Meeting minutes
Brent: we've invited you here to ask for your feedback on ways to shape the process
… we are officially running this under Chatham House rules
… so no attribution. please raise your hand to q.
if you have been involved in W3C, have some emotional attachment, and someone says something not nice about process
… you may be tempted to be defensive, or say "that's the intent" or "that's not in the process"
… my suggestion is we not be in that mode but rather listen
… and internalize feedback in a structured way
… example: support public transit. answer: thank you for the feedback and I will take it in the spirit it was meant
… there are a lot of opinions. it's ok if brainstorming produces flatly contradictory ideas
… we want all the ideas. it's ok if we don't come up with/ a coherent idea in an hour
… we're not solutioning
… all ideas are welcome and useful
… Process doc is quite long. if you want to see it printed, it's impressively long
… so we're not looking at a small slice. might be chartering. or maybe you've done member submission
… if there's an unusual concern you have expereicne w/ let us know
… we have 3 questions: how would you improve the process, what are proririte, are there processes you use elsewhere you'd recommend
… if you have something ot say, say it. Quaker style
… I think W3C is more mature that most orgs. not many have the level of detail and thought
… that's not bad. good for us. but for source of inspriration others might be bare minimum
… i don't mind it's long, as long as it says what it needs. I don't think it's in a bad state
… one gripe, I put this in GH too, any change that encourage groups to feel it's ok to target a rec track spec
… as permanent CR is an anti-pattern
… what they're trying to do is get a spec w/ w3c logo that's not in end state
… I think that's a waste of time. they're using a lot of effort.
… without having the right goal in mind. there's nothing wrong with/ any rec. should be easy to update
… make it easy to update
… i am fine w/ the length. i'm rather new. one thing I find weird is there's no consideration on versioning
… every rec goes through same process even with a small change
… eg: 1.1. but efforts the same. we gave up to do small updates. then we move to only big effort. it's the same effort, all implememtions and features tested again
… i find that a bit weird
… just to join the 2 comments, one thing yes, and
… yes and i think think there's view where it should work for everyone
… that doesn't necessarily force us into these diff status for specs
… but rather insist on underlying principles. PP, wide review, implementation
… that's the 3 pillars. and consensus building that exists at w3c
… w/out necessarily tying to a work flow that's complicated now
… in that you have the ability to do things at diff version levels
… you can have a rec and update a rec. it's a bit clunky w/ the levels
… there are some where you can go back to a maturity level
… or create another version, 1.1. or 2.0 and then start from scratch. i
… it's strange to start from scratch as it's iterating from something
… right now there's none of these paths are easy. could be smoothed
… would be nice to have a story to enforce principles. w different ways, none of which are easy to implement
… one minor point, not sure process or practice. when you do that journey
… and you end up w/ something w/ short name. diff version
… one thing is that it's really confusing what's most recent. latest rec or the replacement rec. there are places where the links don't work
… let's not try to answer that, just gather feedback
… something I've observed is that the process provides uniformity w/ WG
… if you have, then you know how to do work in another
… a lot that's the same is stuff we've all done for while. sometimes things the same
… are basically the same text in charters, we use the same template
… i wonder as a big idea, not specific. there are differences
… they have their own working mode, quirks
… the line between what's uniform and what can vary is a knob to turn
… we have comparison eg: ECMA has a different knob,
… different place. we might copy-paste into docs. if there's stuff in charter template into process if they are uniform/
… the link to charter, process and practice are variables we can fiddle with
… we should consider fiddling w/
<Ian> ls
<Ian> Version management in W3C Tech reports
@ from ECMA did a session. ECMA standardizes stuff. W3C is trying to create a coherent platform
… there 'might be reasons that there's uniformity
I received feed back by someone not able to attend. ITEF provides uniformity
… that they find useful eg; page by group to find minute, documents etc
. .and we do not have that
It's not he process
it's a knob that could be turned
my CG has existed for 5 years, we're going through process of WG
… from a CG i had no idea how to get a sense of how groups are doing things
… it involved going to each group and find their resources to find what parts we wanted to steal
… the next step to WB has been easier as I understand the work. but it's similar thing eg: deliverables,
… or a set of provisions. my impression is that wrt deliverable, groups have moved toward living standard
… approach. and i'm not sure of the reasoning. but for our group, we felt a living standard made sense
… eg: so a long standing CR. that's how we accomplish objectives
(a reminder if you've been involved in process to not try to respond to comments, to correct. we are looking forward to not have pushback but to bring comments)
On the discussion vs. uniformity, w3c used to be unform. they were the same/
… there was no policy. there's a bit in the process. it could be in it, we could turn the knob more or less
we have a system of tools to do work
… those are manifestations of how we do the process. the process is an activity and here's how we do it
… there's been a diversification of tools
… not too far on a limb to say, 15 years ago w3c tooling was way better than any other SDO
… today not sure i can say taht. not a critique. the world has been moving, not sure we've kept up
my focus is on a11y and we have discussed how to get ppl to think about it earlier
… same for priv or sec. sometimes when it comes, it's too late to add a11y to it
… we've been considering how to get it earlier. this come in product development too
… how to get this in process, templates or like when a CG starts, that's when they need advice on a11y etc
… not when they go to CR. not sure tools for that, but it would be useful to get TAG principles, things covered in reviews early in the process
to reinforce, i was gathering feedback in another group and they said it would be great to get feedback earlier
that was something that came up in design review in TAG. ti was a bit demoralizing
… but it would be "looks like you checked check marks"
it should be not just fixing bugs but it can't influence design
some high level design ideas. i'm doing a strategic objective on how we think .
