AB/Director-free Plan/B1

From W3C Wiki
Jump to: navigation, search

In addition to various specific issues filed against specific parts of the director-free branch of the process, we have a few generic ones.

As we triaged the various issues into various themes, these generic ones were classified (by Jeff & Florian) into a category called B1. We felt it was important to get any potentially overarching problem addressed before we deep dived into the details of various other topics.

Most of these GitHub issues talk about many topics and also are partial duplicates of each other. The intent of this exercise is to reduce the duplication, triage any point that can be addressed individually into a standalone issue, and to discuss and hopefully bring to a conclusion any overarching point.

We are proposing to do that by looking at these broad/generic issues one by one, with some proposals about what to do about them made ahead of time. This is meant to invite discussion while having a clear idea of what a possible outcome may be. Other outcomes can of course be suggested, but the goal is to try and come out of this session with resolutions about what to do, not just to discuss.


The full list is here: https://github.com/w3c/w3process/issues?q=is%3Aopen+is%3Aissue+project%3Aw3c%2Fw3process%2F7+label%3A%22d-f%3A+fundamental%22


https://github.com/w3c/w3process/issues/316 (The Elephant in the Director-free Room)

Summary

Issue raised by Mark Nottingham about "almost all of the powers of the Director are transferred to the Team". Includes:

  • a list of areas assigned to the team, with the commentary that it's a lot
  • a comment on how running s/Director/Team/g is not a good methodology
  • an explanation of how Tim delegating to the team isn't equivalent to the Process assigning directly to the team
  • A suggestion to look at each area one by one and determine what the appropriate decision-making mechanism is

After some meta discussion by a few people about how we arrived where we are today, Mark identifies several sub-topics as needing discussion or action:

  1. Starting Charter development
  2. Approving Group creation
  3. Closing a Group
  4. Requiring a Group to stop work on a specification
  5. Specification advancement (or declining to do so) -- including REC
  6. "workshops should be able to be convened by the Team and the TAG".

Dbaron chimes in and agrees with Mark, in particular about the topics related to charters. Mike Champion also stresses the strategic importance of chartering. Both highlight some potential for conflicts of interest if handled only by the team.

Commentary

s/Director/Team/g is not what the AB did, and looking at each area one by one and determine what the appropriate decision-making mechanism is actually the exercise the AB went through when setting this up.

When looking at the director-free branch at the time this issue was filed, it did look like the vast majority of the word "Director" (and there were many) had been replaced with "Team", which certainly gave the impression of a generic change. Many of them where redundant cross references, or used the Director for emphasis on routine tasks where no decision making was involved. These have now be landed on the main branch, using https://github.com/w3c/w3process/pull/254, so the difference between the main branch and the director-free branch is much more limited. And contains significantly fewer cases of s/Director/Team/. Even then, the result still has quite a lot of things on the Team. It is very possibly that we got some of the balance wrong, and we should revisit specific areas where people have concerns.

To make this more easily actionable, it seems preferable to open individual issues for the various areas of concern, and discuss each problem one by one, rather than continuing to comment in this somewhat meta issue.

Item 6 is a simple but concrete proposal, so:

Proposal 1: Florian to open an issue, categorized as B4, to discuss workshops. (addresses item 6)

Another area which has a fair amount of concrete comments is chartering.

Note(subjective opinion): quite a bit of what's concerning about chartering doesn't ultimately have all that much to do with whether it's the director or the Team that's doing it (especially once Formal Objections are no longer handled by the Team), but rather relates to the lack of formal structures/process and documentation thereof. Either way, it's worth looking into.

Proposal 2: Florian to open a new issue about charter development, categorized as B4, quoting relevant parts from this issue. (addresses item 1 and 2)

Other areas are mentioned, but without any clear point expressed about what is wrong with each individual area (not to say that there isn't anything wrong, just that it isn't discussed in this GitHub issue). Some are mentioned as "needing some discussion", some merely listed as examples of the fact that the Team does a lot. None of these should be rejected out of hand, but neither does it seem good to try and guess what Mark's (and other's) concerns are and to put words in their mouth.

Proposal 3: Close issue without addressing items 3, 4, and 5, because there is no content. Specific concerns and improvement suggestions should be filed as individual issues, likely within category B4, or as additional comments on existing issues.

Is the W3C Council the right approach?

These two issues, both opened by Mike Champion to state that he thinks the proposed W3C Council is a bad way to handled Formal Objections, seem hard to separate:

Part 1, the problems (and a suggestion of potential solution): https://github.com/w3c/w3process/issues/331
Part 2, potential solutions (and some more problems): https://github.com/w3c/w3process/issues/391

Summary

Opened by Mike Champion to state that he thinks the proposed W3C Council is a bad way to handled Formal Objections.

Arguments include:

  • 1. It's too large (may be unwieldy to work as an entire body, but biased if it subsets itself, and possibly biased even if it doesn't subset itself, due to politics)
  • 2. Recusals are a messy affair
  • 3. an STV elected body is inappropriate for objection handling
  • 4. AB and TAG too busy already
  • 5. in the absence of consensus, supermajority may be appropriate, but not majority
  • 6. would need to follow some agreed-upon principles rather than making them up on the fly in order to be viable
  • 7. The history of the AB -- even before STV -- shows little indication that it can reach true consensus on much of anything.

