Chartering at W3C

13 September 2023


Anssi_Kostiainen, AramZS, Chris_Needham, ChrisL, csarven, cwilso, dom__, fantasai, florian, Gonzalo Camarillo, hober, jyasskin, Kaz, MichaelWilson, MNot, Nigel_Megitt, plh, rgrant, tantek, Tatusuya.Igarashi
Elika Etemad, Florian Rivoal, Philippe Le Hegaret
fantasai, kaz

Meeting minutes

Pick a scribe

Reminders: code of conduct, health policies, recorded session policy

<hongchan> <hongchan> present+

Quick intro


<florian> Florian Rivoal, Invited Expert, Advisory Board, Process co-editor



Chris Wilson, Google Chrome and Advisory Board

Nigel Megitt, BBC, co-chair TTWG, chair ADCG



<MichaelWilson> Michael Wilson, Google Chrome

<cpn> Chris Needham, AC rep, Media WG co-chair, MEIG co-chair

<hongchan> Hongchan Choi, Google Chrome, Co-chair Audio WG


Sarven Capadisli (Independent), Solid CG chair


Elika Etemad aka fantasai, CSSWG, Advisory Board, Process CG co-chair, Process co-editor


Goal of this session

<rgrant> Ryan Grant, AC rep for Digital Contract Design

<seanturner> Sean Turner (sn3rd)

<seanturner> <--- in room

<anssik> Anssi Kostiainen, DAS WG, WebML WG, Second Screen WG chair

florian: how chartering works and doesn't work
… process says Team pays attention
… must tell to Membership
… propose the Charter to the AC
… and AC review
… if no objections, the Charter is ready
… Team is responsible to talk with various parties
… possible proposals on tweaks
… the basics is very simple
… but the detail is not really described well

plh: the guide we announce, etc.
… we need to have horizontal reviews before AC review
… but not part of the process
… beyond the guide, if we have an existing group, what to do first is telling the Team Contact to work with the group to generate a new Charter
… we have a repository called strategy
… to manage the progress
… but would say it's not easy to follow
… sometimes it's hard to handle

<nigel> Guide page on chartering

plh: e.g., a CG comes to us
… but how to verify it is difficult
… e.g., PAT CG
… Solid CG
… came to us with a proposed Charter
… verifiable identity as well
… we're finishing PAT proposal and sending the proposed Charter to AC
… we have many existing CGs
… sometimes say this group should become a WG
… how to do it is a question

florian: down side is it's very flexible
… the Team could proposal matching to or different from what was wanted, anything is possible
… majority or minority
… would like to open the floor
… encourage to start with problems first

anssik: Thanks for intro
… wrt best practices for my WG for improving engagement in chartering process
… I think some groups don't follow
… initial charter is in GH repo, and we actively advertise this
… try to make it as open as possible
… helps acceptance in AC
… idk if you have that as a best practice, but it's good

anssik: I have a question, you mentioned WASM WG
… charter expired a long time agao
… is there something we can learn from this?
… frustrated Members that WG charters expired

plh: It was my failure that WASM wasn't rechartered properly

<Zakim> plh, you wanted to answer on wasm charter

plh: I lost the Team Contact, and became Team Contact for 7 monhts
… and was too busy
… have help now
… As long as we don't create a new legal entity, we should be fine
… and I understand that there are concerns with this group

jyasskin: 3 fairly minor problems
… 1. It's hard to find charters. Every group puts them in a different place.
… There's no one place to find them.
… 2. Privacy WG chartering has been spinning because people don't know what the right thing to do is
… so waiting instead of picking something and moving forward
… 3. With PATWG, there was an FO around living standards text in original charter
… but some different wording around that charter in the template, which doesn't address FO
… I don't think anyone went back to update the template after the FO to figure out the right text
… Now that we've noticed text in template is wrong, should update it; should have happened already