… one idea re: a comment made, for process for w3c what are core principles for how we design it
if pl have ideas, they can mention tthem
yes. but a minimal set of principles. I'd love to hear thoughts
… tension between breaking into things. eg: TAG and AB vs. what a chair needs
… how to modularize it. eg: how to and Guide. there's also usability etc
… someone said "3 page process doc' and
ECMA's equivalent is @
We could reduce the process in a variety of ways if that's the goal
… we can have in one place, or have links. it's not just UX but it's about how things evolve
I'd appreciate not to read AB or chairs. I just want registry. the level of changes for recs is not there,
… would have been good
an idea, it was mentioned charter template. i thought about registries as we'll
… when we introduced registries we left some walls and let ppl figure it out
… i made a template for registries. may be a concept between process and guide
… of standard practice. we could specificy register and then let ppl do it and see if there was convergence
… that we could turn into standard practice. fi not, then maybe dont' need it
… maybe there are other candidates
Fascinating to hear themes. everything touched on here, has been touched on a in a group
For a template, i learned about them from a breakout, otherwise i'd have no way to learn it
… it wasn't listed, you can't find it in the search engine
I want to take a moment to endorse this.
… the thing about the process doc that we came to a conclusion on is that one thing w3c is doing is
… that others are not is driving to a coherent set of goals
… ECMA has short goals. we spend a year w/ the scope of a TC.
… then it's up to chairs. anti completive rules, notes etc. it's not driving toward a single coherent platform or goal
… doesn't need level of detail
… when it's at ECMA it goes to GA, they say it makes sense and publish it
… as imperfect as it might be, W3C process doc, it does provide guidance
… ECMA's guide is 500 pages
To riff on something you said, we work at diff companies w/ diff processes
… about jointing or protesting about a group
… one bit of feedback i get is that ppl want to see crisp scope. they often don't care about anything else
… that' s hurdle, scope. so it's not like everything can be defined. maybe that's all a charter needs
I love this idea. what's interesting is scope. a yes and
… it's interesting you talk about scope. we thought CG would be easy
… .eg: Incubator CG we thought one big group so no one had to ask lawyers and it's just one big group
… strong relationship between docs. has diff IPR, more tolerable ethan WG
… some anti pattern. if we thought of a super group mentality
… w3c had 5 supergroups. rechartering could go away
… that super group has scope they can do whatever. no recharter
… would a super group w/ well defined scope, could go on
What would SG be?
How's that different than a domain?
I'm thinking of scope IPR friction. i don't have specific topics. that would be highly politic
WebApps thinks its doing that
So does CSS
<DKA> s/webapps/webappsec
You could combine cSS and SVG
Those are interesting ideas
I have direct exp w/ a group in the European Broadcasting Union
… smart media, distrirubiton and spectrum
… one thing it does is creates a place where you need someone in charge
… what that person does is entirely vague. they don't do anytihing.
… it's a gathering place. a reporting mechanism. someone who wants alignment. but they can't bc they are dependant o chairs underneath
… if you do that, you need a really clear reason why. and really think out what's the reason to create a SG
… and what do you want to get out. how do you expect ppl to get out of it
For stable areas you might use it. but for new areas, might offer oppruntiy
This is interesting. i spit standards time here and ITEF. they've done this
… structure is we have diff areas. real time, security. i only pay attention to a few
… one person over area. and assist chair and group.
… another thought is there's a need for a group have scope. to say what working on
… w3c has a tightly constrained timeline. good and bad. hopefully helps ppl get work done
… but does it? it needs a siituatoin where HR ppl can't do it, and then i'm slammed, but can't get to it
… if no end point but scope of work. that takes care of it. if you get review, have to change things in HR
… takes away fear of review
Maybe why CG take time and give to WG. bc CG doesn't have deadline
Do most groups feel that pressure. CSS WG we have to recharter
… everyone assumes of course ewe'll recharter. do others feel it?
