IRC log of html-wg on 2007-05-13

Timestamps are in UTC.

00:03:39 [hasather]
hasather has left #html-wg
00:06:56 [ddailey]
ddailey has joined #html-wg
00:09:09 [ddailey]
I agree with anne -- http://www.456bereastreet.com/archive/200705/help_keep_accessibility_and_semantics_in_html/#comment57 makes very good sense
00:10:25 [Dashiva]
anne = mattur?
00:13:19 [ddailey]
Not sure ... at [16:32] in http://krijnhoetmer.nl/irc-logs/html-wg/20070512
00:13:53 [ddailey]
I assumed this was anne -- but it may have been someone presuming to speak for anne
00:15:16 [hsivonen]
ddailey: anne said he liked the comment
00:15:35 [hsivonen]
ddailey: doh. I misread what you sai
00:15:37 [hsivonen]
d
00:15:40 [hsivonen]
never mind
00:15:53 [ddailey]
hsivonen: I liked it too
00:16:43 [othermaciej]
I dunno who "mattur" is, but he or she makes a good point
00:19:40 [ddailey]
He she or it makes it almost self evident.
00:22:21 [ddailey]
Perhaps the other perspective fears that a descriptivist grammar is mutagenic
00:23:08 [Dashiva]
Indeed
00:24:22 [ddailey]
might there be some way to convince people that the future will not degrade into chaos?
00:25:43 [MikeSmith]
othermaciej - /me wishes you had auto-update for Webkit nightlies
00:26:25 [Philip`]
http://web.mac.com/reinholdpenner/iWeb/Software/NightShift.html ?
00:26:30 [othermaciej]
MikeSmith: NightShift
00:26:36 [othermaciej]
I should add a link to webkit.org
00:27:46 [ddailey]
for example: showing a migration path that includes ample reassurance that life will still be good after 2010
00:28:33 [ddailey]
are nightly builds for webkit equivalent to nightly builds for safari?
00:28:49 [ddailey]
or how closely coupled are the two?
00:30:34 [Philip`]
ddailey: I think http://webkit.org/blog/101/back-to-basics/ tries to explain that
00:30:55 [ddailey]
thanks Philip`
00:30:57 [MikeSmith]
othermaciej - just downloaded NightShift ... very cool -- thanks
00:34:23 [ddailey]
othermaciej and colleagues may be pleased to know that some "customers" now just assume that Safari / OSX widgets exist everywhere. Reminds me of 1985.
00:36:16 [ddailey]
Mac-aware users started assuming a GUI was possible everywhere
00:42:08 [ddailey]
they rather succeeded in convincing others that it was worth looking at
01:17:29 [ddailey]
ddailey has left #html-wg
01:31:49 [gavin_]
gavin_ has joined #html-wg
01:53:24 [MikeSmith]
MikeSmith has joined #html-wg
03:39:13 [gavin_]
gavin_ has joined #html-wg
04:55:13 [hyatt]
hyatt has joined #html-wg
05:46:20 [gavin_]
gavin_ has joined #html-wg
06:11:52 [hyatt]
hyatt has joined #html-wg
06:25:18 [Zeros]
Zeros has joined #html-wg
06:28:43 [dbaron]
dbaron has joined #html-wg
06:40:18 [Roger]
Roger has joined #html-wg
07:04:41 [Lachy]
Lachy has joined #html-wg
07:34:18 [Lachy]
Lachy has joined #html-wg
07:53:19 [gavin_]
gavin_ has joined #html-wg
08:03:02 [myakura]
myakura has joined #html-wg
08:16:10 [myakura]
myakura has joined #html-wg
08:40:06 [ROBOd]
ROBOd has joined #html-wg
08:48:48 [zcorpan_]
zcorpan_ has joined #html-wg
09:07:04 [Shunsuke]
Shunsuke has joined #html-wg
09:45:45 [Shunsuke]
Shunsuke has joined #html-wg
09:45:49 [hasather]
hasather has joined #html-wg
09:58:32 [tH]
tH has joined #html-wg
10:00:46 [gavin_]
gavin_ has joined #html-wg
10:04:14 [jdandrea]
jdandrea has joined #html-wg
10:16:48 [Shunsuke]
Shunsuke has joined #html-wg
10:24:57 [Shunsuke]
Shunsuke has joined #html-wg
11:01:17 [loic]
loic has joined #html-wg
11:14:07 [Shunsuke]
Shunsuke has joined #html-wg
12:07:37 [gavin_]
gavin_ has joined #html-wg
12:31:08 [ddailey]
ddailey has joined #html-wg
12:31:30 [ddailey]
