- From: fantasai <fantasai.lists@inkedblade.net>
- Date: Thu, 08 Dec 2011 14:45:58 -0800
- To: "www-style@w3.org" <www-style@w3.org>
Unless you're correcting the minutes,
*Please respond by starting a new thread with an appropriate subject line.*
Summary:
- UTR50 review period closes in January. If you have feedback on orienting
glyphs in vertical layout, make sure you review it and send comments by
then.
- RESOLUTION: column-spanning elements can only span when the closest
ancestor BFC is established by the multicol (whether by
the columns or the multicol)
- RESOLVED: each column spanning element establishes a separate BFC (option C);
margins between them collapse
- RESOLVED: Glenn and Shane coedit cssom and cssom-view
- RESOLVED: Florian and Sylvain become coeditors of css3-mediaqueries
- fantasai to check in update to http://www.w3.org/Style/CSS/current-work
====== Full minutes below ======
Present:
Glenn Adams
Rossen Atanassov
Tab Atkins (late)
David Baron
Bert Bos
Tantek �elik
John Daggett
Elika Etemad
Simon Fraser
Sylvain Galineau
Daniel Glazman
Vincent Hardy
Koji Ishii
John Jansen
Brad Kemper
H�kon Wium Lie
Peter Linss
Eric Mueller
Edward O'Connor
Anton Prowse
Florian Rivoal
Alan Stearns
Steve Zilles
If you're missing from this list, then next time either a) tell Zakim you're
here or b) say something that's minuted so I know you're present. :)
~fantasai
Regrets:
C�sar Acebal
Kimberly Blessing
Chris Lilley
David Storey
Daniel Weck
<RRSAgent> logging to http://www.w3.org/2011/12/07-css-irc
Scribe: dbaron
Meeting: CSS WG Teleconference
Chair: Daniel Glazman
Scribe: David Baron
Agenda
------
glazou: any extra items? Tab wanted to add item on switching back to
fantasai's current-work listing
glazou: I suggest doing that after the high-priority items.
Bucharest meeting in May
------------------------
glazou: Sent email to list to confirm the dates of the meeting.
glazou: Vincent, can we confirm the dates?
Vincent: May 9-10-11 (Wed-Thu-Fri)
Vincent: FX meeting with SVG would be Wednesday morning
glazou: When to expect hotel recommendations?
ACTION Vincent provide recommended hotels for Bucharest meeting ASAP
<trackbot> Created ACTION-407
jdaggett: is there a wiki page with address of venue, etc.?
ACTION vincent to make wiki page for Bucharest with location of meeting, etc.
<trackbot> Created ACTION-408
<tantek> I updated http://wiki.csswg.org/planning/2012 per the confirmed
Bucharest dates above.
Multicol spanner margins
------------------------
<glazou> http://lists.w3.org/Archives/Member/w3c-css-wg/2011OctDec/0138.html
<glazou> http://www.w3.org/mid/4ED416B1.7070902@inkedblade.net
glazou: previously waiting for fantasai to post blog
fantasai: That's been done
<fantasai> http://www.css3.info/multi-column-margin-collapse/
glazou: We decided to make a decision this week.
rossen: Can we do this as the second item, in 5 minutes?
Update on Unicode TR50
----------------------
<glazou> https://lists.w3.org/Archives/Member/w3c-css-wg/2011OctDec/0249.html
jdaggett: This is text-orientation; action was for Sylvain and Ted to
get feedback from Microsoft and Apple.
sylvain: ... would come back to me with details on latest version of note.
Chatted with Elika last time. Sergei will have another look at it
and I'll share what he says on the list.
sylvain: We've provided feedback in the past; Sergei's been busy with other
things.
jdaggett: At the UTC meeting there were some MSFT reps. Peter Constable ...
he'd tell you who else was there.
jdaggett: They were talking about other proposals.
fantasai: What they were proposing was different from what Sergei was saying
when I talked to him.