Proposed solutions include:

  • a. Have a Technical director instead
  • b. Have a different and smaller body (dubbed Directorate) that isn't TAG+AB, but maybe appointed by TAG+AB+W3M+?
  • c. Maybe the Directorate doesn't handle issues directly, but first appoints a mediator (who maybe reports back, or maybe gets to decide)
  • d. Maybe the Directorate can only approve / reject (the objection or the mediator's recommendation), and status quo prevails in the absence of consensus.
  • e. Maybe there should be no tie-breaker body, and instead either:
  • e.1. go straight from FO to binding W3C Appeal, without an intermediate arbitrating body (possibly with a minimum number of FOs to trigger the appeal, possibly with the appeal being based on super majority)
  • e.2. any FO is blocking, only move forward when there is consensus

Commentary

The current proposition for the W3C Council is not finalized, and has a number of open issues against it on which we have yet to reach a conclusion. It's current design should not be considered set in stone. Quite a few of the points raised by Mike points can addressed through adjustments of the Council. Some are already open issues, some should be filed. The concerns raised in these various points are absolutely valid, and we should get to the bottom of them, but they do not necessarily mean that the Council is a bad idea, only that details matter. Many of these concerns were already raised before/in the AB February meeting. The AB concluded that these were not justification to abandon the idea of a Council, but rather that these were valid concerns to be individually investigated and addressed.

Here's a mapping of the above points that already have related issues:

1,4,b -> https://github.com/w3c/w3process/issues/293

2 -> https://github.com/w3c/w3process/issues/278

5 -> https://github.com/w3c/w3process/issues/280

6 -> There is a draft of such guidelines already, and some open issues about them: https://www.w3.org/2019/06/W3C%20Council%20guidelines%20for%20formal-objections.html https://github.com/w3c/AB-memberonly/issues/24 https://github.com/w3c/AB-memberonly/issues/25

7 -> related, though not necessarily identical, to https://github.com/w3c/w3process/issues/280

d -> https://github.com/w3c/w3process/issues/279 + https://github.com/w3c/w3process/issues/280

The following listed points don't have an existing standalone issue, but we could open one:

c -> The Council is already expected to not handle issues directly at first, with the Team doing the first pass. We could open an issue to see if we want to turn that on its head, and have the issues first go to the Council, who would then deliberately appoint an investigator/mediator, rather than let the Team self appoint.

Proposal 4: Florian to open a new issue with category B2 to discuss whether the Team should be doing the first pass or not.

3 -> related, though not necessarily identical, to https://github.com/w3c/w3process/issues/60 and https://github.com/w3c/w3process/issues/293.

Proposal 5: Florian to open a new issue to discuss the election mode of the AB and TAG and its compatibility with the Council being composed of these two bodies.

Even once we've sorted all these into sub issues to be processed individually, there remains 3 possibilities in Mike's suggestion that need to be considered wholesale, as they cannot be adjustments to the Council but are complete alternatives.

  • a: Have a technical director instead. We've concluded that "finding a new Tim" was not plausible, so any Director would need to fit somewhere lower in the Team's hierarchy. Many of us also don't seem to want to put to much power into the hands of the Team. Having our most controversial decisions be handled by someone who isn't in a "buck stops here" role, and not directly accountable to the membership seems questionable. The minutes for the February meeting, despite having no formal resolution, indicate fairly strong consensus against a director or director-lite approach.
  • e.1: go straight from FO to binding W3C Appeal, without an intermediate arbitrating body. Seems like a departure from a culture of consensus to one of voting, where one wins based on the number of supporters, rather than based on arguments.
  • e.2: any FO is blocking, only move forward when there is consensus. At risk of "consensus by attrition", "consensus by exhaustion", "consensus by bullying"… Also, Formal Objections may be raised against decisions which aren't consensus based. Removing the path to appeal makes various parties (especially the Team) less accountable.

Ideas a, e1, and e2 may be valid approaches, but they are radically different ones, which would likely require rethinking everything in the Process. The AB cannot get "stuck" evaluating all options simultaneously or we will never make sufficient progress. So the proposal is to close them so we can focus on the path we have chosen. If some one or some group of people want to pitch a reasonably fleshed out alternative, they are welcome to do so, but it would need to be decently detailed, not just an open ended suggestion, and in the meanwhile the AB needs actually focus on resolving the known issues. Said differently, if we were to leave a, e1, or e2 open, we should not proceed with the rest of https://www.w3.org/wiki/AB/Director-free_Plan because we will be thrashing too many options at once. Instead, we would spend a few months debating and concluding a, e1, or e2 - and delay our broader progress on https://www.w3.org/wiki/AB/Director-free_Plan.

