W3C home > Mailing lists > Public > www-style@w3.org > April 2018

[CSSWG] Minutes Berlin F2F Tue 2018-04-10 Part I: Snapshot, Scheduling, Decisions

From: fantasai <fantasai.lists@inkedblade.net>
Date: Thu, 26 Apr 2018 23:10:47 -0700
To: "www-style@w3.org" <www-style@w3.org>
Message-ID: <097bcce8-478f-217c-c65e-df39837d145c@inkedblade.net>
=========================================

   These are the official CSSWG minutes.

   Unless you're correcting the minutes,

  Please respond by starting a new thread

    with an appropriate subject line.

=========================================





2018 Snapshot

-------------



   - RESOLVED: Merge text describing cleared-to-ship features after

               editorial fixup.

   - RESOLVED: Proceed with publishing snapshot as-is including the

               open issues.



2019 F2F Planning

-----------------



   - RESOLVED: We will have a late Jan or in Feb 2019 meeting,

               possibly in Sydney.

   - The group generally favored a meeting every 3 months which will mean

     that depending on TPAC there could be 3 or 4 meetings in a year.



Accumulating Compat Issues via Ambiguity

----------------------------------------



   - Everyone agreed that they'd like this to not happen again.

   - Many felt that the percentage margins issue was not typical,

     because this case was intentionally made ambiguous due to

     a split in both author and implementor opinion, resulting in

     a complete lack of consensus on the right way forward.

   - Looking back, this may have been avoided if someone had taken

     the time to specify a property to replace the percentage

     margin hack.




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



Agenda: https://wiki.csswg.org/planning/berlin-2018#schedule



Present:

   Rachel Andrew, Invited Expert

   Rossen Atanassov, Microsoft

   Tab Atkins, Google

   L. David Baron, Mozilla

   Christian Biesinger, Google, observer

   Brian Birtles, Mozilla

   Oriol Brufau, Observer

   Tantek Çelik, Mozilla

   Monica Dinculescu, Google

   Elika Etemad, Invited Expert

   Rob Flack, Google

   Simon Fraser, Apple

   Jihye Hong, LGE

   Dael Jackson, Invited Expert

   Ian Kilpatrick, Google

   Rune Lillesveen, Google

   Chris Lilley, W3C

   Peter Linss, Invited Expert

   Myles C. Maxfield, Apple

   Naina Raisinghani, Google

   Manuel Rego, Igalia

   Florian Rivoal, Invited Expert

   Richard Rutter, Clearleft

   Geoffrey Sneddon, Invited Expert

   Alan Stearns, Adobe

   Shane Stephens, Google

   Surma, Google

   Majid Valipour, Google

   Lea Verou, Invited Expert

   Eric Willigers, Google



Scribe: dael

<RRSAgent> logging to https://www.w3.org/2018/04/10-css-irc



2018 Snapshot

=============



List of features for shipping

-----------------------------

   github: https://github.com/w3c/csswg-drafts/issues/2388



   florian: The snapshot has a section with the general policy of the WG.

            We don't do vendor prefixes but do draft maturity. The WG can

            make exceptions based on market pressure. We don't list what

            the exceptions are. fantasai created proposed text and a list

            of examples.

   dbaron: One of my concerns about the list is over the past 15 years

           we've cleared various features at various times that are

           scattered across minutes. The more complete the list is the

           more serious it is to have them listed.

   fantasai: We should get to the point where the list is comprehensive.

             If there are things not on the list we should add them.

   florian: It's a note, so republishing isn't hard.