Ted: I've got the conversation going internally, waiting to get more useful
feedback for list/wiki.
Ted: Like Sylvain I don't have the knowledge myself.
Ted: My knee-jerk reaction is that if WebKit and IE agree we should go
with that, but I'll have some feedback on the list as soon as I can.
jdaggett: Especially helpful would be if there are things that seem bad
to you about the actual proposal.
Sylvain: What kind of timeline? Something needed before the new year?
jdaggett: Concerned about that, since second round of comments has been
extended to mid-January, but if we're not careful we'll miss that.
glazou: need an action?
jdaggett: have 2 existing
Multicol spanning margins
-------------------------
<glazou> https://lists.w3.org/Archives/Member/w3c-css-wg/2011OctDec/0138.html
<glazou> ACTION howcome to repost his message to www-style
<trackbot> Created ACTION-409
See http://lists.w3.org/Archives/Public/www-style/2011Sep/0509.html
howcome: Certainly an issue we'd like to settle, hopefully today.
howcome: I don't see this as a major issue; it's a corner case, but will
have implications for authors.
howcome: We'd published a blog on the topic. I'm not quite up-to-date with
the feedback on that post, but it's been made.
<fantasai> http://www.css3.info/multi-column-margin-collapse/
fantasai: I can summarize the feedback.
fantasai: The blog post tried to get people to imagine the scenario we're
envisioning, and shows the 2 options we're considering.
fantasai: Most of the comments say collapsed margins for consistency with
the rest of CSS.
fantasai: A few suggest no collapsing.
fantasai: A few wanted not collapsing for consistency with block flow
(which doesn't make sense).
florian: A few said they wanted no collapsing in CSS.
fantasai: And some suggestions for a margin collapsing control property.
fantasai: But most suggestions seemed to want collapsing just like regular
paragraphs.
howcome: I think if the example had column-span set to 2 out of 3 columns,
it might have been slightly different. That's a futuristic case.
florian: Even if we agree that collapsing is better, it doesn't tell us
whether we should prefer solution A or C.
fantasai: I was talking with Kimberly while we were working on this blog
post. The mental model she had (with picture) was that you
have a multicolumn element, then you have a row of columns before
the spanner, then the spanners, and then a row of multicolumn
elements after the spanners. The model was that the row of
columns was a row of columns but it behaved as a block-level
element that was a sibling of all of the spanners.
fantasai: And inside the block-level element you had regular block flow
with the rows of columns being a special block-level box.
anton: At the moment that mental model makes sense because columns can't
have vertical margin because they can't be targeted with a selector,
but in future they might be able to be targeted.
anton: So if the columns themselves had bottom margin, would we expect that
to collapse with whatever comes next?
anton: I'd expect the margins on the columns themselves not to collapse.
anton: I think what's important is the inter-spanner relationship rather
than the beginning/end of the spanners.
* fantasai thinks allowing margins on columns would be like allowing margins
on table cells, i.e. wouldn't make sense even if we allowed
styling those boxes
florian: Even for inter-spanner behavior A and C propose different things:
margin collapsing was the same but floating was not.
<glazou> slower antonp please
anton: Makes sense to allow floats to behave as in normal block flow.
rossen: Would you expect floats to expect flow of column?
?: no, not in flow of column
anton: spanners in A or C are wrapped in a BFC. Question is whether each
wrapped independently or all in one.
rossen: In B you don't have a BFC; they are BFC.
florian: B is ruled out by the poll
...
anton: If there's just one spanner it's still wrapped in a BFC (in A), but
if there are 2 or 3 they would all be wrapped in a BFC.