ChrisL: I'm one of the maintainers of the template
… Recently, was waiting for wording to be resolved
… as I kept updating to match charter drafts, but that created a lot of churn

<nigel> Charter Template

ChrisL: so I've been asked to slow down, make sure there's consensus before making changes to template

jyasskin: Tantek raised objection, idk if he messaged that template should change

cpn: Folowing up on template issue, I've recently raised objections at AC Review stage
… not substantial in terms of the work, but changes made in template
… I think having more of a defined process around managing template sounds like a good thing
… One of the objections was related to something that originated in PATCG
… interesting listening to Sam talking about moving towards more community-driven chartering process
… I'd like to move as much of discussion around charters up front as possible
… so objections can be resolved before final stage
… I don't want to use AC Review FOs to resolve problems
… wrt PATCG, there's still an unresolved objection, but the chairs have chosen not to reopen discussion

florian: One thing I've found uncomfortable in the past
… not necessarily when rechartering, but when initially charter
… unclear who is chairing the chartering process vs proponent of an issue
… e.g. I filed an issue, and they closed it and say "Nah, that's not what we're doing"
… Unclear how this works
… and by the time gets to AC Review, the AC does not get to know that this is how comments were closed

rgrant: WG and non-WG AC Reps that don't find consensus are not engaged in a process of consensus
… when there are FOs, the matter is immediately sent to a Council
… so I'm looking for a process of consensus somewhere

<Zakim> florian, you wanted to react to cpn

florian: to respond briefly, the normal practice is that before going to Council, the Team discusses with participants and tries to find consensus
… but left to Team to determine if there's a path to consensus; if they don't believe, they can go directly to Council
… but first step of handling FOs *is* to seek consensus

<Zakim> rgrant, you wanted to say WG and non-WG AC-reps who disagree are not engaged in a process of consensus

<jyasskin> Florian was describing https://www.w3.org/2023/Process-20230612/#team-fo-mediation about the team mediating before going to the Council.

hober: Main consumers of charters are lawyers of large companies who are trying to figure out whether OK to join group or not
… Risk assessment
… What I hear from lawyers, the tighter and narrower the charter is scoped, the easier and faster it is to do that risk assessment
… Sometimes those assesments take time
… As an engineer, I get frustrated, because I want to join the group yesterday.
… What I heard in the PATCG was, ppl are concerned that if they scope the PATWG charter too narrowly, they'll have to recharter
… too annoying, don't want to do
… I'm sympathetic to the concern, you want more wiggle room
… but the more wiggle room you have, the harder it is for engineers at large companies to join the group in a timely manner
… I'd like rechartering to be something we can do quickly, and cost is low
… so that groups are more comfortable with a tighter charter

<Zakim> cwilso, you wanted to react to hober

cwilso: It actually is. We have to be cautious as engineers at large companies, how much scope we're signing up to

<dom__> [I wonder if this points to maybe different approaches to adopt for chartering/rechartering?]

<Zakim> AramZS, you wanted to say that the charters should be handled in comparison to the state of the template at the time the group is chartered. It's fine to evolve quickly on the template I think, but it shouldn't happen at the expense of chartering process for groups. Also - would like more best practice charters.

cwilso: harder for us to go convince lawyers

AramZS: In terms of template discussion, don't think it's bad that template evolves, even if rapidly
… only if it gets a problem when chartering
… Template is template to build upon, not the only way
… would be good as ppl respond to feedback, building up
… more difficulty of chartering if people are trying to hit a moving target

AramZS: Would also be better to get more best-practice charters
… What goes well with lawyersscope creep of charters is a concept we don't quite understand
… Some level of feedback ...

