See also: IRC log
<anne> http://lists.w3.org/Archives/Public/public-css-testsuite/2009Mar/0042.html
Anne: I suggest talking about test suite [see pointer above]
Chris: SVG understood that CSS would publish it last week, but then they stopped the process when they noticed CSS wasn't publishing.
<dsinger> Dave is puzzled at the lack of reaction to dean's comments to svg
<dsinger> And also puzzled with chrisl
<dsinger> Lack of pib
<dsinger> Pub
<shepazu> dsinger, what do you mean?
<anne> We did not publish because the CSS WG did not formally go on record for publishing.
Chris: We agreed to publish jointly with SVG, but CSS wasn't ready and nothing was published.
DaveS: Why were we not ready?
<ChrisL> Bert, whats up with these four documents?
Bert: I couldn't find any resolution in the minutes, so couldn't publish a 1st WD.
Doug: I couldn't find resolution either, maybe it was just not correctly minuted?
<ChrisL> ok so since we all recall agreeing this a couple of weeks ago lets have a minuted resolution today
Fantasai: There was indeed no resolution, only discussion.
DaveS: We got stuck on talking on one para in 2D.
<ChrisL> hearing no objections here
Steve: We approved 2D provided that para was added.
Doug: SVG really wants to see all 4 published. Can we get resolution on that? 2D, 3D, animation and transition.
<fantasai> Steve: Think we had approval on all but 3d
<fantasai> DavidSinger: We agreed to publish 3d, but make it clear it's on a longer timescale
<ChrisL> so we can publish all four
RESOLUTION: publish all four: Animations, 3D Transforms, 2D Transforms, and Transitions.
Doug: SVG said to Dean already we are very interested in cooperating on all four.
Chris: At last week's SVG's meeting we talked about Dean's comments.
Doug: Yes, the SVG editor has an action to work on them.
<ChrisL> the editor in svg has an action to fold in all deans commentsbefore publication
Doug: Might be good to have joint telcon SVG-CSS. Maybe even a taskforce.
DaveS: But not in the middle of the night for Dean...
<plinss> http://lists.w3.org/Archives/Public/public-css-testsuite/2009Mar/0042.html
Anne: At ftf we decided that if you review a test, you would indicate that. Also decided you could change and somebody else would review that. But not clear from the test itself who reviewed it. So proposal is a small change to test format.
Melinda: So anybody who makes a change should add a link?
Anne: yes, add a "reviewer" link.
Fantasai: "Author" is who wrote the test, maybe has copyright.
Melinda: So are we adding "author" "contributor" or "reviewer" links?
<fantasai> Add "author" if you make a significant contribution to the test
<fantasai> i.e. not fixing a typo or tweaking the title
<fantasai> actually changing the test
Fantasai: [writing in IRC]
<sylvaing> so if I submit a test, then fantasai edits it then anne approves it we'll have...
<fantasai> The problem I had with the reviewer link idea, is that it's not clear when the complete test is reviewed
<sylvaing> a link rel=author for fantasai's edit
<fantasai> for example
<fantasai> a test is submitted with some problems
<fantasai> I review it
<fantasai> it's mostly good
<sylvaing> then a link rel=reviewer for anne ?
<fantasai> but this one part needs a fix
<fantasai> I can fix it and then ask the author to review my chang
<fantasai> in that case
<fantasai> we're both actually reviewers
<anne> sylvaing, yes
<sylvaing> ok
<anne> sylvaing, but you can review it yourself as well
<fantasai> that doesn't help fantasai programmatically figure out whether the test is *approved* yet or not
<anne> (dates should be clear from SVN)
Melinda: So probably the review needs a date field as well.
<fantasai> Anyone can review
<anne> fantasai, "reviewer" means approved
<fantasai> peers will approve the tests and move it over, and that might mean rubber-stamping a review by someone competent
Chris: We have a list of who reviews which chapter in principle.
<anne> fantasai, is what we decided
<sylvaing> right, not sure I as microsoft should review our own tests. am open to reviewing other tests
<fantasai> e.g. if jdaggett is reviewing the fonts test,s I'll assume he's rgith
<fantasai> if someone I've never heard of reviews them, I will probably take another look first
Melinda: No record of who "approves" a test?
<fantasai> before copying the tests into CVS
<fantasai> no, just cvs record for that bit
<fantasai> So
Melinda: So whoever approves must do a CVS check-in?
<fantasai> there's two levels of "review" one is mainly about reviewing the test
<fantasai> the other is mainly checking that the test has been appropriately reviewed
<fantasai> the first level is where we send comments to public-css-testsuite
<fantasai> and mark reviewr in the test case itself
<fantasai> the second level is mainly about copying it into the main repo
<fantasai> it should mean that you looked at the test and approve of it
Peter: Confused about the "reviewer" link: is that marking review or approval? Or both?
<fantasai> at least
<fantasai> that was the goal
Peter: And the "contributor" link?
<fantasai> the main purpose of the reviewer link isn't to say who revieed the test for posterity
<fantasai> it's so that someone knows the test has been approved
<fantasai> "contributor" doesn't exist
<fantasai> we are using "author'
Sylvain: OK, so there is no "contributor." Fine.
<fantasai> The *point*
<fantasai> of this link
<fantasai> was to mark "this test has been reviewed"
Peter & Melinda: Is this for all existing tests as well?