I'v never thought charter end date was other than date. deliverables are aspirational
The sound of them whooshing by
The groups i've been in felt they did matter
Context dependent
Talks w/ ppl that the charter end date is a forcing function and as members are heavily deadline dependent or they might put it off
Had a session today, and a thought about encouraging engagement
… from hearing from ppl we don't hear from much. it occurred to me that one thing i've done a lot of
… around here that isn't high on my list is FOs
… i've sat on a lot of council. I've thought a FO means something bad happened.
… a way to measure badness. the more here, the worse things are going
… in some way it's a measure of engagement and care. one way to measure i we're being provocative enough
… no objection, are e doing anyting?
Ways to define consensus. some require unanmity
… others are positive endorsement required. if you ask, at least one person has to say yes
… a more positive spin
What we have currently is a mix. we do need active support and everyone needs to live w/ it
A majority in saying yes, no one saying no
or substantial amount saying yes
I want to insist on rechartering could be something we can streamline
… the scope does not change. something time consuming for everyone
… group chairs dont' want to do it. they have other priorities would rather progress the group
… team contacts dont want to do it, it's tedious
… take sa lot of time for HR, and Comm team to send to AC
… i think it bothers the AC bc they don't vote. sometime for rechartering the companies in the groups don't even vote
… probably signals something very formal for something expected, for something that they'll get to the end of might not be good
Useful to have principles ot see if we can optimize principles.
… eg: SG reduces charter need. it would stabilize. eg: CSS if it were a SG might be more patent exclusions.
… if not really, then good. that's one area. another one migh the specs themselves
… .eg: web assembly, a11y review it and say "its so low level we don't need to review"
… not limiting ability, optimizing and WebAssembly keep going and HR ppl put more time into things that deserve it
… what's the reality. then if WebAssembly says they might go to REc
… they might and then we don't have to fake it. the process works
Comment was made that sticking in CR is an anti pattern
… trying to leverage that
… there are interesting ways to be principle based and more flexible e
Fascinated by idea of optimizing HR. we have change classes, substantive or informative
… we could have change class which are a11y or sec impacting.
… or some assessment to change spec, someone like Team contact, if adding user interface. before complete, let's just check in
… on a small scale, the HR group has a minimum thing to review
… at the end they'll have seen things. do it as we go along
I've seen that work in product dev. checkpoints for adding new features
… where aspects like UI would trigger a11y or sec review. and would flag ppl who can do it
… doing it smaller makes it better. reduces load of HR> and make more effective
… helps train ppl working on spec on things they need to work on
… and ppl who have been doing it, start to learn, might need a threat model
… not just big scary ppl will tell me where i'm wrong. and finding ways to trigger on additions vs. big gates in a process
We may not need process updates to drive this. to find workflow s
Where does HR fit?
It has to happen but does not outline how
Lot of ppl assume HR is a laucnh gate and not something early
Observation: we've talked about optimizing or skipping HR. but when i talk to others at other SDOs
… the thing we do w/ HR is the single distinguishing characteristic
… some maybe dont' bring work here but some like it
Not talking about removing
Careful how tinker
I hear if cookie cutter process, find optimization. we want to keep principles
Review is too light. it's impact, it's all that not just saying good enough
Process embraces this, whether we do in practice or not. we can already and many charters encourage that behavior
… its interesting to figure out why
Seems ppl don't know how to start or who to ask. seems a notification goes out at some gate
… vs. pinging ARiA or a template which makes it easier and more visible
… lots of ppl don't know how to do this stuff
I remember on the TAG ppl would want to hook into design review tracker
… bc it was a way to find out when ppl were looking for review
… each group had their own way to do it . felt like a hack. a more uniform entry point migh mean
Potential standard practice
More kick off than review
An invitation to dialogue
… last minute curve-ball, process doc defines a lot of things.
… we've been talking about our work to get a doc along maturation process
… there's ton of other stuff there. kitchen sink doc of governance. defines AC
… we asked how to improve AC participation. maybe we can redefine AC. look at table of contents
… all sorts of stuff in there. don't hesitate to tell us. we're open to zany ideas
In the Process CG
It's a CG for a reason
We need usage data to figure out how used and order it that way :)
Thank you all so much
We are one process reform away from not having an AC
There are tons of things in there we take as fixed and immutable
One group asked why we have CR, why REC. i said none are natural laws
… they happen that way bc we make them that way