<Zakim> anssik, you wanted to ask about joint deliverables

anssik: Like to folow up on what Tess and Chris said
… I think we have a possible solution, discussing over last year
… example: 2 WGs want to define 4-5 specs as joint deliverable
… I'd like to say that scope of some groups be large, in part for ?? reasons
… I was following up on Tess and Chris wrt broadly-scoped WGs
… we have a concrete case of 2 WGs that want to take up substantial joint work around 4-5 specs
… scope of these WGs, we want to split a bit, that's due to historical reasons
… previously supergroup that adopted new work as came through incubation
… we believe joint deliverable will be a good way to solve scope creep issue
… but one blocker on our way is that joint deliverable as a concept isn't defined anywhere in W3C Process
… which was surprise to me
… that we don't have agreement or rules on how we would do this
… I feel a bit bad, not able to deliver this agreement for joint work for my summer vacation
… found out it took 10 years to not come up with solution
… would like to solve this problem, that joint deliverables might be a solution

sarven: Wrt how chairs are chosen

[audio cut]

nigel: My first point is, although it's the Team formally who need to prepare the charter

<csarven> w3c/tpac2023-breakouts#43 (comment)

<csarven> Self-Review Questionnaire for Chair Candidates: https://github.com/solid/specification/discussions/568

nigel: as you described, Team worked with group
… I think it's confusing, needs to be a person who takes lead and owns it
… gathering views of people etc.
… Over many years of rechartering, there was an expectation that chairs would prepare the charter
… I think that situation has become less clear over time, needs to get more clear
… Charter is a deliverable document, needs an owner, and we need to know who that is

nigel: Wrt consensus, when chartering for first time
… not a defined group that you're trying to agree among
… If you have 10 people and 1 doesn't agree, can recast it as 9 people who agree
… that's an anti-pattern

nigel: There's a point jyasskin made about findability
… has been suggestion is to put all in repo
… really bad idea, difficult to work on comments per charter
… one document per repo is a better idea
… orthogonal to findability problem

nigel: Charter template adds a lot of pressure to conform
… that's a new thing, doesn't feel right
… too much pressure to conform
… idk how to address that balance

<jyasskin> +1 to one document per repo

nigel: There's a part of charter, if group doesn't intend to move to REC, add this text
… I think if a group doesn't intend to go to REC, needs a lot more highlighting than that!
… if you're working on a standard, why not making a standard?

nigel: Wrt deliverables and timeline, there's duplication
… either have a list of deliverables with times
… or a timeline
… having both creates potential for error state
… happens like 100% of the time

<plh> One repo for ALL

nigel: very unhelpful

nigel: Lastly, I don't understand how the joint deliverable idea
… how it works, what does it mean?
… Making things confusing for people
… should be a better way, much clearer to have specific ownership of each REC-track document
… then you know what's happening
… if you need more people, have them join the group
… or have formal requirement to get review

<AramZS> +1 to one document per repo for sure

nigel: that feels like safer places from management place
… I'm sure you're solving a problem there, not sure what it is, but going to create a lot of other problems

<Zakim> anssik, you wanted to react to nigel

anssik: I agree that we need joint deliverables defined before we use them
… can't create the rules as you're flying

plh: We've been using joint deliverables for the past 20 years

anssik: We need a written agremeent. Doesn't need to be complex

<Zakim> dom__, you wanted to suggest evaluating rechartering shortcuts

dom__: 2 points
… 1 pattern I'm hearing is that rechartering has created a push for looser charters, which create problems
… so maybe one space of exploration is creating a simpler rechartering for adding deliverables
… Stepping back I hope we get out of this breakout, how do we go about these topics
… Not just Process document, not just Team operation thing, also involves WG chairs trying to understand their roles, same for CGs
… If we can get a sense of right mechanism, that would be helpful

ChrisL: wrt legal thing, one change to template recently to change "scope" and "motivation" sections
… since actual scope is what legal need to review

csarven: Wrt how chairs are chosen
… want the Team to be more transparent about that process
… I understand it's a private decision, not sure whether AB is involved in decision or aware of discussion?
… if there is a charter being proposed, coming from a CG, I think the CG and community can benefit from
… knowign decisions made beyond proposing the charter

<AramZS> I do wonder if there is a problem with too much history etc being put in charters I'd suggest formalizing some sort of history / background artifact that is optional to groups instead of trying to *formalize* putting that information in the charter itself just in a different section.

