W3C home > Mailing lists > Public > www-style@w3.org > November 2016

[CSSWG] Minutes Lisbon F2F 2016-09-20 Part IV: Incubation in the CSSWG

From: Dael Jackson <daelcss@gmail.com>
Date: Tue, 15 Nov 2016 21:27:50 -0500
Message-ID: <CADhPm3v4ctYMu5zEzZkWnncenLjB-D6NjL8rU=brNUCiGpxrFQ@mail.gmail.com>
To: www-style@w3.org
=========================================
  These are the official CSSWG minutes.
  Unless you're correcting the minutes,
 Please respond by starting a new thread
   with an appropriate subject line.
=========================================

Incubation in the CSSWG
-----------------------

  - There was willingness to do incubation with some concerns that
      it could end up siloed or without the right people doing the
      incubation.
  - There was a strong desire to develop rules around when a
      proposal should go to incubation and when it should be
      considered ready to come out of incubation.

===== FULL MINUTES BELOW ======

Agenda: https://wiki.csswg.org/planning/tpac-2016#agenda

Scribe: Greg Whitworth

Incubation in the CSSWG
=======================

  Rossen: Since we have a few people from Google in the room [we'll
          do this topic]
  Rossen: As you know we recently re-chartered
  <ChrisL> https://www.w3.org/Style/2016/css-2016.html
  Rossen: The current charter has we are allowing and encouraging
          incubation of new ideas in the WICG but not requiring it.
  <astearns> relevant text from the charter: "The CSS WG may
             incubate speculative new work in the WICG, and may
             adopt promising CSS work developed in WICG, provided
             that RF patent commitments are in place for such work."
  Rossen: There was a topic today that was incubated elsewhere with
          a relatively complete spec which was desired to be adopted
          under fonts L4. There was an objection by TabAtkins.
  Rossen: Then a further statement that any new work should go into
          WICG.
  TabAtkins: I don't think the spec is defined.

  ChrisL: I will say it's a pretty good spec, I've reviewed it, I've
          seen far worse specs that have gone to FPWD.
  dino: The topic is incubation not this particular spec.

  Rossen: I would like to hear a more formal policy behind this.
  slightlyoff: As many of you know, we work across many standards
               organizations. W3C is only one of them, I ended up
               with the role of ensuring the health of those groups
               was good
  slightlyoff: and it wasn't and most of it looked like design by
               committee.
  slightlyoff: The proposals need a lot of momentum and community
               backing.
  <ChrisL> I would like to see it minuted that slightlyoff said that
           this spec *did* have a lot of momentum behind it
  slightlyoff: This is a very large investment.
  slightlyoff: This is the wrong place for design, my teams come
               prepared for resolving issues, not for designs.
  slightlyoff: Starting with the CLA policy that the CG has in
               place, is perfect for these proposals.
  slightlyoff: We've done quite a few things for this already.
  slightlyoff: intersectionObserver, web payments, resizeObserver
  slightlyoff: This is to get the i's dotted and t's crossed.
  slightlyoff: But it's not just us, Microsoft has been there as
               well.
  slightlyoff: Allow for iteration where you can bring well formed
               proposals that can become standards.
  slightlyoff: I did recommend a blanket proposal for that inside of
               Blink, thus TabAtkins request.

  tantek: I like the WICG, I think this working group has tended to
          moving in that direction but haven't formally adopted this
          process.
  tantek: It would front load some of those questions that we ask at
          CR time, it should help us move more efficiently.
  tantek: It keeps from there being specs that live in limbo for a
          long period of time.
  tantek: WICG is one way of doing that, but there are many options.
  tantek: in general I'm in strong support of adopting an incubation
          first model.

  cwilso: I'm actually co-char of the WICG and partner in the
          incubation first, Alex mentioned Microsoft but it actually
          started at Microsoft from Adrian Batemen.
  cwilso: It enables us to embrace graceful failure.
  <fantasai> I don't think the CSSWG has a problem with allowing
             specs to fail. Our failures just end up as abandoned
             drafts that we then put big red obsoletion notices.
             Granted, said notices are not very graceful.
  <fantasai> But I don't think graceful styling of obsoleted specs
             was the concern here.
  cwilso: We are not saying the WICG is not the only group, it
          speeds up the need to create a new group, etc.
  cwilso: If you feel the problem needs to be solved and you've
          scoped it down to what/who needs to be involved and you
          know what the solution looks like
  cwilso: that's really rare, but if it is then you don't
          necessarily need it to incubate.

  Florian: During the short charter review I believe you asked for
           mandatory.
  Florian: I'm not opposed to incubation, but I'm against to a
           mandatory.
  Florian: If Google wants to prove that incubation leads to better
           results, then do so.
  Florian: but the way this was handled was very unpleasant.
  <fantasai> +1
  frremy: I had a very same point to Florian, I think it helps the
          community, it's a very good workflow
  frremy: They are an external partner that has a sound plan as they
          already have it designed for Windows/Mac that they'll
          point to CSS.
  frremy: In this specific case I don't think it's necessary.
  frremy: They came to this group to ask us to work with them, and
          then we asked them to go to WICG.
  frremy: It seems like it's sending the wrong message.
  frremy: Exploring is for the WICG and I think the CSSWG is for
          refinement.

  astearns: I'm all in favor in moving things to the WICG when we
            can, so they can be done in a small tight group.
  astearns: Instead of setting up a subgroup or a TF, I prefer WICG.
  astearns: For this particular one it feels like a process speed
            bump, because things are passed the incubation phase, or
            relatively passed the incubation phase.