<florianr> In A, spanners are not individually BFCs, but their are together
wrapped in an anonymous one
<florianr> in C, each spanner is a BFC
H�kon: my preference is C
<fantasai> D, each column row is a BFC and everything else just behaves
like regular block flow
rossen: In the blog post the example is oversimplified; just text and
spanners. Would like to see example that's more complicated, e.g.,
tables in the column and the spanners coming from deep inside the
tables.
florian: That's probably something we don't want to support at all.
rossen: Then I'd want spanners to come only from the BFC level of the column.
florian: After talking w/ implementors, would be comfortable with that.
sylvain: Things become really weird otherwise.
florian: The property just doesn't do anything when you apply it on something
"too deep"
anton: Restrict it to the BFC. A spanner can't escape from a BFC.
florian: We can argue back and forth; we certainly want to forbid things
that are way too deep like inside a table.
rossen: The first time I looked at it, the deeper structures were the
problem I ran into it. That makes collapsing pretty hairy.
rossen: Everyone seems to be ignoring the general case.
rossen: Either we say this is level 1 only or ??? ???
florian: I don't think anyone wants to span things that come from deep down,
and I think we should resolve on that.
H�kon: ...
Sylvain: We just need to define what Rossen means.
fantasai: I suggest we resolve that the spanner has to be in the same BFC
as the main level of the column content.
H�kon: someone suggested making each column a BFC
anton: but I've gone off that idea
glazou: I'm almost hearing consensus.
Rossen: BFC or non-BFC-ness of spanners themselves... not resolved
florian: we should resolve on that first
Sylvain: I've heard a couple of definitions of level 1 already.
Sylvain & florian talk at the same time
Sylvain: We agree that spanning should be scoped at some level.
Rossen: to the BFC of the column
H�kon: definition of spanning is that the element spans across all columns
of the nearest multicol ancestor of the same BFC
<howcome> The element spans across all columns of the nearest multicol
<howcome> ancestor of the same block formatting context.
anton: we might need to tinker with that wording
anton: question is whether spanner can escape inline-block
<howcome> Proposed definition of spanner: The element spans across all
columns of the nearest multicol ancestor of the same block
formatting context.
dbaron: How could a multi-column not establish a BFC?
RESOLUTION: column-spanning elements can only span when the closest
ancestor BFC is established by the multicol (whether by
the columns or the multicol)
?: need an action for somebody to propose wording for this
Rossen: I can write it
ACTION rossen propose wording for column-spanning elements can only span
when the closest ancestor BFC is established by the multicol
(whether by the columns or the multicol)
<trackbot> Created ACTION-410
glazou: Back to the choice between A and C.
fantasai: and D
fantasai: D is where the multicolumn element establishes a block formatting
context and column spanning elements are treated as regular blocks
and each column row is a block level BFC within the multicol BFC
fantasai: It's like the table element having an outer table that has the
captions and the table box
florian: This D model wouldn't play nice with the ability to select
individual columns and do things with them (in the future).
rossen: Especially if multicolumns are going towards ??? columns that
H�kon was proposing.
H�kon: Can't we just pick one of A and C ?
fantasai: That's just what came out of the discussion I had with Kimberly.
Though we didn't discuss floats.
Rossen: On option A, if we were going to go with the non-BFC behavior where
floats can affect subsequent spanning elements, what would be
drawing order and who would be drawing floats?
Rossen: You meant(?) anonymous BFC; in our implementation we have nothing
of this sort.
Rossen: Option C seems fairly consistent: spanners will collapse margins
between themselves and keep everything else to themselves.
Florian: There is another difference between A and C: if the spanner
itself has children do the margins of the children collapse with
things in the next spanner. In C they don't; in A they do.
Rossen/Florian: Also if there is an empty spanner between 2 non-empty
spanners
Anton: It depends what you want these spanners to be. If you want them
to look like normal block formatting then they ought to collapse.
If they're each individually special then it's fine that they don't.
Rossen: I think they're each individually special.
+TabAtkins
Rossen: At the end of the day we're taking the spanners out of the flow
and collapsing the margins between the spanners themselves.
Rossen: Whether or not we have to treate empty spanners the way we treat
empty blocks today. But if we're taking them out of the block
flow and collapsing margins in between them, then I don't see a
reason to make them non-BFC.
Rossen: There's no other precedent for taking anything out of the flow
that isn't a BFC.