<nigel> +1 to Dom's suggestion of easier process to add deliverables to charter

csarven: Document on role of the chair, and the document is along the lines of self-review for chairs nominating themselves and for Team to consider
… Slight gap between CG proposing a charter, Team updating charter with chairs
… no communication with CG as to why those particular chairs were chosen

<AramZS> Yes +1 to making it easier to add deliverables to the charter.

csarven: I have a list of question, maybe publish somewhere
… so that chair selection is more clear

<csarven> w3c/tpac2023-breakouts#43 (comment)

<csarven> Self-Review Questionnaire for Chair Candidates: https://github.com/solid/specification/discussions/568

<rgrant> it seems that one of the reasons for team's involvement in chartering is tht there is not a clear quorum for consensus. rechartering, especially for maintenance, does have a clear quorum for consensus, and the consensus should be within the WG. when the team takes a charter out of the WG without consensus they have the opportunity to present the worst worded charter to AC-review rather than a charter with the weakest technical objec

<rgrant> tions.

<Zakim> rgrant, you wanted to say rechartering, especially for maintenance, does have a clear forum for consensus

<tantek> +1 csarven great questions. and https://github.com/solid/specification/discussions/568 looks really useful too

rgrant: Seems reason for Team's involvement in chartering is that there's not a clear quorum for consensus
… when chartering
… but for rechartering, I think should be in WG
… when Team takes a charter out of WG, they might take a bad charter
… for that reason, the consensus required for rechartering should come from within the WG

<Zakim> fantasai, you wanted to speak about joint deliverables

<csarven> tantek: I'd be honoured! Please copy/revise/reuse as you see fit. Would love to see something along these lines out there in W3C docs.

<florian> fantasai: external review is not a good substitute for joint deliverables

<tantek> thank you so much csarven. really appreciate it 🙏🏻

<florian> fantasai: because it doesn't give IPR commitments

<florian> fantasai: only the publish group is bound by the patent policy

<florian> fantasai: [missed], it's not that complicated

plh: Wrt charter template, reason we do it
… I see some AC reps trying to push through requirements outside the process
… a good example is the Ethical Web Principles
… this is not adopted by the Membership, only by TAG
… it's in the template, but I'm very uncomfortable with it
… but also I understand why it's there
… I think solution is to push TAG to adopt as Statement
… but putting in template is problematic I believe
… I have plenty other examples

plh: I wonder if we should announce, when we make changes, to the AC
… currently no opportunity to push back on template changes

plh: AB has made it clear that the template is not requirements, they're guidelines
… Welcome to not follow, if good reason. Should document the reaosns

plh: Another thing, to help solve problem of lawyers, moving more of the text into separate documents
… sections that don't get modified
… e.g. testing policy, linked from every charter, but it's a separate document

<AramZS> jyasskin: no

plh: says it can't be changed without AC review, to reassure lawyers

plh: Yes, I encourage staff to work with community as much as possible
… to work with CG and WG and try to find consensus
… but if no conensus found, not a blocker for the staff
… Can't allow a member of community to block
… next step is AC Review
… For DID charter, we were fully expecting to get FO

<rgrant> that's a process objection

plh: Btw, formally not able to FO prior to FO atm
… need to try to solve that
… My current motivation is, whether I receive before or during AC Review, should send them all to Council

plh: Would like to simplify rechartering, but when we present e.g. adding a deliverable, the AC will sometimes comment on other parts, even FO on them

<anssik> [ Off the queue Q: Should informing the AC of template changes apply to the CG charter template too? It is used by many productive CGs. ]

plh: e.g. DID WG. If we propose a pure maintenance charter to AC, would get FO today

jyasskin: Don't have time to talk about solutions, is someone write them up as issues and CC folks?