Proposal 6: We are not collectively exploring a, e1, or e2. We are exploring other issues from these #331 and #391. Close #331 and #391.

https://github.com/w3c/w3process/issues/360

Summary

This issue, raised by "danield06" combines multiple points:

Commentary

The first point is largely the same as https://github.com/w3c/w3process/issues/316 (discussed above)

The second point is largely the same as https://github.com/w3c/w3process/issues/331 + https://github.com/w3c/w3process/issues/391 (discussed above)

Proposal 7: close subpoints 1. and 2. as a duplicates. Specific concerns and improvement suggestions about the Council or TAG Appointment Committee should be filed as individual issues within category B2 or B3 (or additional comments on existing issues).

Deepening our knowledge of how other standardization bodies work, beyond what various AB members already know personally, seems useful, but it is not a concrete issue against our DF branch.

Apart from the formal issue triage, it of course is wise for all of us to learn more about other standards organizations.

Proposal 8: Research and sharing of information is encouraged, but it is not a blocking step in the Director-free project.

OR, the AB could decide to leave open the exercise to have a totally different plan based on other standards organizations. But in that case, we should not proceed with the rest of https://www.w3.org/wiki/AB/Director-free_Plan because we will be thrashing too many options at once. Instead, we would spend a few months debating and concluding on what other organizations do - and delay our broader progress on https://www.w3.org/wiki/AB/Director-free_Plan.

https://github.com/w3c/w3process/issues/392

This issue, raised by Mike Champion, is similar to Mark Nottingham's https://github.com/w3c/w3process/issues/316, and express concerns about too many things being assigned to the Team and its CEO, and in particular their role in chartering. It also touches about concerns on the Council, as did Mike's earlier https://github.com/w3c/w3process/issues/331 and https://github.com/w3c/w3process/issues/391

Points not previously raised:

  • A W3C Council and a TAG appointment Committee are too weak and fragmented to counterbalance the team

Proposal 9: Florian to open a new issue to discuss whether the powers currently attributed to the Council and the TAG appointment Committee should be fused into a single body, and if so, how that body should be appointed.

  • Determining whether there is sufficient implementation experience is too important to just assign to the Team

Category K is a large category that has not been developed, which relates to issue #392. Several specific proposals arise from this issue (9, 10, and 11 below) that should be merged into the broader exploration of K in https://www.w3.org/wiki/AB/Director-free_Plan.

Proposal 10: Determining whether there is sufficient implementation experience, for every spec, is too much work to assign to a volunteer body, so it double down on it have to be on the Team.

However, this does not imply that there should be no constraint or counter-powers. An existing one is the AC Review. We can add / formalize others:

Proposal 11: Decisions of the Team to advance / not advance are decisions, and therefore can be formally objected to. Florian to open an issue in category K to find a way to make this explicit in the Process.

Proposal 12: The principles by which the Team makes decisions to advance / no advance should be documented into an explicit policy. First draft probably to be written by Ralph/PLH to document their current understanding, then refined with the AB and TAG, and linked to from the Process. Florian to open issue in category K.

  • Signing MoUs is too important to assign to the CEO

Proposal 13: Whether MoUs should be assigned to the CEO, the Board, or elsewhere is worth talking about, and fits within category B4. Florian to open an issue for that.

  • It is unclear who will replace the director in his role as a vision/mission setter

Proposal 14: The answer to this question has not yet been provided, but the need to answer it has been identified as topic N in https://www.w3.org/wiki/AB/Director-free_Plan#Relevant_Material. Consider this topic as raised, and get back to it as per https://www.w3.org/wiki/AB/Director-free_Plan

  • The way forward must have broad and early buy-in from the web standards community, not be cooked up in private by the AB/W3M and presented as a fait accompli to the AC

The AB, in its role shepherding the process, plans to work in a manner that makes it not "cooked up in private". Specifically:

  • Drafts being public since the early days and continuing to be going forward
  • AB minutes being Member visible
  • Issue processing continuing in GitHub
  • Update to AC in the spring
  • After the AB has made sufficient progress, taking the entire work through the W3C Process CG and AC review - before bringing it to AC ballot.

But, we are trying to make progress in the smaller group because otherwise we worry that we will not make any progress.

https://github.com/w3c/w3process/issues/393

Summary

This issue raised by Tantek presents itself as a tracker of other Director-free related issues, with some light commentary as to what the goals and aspiration of the project are.

  • This is misinterpretation of the issue. This is not a tracker. This is a more general issue with a few specific examples. Solving those issues is necessary but vastly insufficient to solve this issue. Tantek Çelik (talk) 13:45, 22 September 2020 (UTC)

Commentary

A list of issues isn't in itself a topic we can discuss or attempt to reach consensus on. For our shared understanding of what's in the director-free project, we're using the "Director-free" github project, and all issues in it can be found at https://github.com/w3c/w3process/issues?q=is%3Aopen+is%3Aissue+project%3Aw3c%2Fw3process%2F7, with various labels categorizing things per topic (as per https://www.w3.org/wiki/AB/Director-free_Plan). For each person's shortlist of the issues they personally find most important to achieve, private notes are more appropriate than an issue in our common tracker.

Proposal 15: Close the issue.

  • I reject this proposal because it is based on a false assumption and minimizes the actual meaning of the issue. Tantek Çelik (talk) 13:45, 22 September 2020 (UTC)