Anton: I'm not entirely convinced we're taking stuff out of the flow here.
Anton: If you've got a spanner you're ending the columns and then starting
them again.
Anton: They disrupt the multicol, but they're not out of the flow.
dbaron: I think they're out of the flow.
Rossen: They're out of the flow in our implementation apparently.
fantasai: I see this more like a block-in-inline case.
fantasai: Jumping out to an outer flow.
<Bert> I think D works, but there appears to be very little difference
with A in the visual effect.
<fantasai> Bert, except when there are margins on the children of the spanner
<Bert> Ah right. So then I think I actually like D better. :-)
<fantasai> Oh, wait, no I think you're right!
<fantasai> A and D are equivalent
<fantasai> C and D are different
H�kon: I think we have consensus for C. I don't hear anyone arguing for A.
various: does anyone object to C?
SteveZ: Only mildly.
SteveZ: I think one of the things fantasai just said: treatment of blocks
in an inline. If you look at is a headings it doesn't make much
sense. But if you look at it as temporarily switching back to
single-column, it seems like the user would want those pieces to
behave as inside one single column.
Florian: If you want that, you can have a containing element be the spanner
rather than make several consecutive elements spanners.
H�kon: As long as you can insert a div around ...
Steve: I'm not strong on this, just wondering what users will expect.
H�kon: Float behavior will be different, that's true.
glazou: Given constraints, I think authors won't meet expectations anyway.
fantasai: Bert points A and D are equivalent, so I'm leaning towards A.
fantasai: If we take the premise that a column row stretches across the
entire column and clears all the floats before it, then A
expresses that behavior.
fantasai: Interrupting the column flow and going to a flow that stretches
across the entire block... can resume multicol afterwards.
fantasai: Within the anonymous BFC everything is a regular block.
fantasai: Nothing different from ... except border of multicol box goes
around everything.
dbaron: Considering the possibility that we might later move to having
column spans that don't cross all columns, I think it's much
better to think of each column spanning element as isolated --
I'm scared of doing otherwise. Thus I prefer option C.
H�kon: implemented ... . ...
Florian: Could have strange cases: span:all followed by span:3 would lead
to weird results if you have 3 columns
fantasai: I'm happy with C.
Steve: I can live with C.
glazou: anyone objecting?
Florian: Alex, but he's not here?
RESOLVED: each column spanning element establishes a separate BFC (option C)
Editorship of cssom and cssom-view
----------------------------------
glazou: Anne left the group.
glazou: We need editors for these documents.
glazou: Proposal from Glenn to be coeditor.
Tab: At last TPAC Shane Stevens offered to edit as well.
jdaggett: Florian offered?
Florian: For Media Queries
Sylvain: I can help with MQ too.
<glenn> is here
jdaggett: OM is kind of a key spec; also to some extent OM-view. Is that
the right spec for people new to the group? Would be more
comfortable with somebody with more familiarity.
Tab: With Shane, I expect I'd be acting as a mentor for that spec.
jdaggett: I'd feel better if he was working on different specs.
Glenn: I've implemented cssom and cssom-view and CSS formatting semantics
in 2 or 3 products.
glazou: Tab, what's Shane's opinion?
Tab: He's fine with Glenn being a coeditor.
Glenn: I'd suggest both Shane and I be assigned as coeditors as a starting
point, and if others want to help we can change that over time.
Florian: I think Shane said he was interested in documenting existing
compatible bits.
Tab: Yeah, he dosen't want to start working on new stuff until we get
the existing stuff documented & stable
Sylvain: I'm more concerned about what we're working on rather than who's
working on them.
Sylvain: These are fundamental specs, but were neglected for a long time.
Glenn: One reason I voluneteered because I'm working with an external org
called DLNA, which normatively references both of these, and
identified these as needing significant work to get to CR.
Sylvain: One issue recently was that DLNA was depending on working drafts,
and complained when they changed. Are we going to have shenanigans
of that sort if a draft changes what the previous draft said?