  cwilso: The question I would suggest to the wg is: it's not to be
          a process speedbump but the new normal is that everything
          should start in incubation. The wg may want to define what
          should be in the WICG.
  <frremy> (seems fair to me)
  <astearns> I definitely agree we should have agreed-upon
             guidelines, so we don't rehash this every time new work
             appears

  fantasai: Alan gave an example of step sizing.
  Florian: But this wasn't the case, it fits in this group.
  cwilso: That is not the bar, that you may need a new working group
  hober: Who do you mean by we?
  cwilso: um, the WICG. Started with Microsoft, myself, Alex,
          Google, etc.
  cwilso: The bar for incubation is much higher than starting a new
          working group for new features, it was to get a decent
          proposal to save the groups time.
  <cwilso> hober: Thanks for asking for clarity on "we". that was
           important, hadn't even thought of it.

  dbaron: I think one of the problems incubation solves, the way
          large groups discuss is not effective.
  dbaron: We have discussions about technical stuff and make
          decisions around stuff we at times don't even understand.
  dbaron: I think it's best to move the problem solving to the
          smaller groups so they have the right people that can make
          the right decisions.
  * ChrisL doesn't think WICG is a "smaller group"; more a
           super-super-group
  dbaron: Now there is the flip side, that the group chosen to solve
          that problem are the wrong set of people
  dbaron: So we can't just assume that a spec that was done in
          incubation is ok, may need to be completely redone.
  <fantasai> +1
  <astearns> wide review needs to happen in incubation as much or
             more than it does in a WG
  dbaron: I would like to ensure that incubation feedback is done
          earlier enough and high level enough that we aren't stuck
          with something that was effectively done by one company.
  dino: Would round display of something that fits your narrative.
  dbaron: Yes, it did get completely done.
  dbaron: Have the high level overview but not argue over syntax of
          each prop.
  <dbaron> one other sentence: the set of people in this group isn't
           actually the right set of constituents that need to be
           talked to

  fantasai: One of the points I wanted to hammer on is it encourages
            people to go off in a corner and not get a lot of
            feedback.
  fantasai: One of the benefits of this room is that you get a broad
            set of ideas.
  fantasai: There's a tension in standards around parallelizing and
            fragmentation; one of the advantages of discussing
            things in the entire WG is getting the broad expertise
            of a) people with different perspectives and b) people
            who have experience with many different parts of the
            technology and know how to integrate it and keep it
            self-consistent.
  fantasai: I think pushing everything into incubation will have
            another set of problems
  fantasai: Note to some extent we've been doing small-group work
            already with individual threads and side meetings on
            an ad-hoc basis.

  <cwilso> I would note that discussing something in WICG doesn't
           mean you can't be in the WG, nor that you will not
           discuss said incubation inside the WG or with other WG
           members.
  <astearns> +1 to cwilso's point
  <tantek> cwilso, indeed, I think CSS incubation could occur in the
           CSS WG itself, as long as the CSS WG formally adopts
           incubation as part of our work flow
  <cwilso> tantek: I'm not sure I agree, but depends on setup that
           is beyond the next 30 seconds of discussion. :)
  <tantek> cwilso, sure, it's going to take some very deliberate
           work (especially on the part of the chairs) to institute,
           guide, and enforce a policy of incubation in the CSS WG.
           That's a non-trivial challenge. But doable.
  <tantek> cwilso, in addition, I think culturally there's a good
           chance that CSS WG will come up with a compatible at
           least in spirit/methodology way of incubating that is
           more efficient than WICG processes literally.

  SteveZ: I actually agree with one point that Google made, we ought
          to have criteria that decides what does/doesn't go into
          the CSSWG.
  SteveZ: We've come up with a bunch of criterion for other things,
          we should do that for this. We should spend some time on
          this.
  SteveZ: I think it has to do with a number of the things that
          cwilso mentioned, like what the tech looks like
  SteveZ: such as regions, provided where we had discussion at a
          plenary session and got feedback.
  SteveZ: I think there are things that should be in incubation, but
          before we make forced decisions let's look at our history
          and come up with that.

  Rossen: I hear a lot of sympathy and willingness to incubate.
  Rossen: There were strong preferences to where to incubate
  Rossen: What are the consequences of not working on it where you
          prefer.
  slightlyoff: I would like to see the WG change and move to the
               incubation, we would have a hard time to follow this
               group if it doesn't change.
  slightlyoff: It worked for service worker, I'm less worried about
               where it happens more on how it happens.
  Florian: Out of principle will Google block us
  Florian: To the extent that Google wants to lead by example,
           that's okay
  Florian: But will Google block any progress that is willingly
           happening now out of principle?

  Rossen: Please, let's not do this.
  astearns: I'm willing to work on it.
  Rossen: That would be great, it would be good to cwilso and
          slightlyoff thoughts on the matter.
  astearns: We will have success criteria on when something gets
            passed off to the CSSWG.
  fantasai: I would like to see it come back to us as well, not just
            after it's all considered "done" and competed in
            isolation somewhere else. I don't want silos.
  Rossen: We will come up with that process.
  Rossen: When it comes to the CSS Font spec we'll discuss this
          later, I want to close on this issue.
  Rossen: Thank you slightlyoff and cwilso
Received on Wednesday, 16 November 2016 02:28:53 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 16 November 2016 02:28:53 UTC