   <astearns> section under discussion:

             https://drafts.csswg.org/css-2018/#experimental

   dbaron: We may find two or three missing over the course of a month.

   fantasai: That's okay. We'll add them.



   florian: Review the list?

   fantasai: I think my list was things we've already cleared. There's

             stuff we didn't clear but already shipped.

   florian: [reads list]

   florian: Chris suggested we add conic gradient

   florian: :focus-within?

   fantasai: I don't think explicitly cleared, but it's shipping.

   florian: That's the list. fantasai has an edit proposed. If everyone

            is happy we'll merge.



   dbaron: Is this the edit with duplicate text?

   fantasai: Yeah, we have to clean that up.

   dbaron: I'm fine with merging after the duplication is sorted out.

   florian: Are we okay resolving to merge this after editorial improvements?

   astearns: And going forward as we approve things edit this in.



   fantasai: And if there's something that needs to be added to the list

             let us know.

   dbaron: There's things that should be on the list but I couldn't find

            minutes.

   dbaron: Some were over 10 years ago.

   fantasai: Just add a comment and say "I think we cleared but I couldn't

             find minutes".

   fantasai: We can put them in and we'll either find the minutes or re-resolve

   Rossen: Other suggestions?

   Rossen: Objections?

   RESOLVED: Merge this text in after editorial changes



   florian: There are 4 open issues. Link in the agenda. Do we want to publish

            before dealing with them?

   florian: One is an issue in bikeshed, the indexes are slightly wrong, I

            wouldn't block on that.

   florian: Document conformance should be added from 2.1, that should happen.

            But do we want to block on it?

   florian: Last is sort of related, issue #1139. Nothing lists what the fields

            in the propdef table means. Snapshot could be that.

   ??: I think that's the right place.



   florian: It's a new snapshot because it's a new year.

   Rossen: Sure. It will be more beneficial to get the snapshot out.

   Rossen: Was issue should we block on those?

   florian: yes

   Rossen: Obj to proceed with publishing snapshot as-is including the open

           issues?

   RESOLVED: proceed with publishing snapshot as-is including the open issues



2019 F2F schedule

=================



   Rossen: Both myself and astearns have been asked many times why we have

           so much agenda, we can't make it through everything, should we

           go back to 4 meetings?

   TabAtkins: We have an open afternoon right now.

   astearns: It'll fill.

   astearns: Given the last month of telecon that were too long for the time.



   astearns: Unfortunately TPAC is in September so it makes it awkward to put

             3 in before that.

   dbaron: TPAC will be about a month earlier then this year.

   Rossen: If we needed to get 4 next year that would make it hard unless we

           do one in December?



   Rossen: The fact is that at least since Seattle 2017 we've been running

           every meeting with parallel tracks as much as we can so we can

           absorb more agenda and have more focused areas of discussion.

   Rossen: Even with these efforts we're still running tight on the schedule.

   Rossen: Is 4 meetings in general something we should think about?

   Rossen: If we do go back to 4 we have to start thinking about what the

           first meeting of 2019 would look like.

   fantasai: Given the way TPAC is scheduled it's likely best to think in

             terms of how many months between a meeting, rather than how

             many between TPACs.

   florian: Yes and no. 3 months after TPAC is December.

   fantasai: TPAC is end of October?

   dbaron: Yes.

   fantasai: So we might do 3 meetings in 2019 because TPAC, but then it

             might be 4 the next year.

   fantasai: The problem with this year is it's been 6 months since the last

             F2F. If we space out better it'll work better. I'm happy for 4

             a year but if it's a short year one every 3 months-ish will be

             a better target.

   Rossen: We know we have TPAC and TPAC-next. end of Oct 2018 and Spet 2019.

           That's a shorter 11 month period.

   fantasai: If we want to space every 3 months we should do Jan next year.



   dbaron: I was trying to work out even spacing for 2 between TPAC and it's

           early Feb and late May/early June.

   fantasai: And 4 would be Jan/May/July

   <dbaron> I think even spacing would be early-mid February, and then late

            May / early June

   <dbaron> Strict averaging would say meeting the week of February 11,

            and then the week of either May 27 or June 3



   astearns: Would anyone be able to host late Jan/early Feb?

   Rossen: That's the 2 meetings?

   astearns: That's even with the 4 because we can't move the first any

             further back due to holidays.

   TabAtkins: If we'll do Feburary we'd be happy to do another Sydney.

   fantasai: You've got no one based in Sydney.

   shane: I'll still help.