Glenn: You'd formally objected to a change in css3-fonts because it was
incompatible with a prior editor's draft.
Glenn: Formally, DLNA policy does not allow normative ref to anything other
than final spec (REC in W3C). They're interested in participating
to move all the dependencies forward.
Glenn: The css3-fonts issue I brought up while representing Samsung has
been closed as far as I'm concerned. I'm now representing Cox
Communications in this WG.
Glenn: I wish to help to move to REC as fast as possible not only these
specs but other specs I can help with.
glazou: What this WG would like to see is a schedule for these documents.
They've been orphaned for a long time; they're crucial for CSS.
glazou: Do you think this is doable? Doing the steps to move these
documents along the rec track.
Glazou: At least by the end of the year, I'd like to provide a proposed
schedule for the work.
Florian: Seems like Shane is interested in documenting the stable bits,
and Glenn interested in moving to Rec
Sylvain: What is the work? Reduce to what's implemented? Values API?
Tab: I think 2.1-style: reduce to what's implemented
Tab: New stuff needs to be in CSSOM level 2.
glazou: I agree with that
dbaron: I'd be concerned about moving the new stuff to REC "as fast as
possible", but if we're splitting that out I have no concern.
glazou: I think I'm hearing consensus.
RESOLVED: Glenn and Shane coedit cssom and cssom-view, Florian and Sylvain
become coeditors of css3-mediaqueries, and we will revisit
schedules in 2 weeks
<glenn> thanks, looking forward to accomplishing this work in a timely manner
Moving stuff from css3-text to level 4
--------------------------------------
<glazou> http://lists.w3.org/Archives/Public/www-style/2011Nov/0395.html
fantasai: That'll take more than 2 minutes
Tab: unless we agree to just do what fantasai's saying
current-work
------------
Tab: It's been more than a month; new page should be up.
Tab: what do we need to do to make this happen?
Tab: Bert, if there's still work you need to do, let me help you with it
Bert: I started discussing with fantasai on common list of drafts.
Long to do list. The content is what I'm worried about.
Tab: The content that fantasai proposed is much more useful than what's
there right now.
Tab: we can tweak it after it's up
glazou: I agree with that.
Bert: I'm not so sure that content is useful.
glazou: We had a resolution on that
* fantasai notes we don't actually, we only have an action item
Bert: You had a resolution that you think it's more useful.
glazou: the group
Bert: I'd like to find some way to integrate that. The original text that
Elika proposed is not complete, and there are some distinctions that
I think shouldn't be made.
Tab: Let us do that afterwards.
glazou: It's better.
Bert: I don't see that -- there are so many categories: what do they mean?
Tab: They mean the English names of the categories, and arranged in order
of stabilization.
glazou: we have a resolution
fantasai: not technically...
Bert: I have a lot of freedom in making these pages, but I still have
responsibility there: I'm making them on behalf of the W3C not on
behalf of the working group.
Bert: We don't have a list that ... happy with ... agreed on.
fantasai: You sent me a list that was a slightly modified version of mine.
We seem to be pretty close with the exception of naming one of
the sections. Why can't we move to that?
Bert: I haven't looked at your list with that in mind
fantasai: Only changes I suggested to that was keeping hte section on
abandoned specs, and switching the location of css3-ui in list
Bert: I think that list is fine.
glazou: This is part of the outreach of the wg
glazou: We have to close that issue
glazou: Bert, I want you to do that change as soon as possible.
Bert: It will be done within the next week and a half.
glazou: That's the best you can do?
Bert: I have things to do tomorrow and the day after.
Tab: Can one of us publish the page, then?
Bert: I guess fantasai can.
glazou: Let's do that.
fantasai: I don't know the structural dependencies
Bert: keep ids 'completed' and 'high-prio'
ACTION fantasai update the current-work page
<trackbot> Created ACTION-411
Received on Thursday, 8 December 2011 22:46:29 UTC