efficient as if everything were decided by one very active leader with total control over direction. But even TimBL was never that. While I favour W3C have a formal technical director on staff, i don't think that would change the vital role that AC review - and the subsequent discussion - plays in dealing with these questions when they arise.
Those aren't our only choices, luckily. There are a wide variety of governance structures we can choose from --with huge differences in their properties. The most obvious in this situation are:
- A benevolent dictator (TimBL)
- A petition system involving every member (the AC)
- An elected board/body (the TAG? a new thing?)
The politics in each case work in profoundly different ways.
A benevolent dictator worked for a long time, but I think we all recognise that it has a dependency on finding the right person, which is difficult. It also makes us vulnerable to that person's absence.
A petition system (which is effectively what is currently specified) creates incentives for partiicpants that may not lead to good outcomes. Because so many AC reps are disengaged, I'm concerned that outcomes will be fragmented, factional, and vulnerable to manipulation.
A board, if we get it right, will take ownership of the vision and scope of the organisation, but still gain legitimacy through being elected. They can make the difficult decisions and assure that a strategy is consistently applied, as opposed to the piecemeal approach that's implied by the AC. It can adapt over time as members come and go, but maintain consistency though the overlap.
This is not about replacing the AC as a backstop to assure that members are on board with new work, BTW -- if we had the TAG (for example) inserted into the process, I don't think we'd remove the AC step. It's about whether using the AC as the sole steering function for new work is wise for the future of the Web.
That screenshot doesn't define the mission nor scope of W3C. It certainly touches on values, without trying to define those values either. They are used as a justification for how the Process is being set up since the Process is a reflection of those values. I agree that consensus/quality/fairness being values of W3C needs to be referenced. We can certainly add a link to something like https://github.com/w3c/AB-public/tree/main/Vision (or, better, make sure it's properly integrated on w3.org if that's not the case already).
That screenshot doesn't define the mission nor scope of W3C.
W3C revolves around the standardization of Web technologies
Could the discussion of what / where the mission and scope of the W3C are be split into a separate issue?
Tying this into some discussion that happened on a members' call:
- AC review requires a company to take a position, rather than a person. This often pits company-vs-company, which can discourage open discussion about the merits of a particular question. Open participation as individuals rather than company representatives would help here -- e.g., on a W3C-wide feedback list, rather than one restricted to AC reps.
- The judgement of community sentiment in anything less than a FO is opaque. That creates an incentive to FO, even when it may not be necessary.
I think @mnot raises 2 really important points. These are both about the fact that we have no good way of communicating within the larger community.
Where do we stand in terms of resolving this issue? I like mnot's framing in https://github.com/w3c/w3process/issues/620#issuecomment-1262993087.
A benevolent dictator worked for a long time, but I think we all recognise that it has a dependency on finding the right person, which is difficult. It also makes us vulnerable to that person's absence. A petition system (which is effectively what is currently specified) creates incentives for partiicpants that may not lead to good outcomes. Because so many AC reps are disengaged, I'm concerned that outcomes will be fragmented, factional, and vulnerable to manipulation. A board, if we get it right, will take ownership of the vision and scope of the organisation, but still gain legitimacy through being elected. They can make the difficult decisions and assure that a strategy is consistently applied, as opposed to the piecemeal approach that's implied by the AC. ... This is not about replacing the AC as a backstop to assure that members are on board with new work, BTW -- if we had the TAG (for example) inserted into the process, I don't think we'd remove the AC step. It's about whether using the AC as the sole steering function for new work is wise for the future of the Web.
I'd be comfortable with a resolution along the following lines:
-
The Team have the primary role of crafting/reviewing/mentoring charter development to ensure they align with the mission and values of W3C. BUT they must be more explicitly accountable to the Board for vetting charters against mission/values than they have been in the past. (To be blunt, it is NOT OK to draft charters to promote a team member's pet theories, as a way to recruit new members who don't share core values, etc.)
-
The AB and TAG are explicitly informed about charters in progress and invited to comment, but the charter process isn't blocked on them responding nor do they get an automatic veto. I like at least the metaphor of Green / Yellow / Red lights, with the TAG or AB being able to go on record as saying a charter is OK, worrisome, or dangerous... and having that input be considered by the team and AC as the charter gets final consideration
-
I would prefer that AC charter ballots not be so heavily stacked in favor of approval as they are now. Dissent on principled or pragmatic grounds should be encouraged even if approval would not "drive W3C over a cliff". (This the guidance I got from the team when I was an AC rep dissenting against "too many" charters for their taste.) Either dropping the requirement for formal objections to be made public might help, or (conversely) making both positive and negative votes a matter of public record and future scrutiny might as well.
I would prefer that AC charter ballots not be so heavily stacked in favor of approval as they are now.
I disagree.
the W3C should a priori welcome new projects driven by its members, unless:
- deliverables will necessarily and demonstrably conflict with the mission, scope or process of the W3C
- there are no member resources to conduct the project
- the charter does not meet the contents requirements
Suggestions to charters should of course be encouraged and welcome, but the fact that a member "does not think the project is worthwhile or succeed" or "would do it differently" should not be grounds for not approving the project.
The AB, TAG and board should only get involved if member(s) believe that the charter fails (1), (2) or (3) above.
the W3C should a priori welcome new projects driven by its members, unless:
- deliverables will necessarily and demonstrably conflict with the mission, scope or process of the W3C
Values. Charters should be disapproved, or at least be subject to intense scrutiny, if their aim or likely effect is in direct conflict to our values.