<fantasai> without creating a new system for recording which test have been reviewed
Steve: Do the tests already have "author" links?
Melinda: Yes, I think they do.
<fantasai> yes, they have "author" links
Fantasai: Yes, all tests have "author." I put them in. Goal is not to know who reviewed, but to approve. There were other ideas, such as a wiki page. I don't really care about the mechanism, but we need some way to track. Depending on who is the reviewer I may or may not do another review myself.
Melinda: So "reviewer" means approval.
Steve: Except when reviewer makes changes in the process.
Melinda: But then you would use an "author" link rather than reviewer, wouldn't you?
Steve: OK, I see.
Fantasai: The fixes by a reviewer need to be reviewed, by the original author, e.g. We could put a date in comments or something.
Steve: Idea is to know that last reviewer is not not the same as the last author. Author has signed off that *he* believes it is correct. Just need somebody else after that. Can we put a date field in the link?
Fantasai: A comment on the same line might work.
Steve: Or just an unknown attribute?
Fantasai: It needs to validate.
Fantasai & Steve: OK, so a date in a comment, then.
<fantasai> <link rel="author reviewer" title="Elika Etemad"/> <!-- 2009-03-17 -->
<ChrisL> it will do
<fantasai> meaning "everything looks ok except the stuff I changed, and the stuff I changed needs to be reviewed"
Fantasai: rel="author reviewer" (plus a date) means I reviewed everything except for the parts I changed.
Steve: and thus rel="reviewer" means you did not change anything.
Steve: Can you point to example?
<fantasai> http://wiki.csswg.org/test/css2.1/format
<plinss> http://lists.w3.org/Archives/Public/www-style/2009Mar/0135.html
Fantasai: I agree with the comments Bert sent.
<fantasai> http://lists.w3.org/Archives/Public/www-style/2009Mar/0183.html
Peter: It looks like an itneresting idea. Would it fit? and if so where?
Steve: We have to talk about Grid and Template at the same time. Maybe not urgent and better for a ftf meeting.
Daniel: Is MS still working on Grid? Haven't heard from Alex in a while.
Sylvain: Yes, still interested, but CSS 2.1 takes all resources right now. Alex should be at ftf in June.
Steve: I haven't reviewed the new proposal yet.
<fantasai> Melinda: I think we should put these three proposals side by side and compare their pros and cons
Melinda: Putting the three together at a ftf seems a good idea, indeed. Towards some combination of them.
Steve: And GCPM seems to have some stuff as well.
Bert: Yes moveto/pullfrom and similar ideas.
<fantasai> It seems to me this matrix proposal is just like template layout, except with the added ability to overlap elements
Melnda: There is a need for improved layout techniques, but we need to be clear about our objectives.
Steve: Peter, is that what you meant when you asked about how it fit?
Peter: Yes, ftf seems reasonable. But also wants to know who is interested at the moment.
Fantasai: We should look at the proposal and focus on use cases, but not focus on syntax too much now. Maybe the matrix things can be done by extending layout elsewhere.
Sylvain: I heard there was interest in this stuff.
Steve: It has always been clear that people want this. Less clear if there are implementers for it.
<sylvaing> i.e. web designers came up during and after the CSS3 panel at SXSW to express interest in Jonathan Snook's proposal
Steve: I mean: as a priority.
Peter: So to summarize: I hear interest in evaluating the proposal. Don't hear anything about it being implemented soon. I suggest we pencil it in as a topic for the ftf.
Steve: Somebody should respond to Jonathan to say we probably won't talk about it until June.
Peter: Who will represent the matrix proposal?
Steve: We can invite Jonathan... Other question: are there patents involved?
Bert: Anybody know Jonathan Snook?
All: No, never met him.
<sylvaing> http://snook.ca/
<plinss> http://lists.w3.org/Archives/Public/www-style/2009Mar/0133.html
<plinss> http://lists.w3.org/Archives/Public/www-style/2009Mar/0194.html
Fantasai: David Baron posted proposals.
<Melinda> *jonathan is a member of the CSS 11 ;-)
<fantasai> http://lists.w3.org/Archives/Public/www-style/2009Mar/0194.html
<fantasai> I'm in favor of option 1
Steve: I like option 1
Fantasai: Same question for other keywords, such as 'inherit'
<fantasai> so any objections? :)
Bert: Leaning to option 1 as well.
Melinda: So what does this mean for 'inherit'?
Steve: Can't use it is a counter name.
Chris: Can you escape it? With a backslash?
Fantasai: No.
Sylvain: What's the use case for 'counter-increment: none'?
Fantasai: I can't think of a reason for a counter named "none", but I can certainly see a case to explicitly set 'counter-increment' to 'none' to stop the counter from incrementing.
Sylvain: I don't get the 'counter-increment: none 1' rule.
Peter: That is just invalid. I think that's implied by the prose, but not explicit.
Melinda: We need some words to describe 'none' then.
Peter: Yes, agree.
... Bert, can you write text?
Bert: OK.
Fantasai: I'll note it in the issues list.
Peter: Should be enough if Bert sends it to www-style. We'll see what comments, if any, it gets.
<szilles> Steve has probable regrets for next week due to AB meeting
RESOLUTION: accept proposal 1, Bert to come up with wording, submit for review to www-style; no objections means accepted.