   Rossen: We have Sydney as a potential host in early Feb/late Jan. Any other

           ideas or offerings?

    * tantek SGTM



   Rossen: Should we resolve on a end Jan/early Feb 2019 meeting? That's long

           enough after Lyon TPAC.

   Rossen: We've got one offer for hosting and we'll figure out if anything

           else comes.

   fantasai: Maybe Apple can host?

   myles: Our new building doesn't have anywhere to do that.

   TabAtkins: We discussed yesterday and Mozilla Toronto might host in the

              summer.

   dbaron: It's nice in the summer.



   Rossen: Proposal: End Jan/Early Feb 2019 for the first winter meeting.

   Rossen: If we only want one meeting between that and TPAC it may be late Feb.

   Rossen: So we have a tentative Feb 2019 meeting and we'll have to work out

           logistics.

   Rossen: Objections to winter 2019 meeting, potentially in Sydney?

   RESOLVED: We will have a late jan or in feb 2019 meeting, possibly in Sydney

   Rossen: Based on Sydney we can see if we need to squeeze two more in that

           year or just one.



Accumulating Compat Issues via Ambiguity

========================================



   Topic meta-issue: when we don't act on some ambig we spec due to compat.

   latest example, having flex/grid % margins be based on oldest (block flow)

   model, rather than anything in the past 20 years (positioning, flex, grid).



   Rossen: I'm not sure who put the issue in.

   tantek: I put that in during the telecon where you said Edge will support

           horizontal percentages for Grid.

   astearns: As I recall we set ourselves up for this. We left an ambiguity

             in the spec and let it fester.

   tantek: Worse, we created an ambiguity in the spec.

   astearns: If anyone has ideas on how to avoid in the future.



   florian: Implementations were already disagreeing when we made it ambiguous.

   TabAtkins: Agree. That's when we officially undefined it.

   astearns: Maybe we shouldn't do that.

   fantasai: This was the first time since I've been here where we had a

             disagreement where we couldn't agree what the spec should say.

   fantasai: There are places where we make something undefined because we

             need to move forward, but we plan to align. Every time we've

             had technical discussion we've figured out what the spec should

             say. This is the first time WG was completely split.

   tantek: We had agreement at NY F2F and we resolved and changed the spec

           and then we undid it in Paris. So we started with agreement and

           then, correct me, but TabAtkins brought forward a case and saying

           we disagree.

   shane: I don't think that's what happened in NY.

   TabAtkins: Our implementation disagreed and we said we'd check it out.

   <tantek> Resolution from 2015 NYC f2f

            https://lists.w3.org/Archives/Public/www-style/2015May/0314.html

   <tantek> "RESOLVED: Flexbox top/bottom margins and padding resolve

             against height."



   Rossen: Agreement on symmetric resolution of percents was led by us,

           based on how we wanted the overall model of grid to be.

           One of the guidance principals was we wanted it as symmetric

           as possible. When we discussed we generally agreed this is

           the right modal to have. If it was in a vacuum I don't think

           we would have changed.

   Rossen: However we had a requirement to keep flexbox same as grid.

           When we wanted to backport flexbox it was already widely

           implemented. People were using the percent padding hack for

           aspect ratio and that was a problem for Google when they

           tried to flip to the new model. Google came back and said

           they'd break everything.

   Rossen: It wasn't a disagreement, it was a statement of facts that

           we are going to break the web. What tantek is pushing for

           is we have a model that makes perfect sense and we have new

           people using grid and flexbox and let's have something that

           makes sense.

   Rossen: Fastforward to 2 months ago, we were the only ones shipping

           grid for a long time. When other implementations shipped grid

           we had more and more pressure to fix the bugs.



   <astearns> as much as is possible I would like to limit the

              re-litigation of this particular issue, and instead see

              if there are suggestions on how to avoid this impasse

              in the future

   Rossen: I think this is more of an outlier then a practice we have.