<Zakim> nigel, you wanted to respond to the need to document why not following the charter template

nigel: Responding to thing about not using charter wording, have to explain why
… that's example of pressure to conform
… as long as you're working in the Process, shouldn't need to justify yourself
… Classic example of going too far with template

nigel: Another point, AC Review is too late
… FOs in AC Review go to Council quite often, add time delay and stress
… we need a group that owns development of charters, can understand its own consensus before AC Review
… which should be a final stage. Should be exception to get FO
… smoother for everyone

<rgrant> +1 to AC-review is an exception after the proper group seeks consensus

hober: I feel tension between template vs objections
… charter template is to capture best practices for not getting objections

<Zakim> AramZS, you wanted to say 'Good reason to alter from the template' is in the eye of the beholder, and is not a very great process step. If deviations from the template need to be justified we need clearer guidance on: how, to what extent, what justification information is expected?

<tantek> +1 hober

AramZS: Wrt charter template, I understnad what's being said here
… wrt avoiding objections
… if that's the route we want to go, if deviating
… what the justifications, how to work with group that deviates
… needs to be better document
… what's a good documentation? What makes it a good justification?
… if using template as part of process, need to be clear

florian: This is clearly not a converstaion that's over
… priority project of AB for the entire year
… working from these minutes, we'll try to summarize the various points raised
… debrief to AB
… eventually won't be surprised if changes to Process

<dom__> [Ian and I had explored not just maintaining the template but also the reasons that led to its content]

florian: but atm don't jump into solution space

<tantek> +1 florian, digest what has been discussed today instead of jumping into solution space immediately

florian: next step will be discussion in AB, and eventually this will get into Process and Guide
… but middle zone in between, and that will be part of what AB talks about
… do we put all of this discussion into Process CG? Discuss first in AB?
… not clear, but input from today will give us a sense of scope for what we do

<AramZS> dom__: Sorry! I think the summary from my question on the queue is a decent summary of what I was trying to say, I tried not to go too far off topic.

florian: Thanks everyone for your input, conclusions coming but not in 5min!

Minutes manually created (not a transcript), formatted by scribe.perl version 221 (Fri Jul 21 14:01:30 2023 UTC).


Succeeded: s/Nigel Megitt/Nigel Megitt, BBC, co-chair TTWG, chair ADCG

Succeeded: s/Chris Wilson/Chris Wilson, Google Chrome and Advisory Board

Succeeded: s/Fantasai/Elika Etemad aka fantasai, CSSWG, Advisory Board, Process CG co-chair, Process co-editor

Succeeded: s/, Team//

Succeeded: s/Sarven/Sarven Capadisli (Independent), Solid CG chair

Succeeded: s/propose a group too/proposal matching to or different from what was wanted, anything is possible/

Succeeded: s/[missed]/the AC does not get to know that this is how comments were closed

Succeeded: s/step/step of handling FOs/

Succeeded: s/up on ?/up on what Tess and Chris said

Succeeded: s/?/scope creep of charters

Succeeded: s/people are trying to [missed]/people are trying to hit a moving target

Succeeded: s/?/Chris/

Succeeded: s/come up/not come up

Succeeded: s/... There's/nigel: There's/

Succeeded: s/join de/joint de/

Succeeded: s/haveplenty/have plenty/

Maybe present: anssik, cpn, nigel, remote, sarven

All speakers: anssik, AramZS, ChrisL, cpn, csarven, cwilso, dom__, florian, hober, jyasskin, nigel, plh, remote, rgrant, sarven

Active on IRC: anssik, AramZS, cpn, csarven, cwilso, dom__, fantasai, florian, hober, hongchan, Ian, igarashi, jyasskin, kaz, MichaelWilson, nigel, plh, rgrant, seanturner, tantek, Wolfgang