           I don't think we've had this drastic reality. Yes, we must

           avoid this on all counts.

   florian: I don't see how we find a process to avoid getting stuck

            on not agreeing. Once you have the split, what do you do?

   tantek: Other reason given to have the split is the potential utility

           of the aspect ratio hack, if that's something we want we

           should have a feature for that. We as a group didn't follow

           through and that's also partly to blame. If we were to have

           a good proposal that might have been a way out of the compat

           mess. That didn't happen.

   tantek: I understand we've been backlogged on agenda and bottlenecked

           on editors, but with hindsight that might have been avoidable.

   fantasai: There was also a fair number of comments in the bug where

             author feedback was also split. A bunch of authors wanted

             it to work the way MS did and a bunch that wanted Chrome.

   fantasai: Some were about aspect ratio, but some where about other

             aspects of using it. If there had been strong feedback one

             way or the other that would have pushed us a way to go.

             We had a split in the WG and a split in feedback. I don't

             think that's a pattern we tend to follow. Yes, we were

             backlogged, and in hindsight there are always things we

             could have done better. But I don't think this is a

             systematic problem that we have.



   astearns: We can close this saying we don't think this is a good

             pattern and we don't think we normally do it, but if

             anyone sees us falling into it please call it out.

   tantek: Because we had split author feedback combined with that

           the browsers hack were pressured.

   astearns: When we find outselves in a place where we want to leave

             something undefined we need to do analysis.

   tantek: We did analysis. We knew we needed the feature.

   fantasai: People disagreed on what was the right thing.

   florian: It was not the only issue. If aspect ratio had been solved

            it may have helped.

   tantek: To solve it we needed and aspect ratio solution. It might

           not have been enough.



   tantek: If we're ever tempted to undefine we should use this as a

           prediction of what will happen.

   fantasai: We knew this would happen. We'll put both and someone

             will have to switch due to web compat.

   florian: There was a note from the time saying it was sucky.

   tantek: You don't have 2 major implementations sticking with one

           way of doing it for no good reason

   fantasai: We knew it would come to a web compat fist fight. We

             could not come to a consensus. We discussed many times

             for hours. You can't tell us we didn't try. We decided

             this will suck, we have two impl, and web compat will

             drive it in one direction or another. We've never had

             that problem before and I don't expect us to have it

             again.

   fantasai: We'll always have web compat bugs. They're not due to

             intention of the working group. That's not what happened

             here (this was intentional) and makes it different from

             every other case.



   Rossen: I think the lessons have been learned. I don't think we can

           have anything in a more formal way to avoid these sort of

           issues. I don't believe we've made such issues intentionally

           besides this one. I'd encourage anyone who wants to come up

           with a process to put your thoughts on the wiki to say this

           is a dangerous one and let's not repeat the padding/margin

           issue.

   Rossen: It sucks when this happens, users the most, because they get

           crappy output, but that's where we ended.

    * fantasai doesn't think there was a lesson learned, we understood

               exactly what we were getting into at the time, there was

               just no other way forward



   tantek: There's an additional conclusion which is when there is web

           author demand for a feature and that's expressed via a hack

           that should be a strong signal to the group to make it official.

   TabAtkins: tantek that's a spec you could have edited.

   fantasai: You're complaining why didn't anyone submit a patch.

   tantek: Any of us could have done it. I'm making a forward looking

           statement where if we see this in the future it behooves us

           to pay attention.

   Rossen: We can make a new flag in github saying to elevate this.

    * astearns [HACK DETECTED - MUST FIX!!!!!]

   florian: You can't with a missing feature.

   Rossen: Why not?

   tantek: Prior example was advancement with animations and transitions

           back in 2000s with webkit. It was massive changes to the platform

           where we didn't act soon enough and we had to endure a lot of pain.

   Rossen: Let's move on.
Received on Friday, 27 April 2018 06:11:21 UTC

This archive was generated by hypermail 2.3.1 : Friday, 27 April 2018 06:11:22 UTC