ddailey has left #html-wg
13:20:05 [Shunsuke]
Shunsuke has joined #html-wg
13:37:18 [dk]
dk has joined #html-wg
13:56:42 [gsnedders]
gsnedders has joined #html-wg
14:30:32 [Roger]
Roger has joined #html-wg
14:32:07 [gavin_]
gavin_ has joined #html-wg
14:51:31 [zcorpan_]
zcorpan_ has joined #html-wg
15:07:11 [Lachy]
Lachy has joined #html-wg
16:35:13 [zcorpan_]
zcorpan_ has joined #html-wg
16:39:05 [gavin_]
gavin_ has joined #html-wg
17:47:13 [dbaron]
dbaron has joined #html-wg
18:11:18 [zcorpan_]
zcorpan_ has joined #html-wg
18:14:08 [tH]
tH has joined #html-wg
18:26:43 [kingryan]
kingryan has joined #html-wg
18:45:53 [sbuluf]
sbuluf has joined #html-wg
19:20:11 [hasather]
hasather has left #html-wg
19:20:40 [hasather]
hasather has joined #html-wg
19:32:31 [dbaron]
ok, I'm unsubscribing from www-html
19:32:48 [dbaron]
tired of the duplicate messages, and participation by reading occasional messages doesn't seem to be welcome
19:35:12 [anne]
FWIW: I agree with you that it would be nice if HTML5 defined all features of HTML in use on the web
19:35:37 [dbaron]
in use on the Web is a little strong
19:35:50 [dbaron]
that includes a lot of proprietary stuff that is very rarely used
19:37:49 [Lachy]
dbaron, what gave you the idea that your input wasn't welcome? It just seemed like a slight misunderstanding of what was being discussed.
19:38:08 [dbaron]
how was I supposed to know that that was being discussed?
19:38:18 [dbaron]
It's not mentioned in your message or any of the ones preceding it in the thread
19:38:36 [dbaron]
you can't have a 300 person discussion without being clear about what you're saying ,sorry
19:39:32 [Lachy]
I just thought it was implied that we were talking about what authors could use
19:40:02 [anne]
Hmm, yeah, I guess rarely used stuff should be excluded somehow. Like <t:video>...
19:40:50 [anne]
I meant things like <frameset>, <marquee>, <body vlink>, etc.
19:47:16 [othermaciej]
othermaciej has joined #html-wg
19:59:39 [tH]
tH has joined #html-wg
20:02:17 [zdenko]
zdenko has joined #html-wg
20:06:48 [schepers]
dbaron, please don't unsubscribe... people like you are crucial to HTML5's success
20:13:29 [anne]
www-html isn't
20:14:48 [schepers]
oh, right, ok
20:16:34 [tH]
tH has joined #html-wg
20:18:28 [hsivonen]
oh. there's been a whole thread on www-html since I unsubscribed
20:18:59 [anne]
same thing happened when I unsubscribed a month ago :)
20:20:37 [hsivonen]
anne: well, I unsubscribed yesterday
20:21:48 [asbjornu]
asbjornu has left #html-wg
20:25:24 [zcorpan_]
zcorpan_ has joined #html-wg
20:25:26 [gavin_]
gavin_ has joined #html-wg
20:45:40 [myakura]
myakura has joined #html-wg
21:00:24 [jdandrea_]
jdandrea_ has joined #html-wg
21:14:55 [hyatt]
hyatt has joined #html-wg
21:16:43 [hyatt]
dbaron: you there?
21:19:38 [dbaron]
hyatt, yes
21:23:35 [hyatt]
dbaron: i sent out some more details
21:23:41 [hyatt]
about the idea for this baseline property
21:23:46 [hyatt]
i actually don't think it would be inherited.
21:24:00 [hyatt]
dbaron: in terms of webkit, we basically have a method on our renderobjects called baselinePosition
21:24:14 [hyatt]
for inline flows/blocks it's font.ascent() + leading / 2
21:24:22 [hyatt]
for replaced elements it is margin top + height + margin bottom
21:24:30 [hyatt]
for inline blocks it is the last line box's baseline etc.
21:24:43 [hyatt]
i basically want a CSS property that lets me define that value
21:24:49 [hyatt]
instead of it being hardcoded for the various element types
21:25:02 [hyatt]
so i don't think the property would need to be inherited
21:25:05 [hyatt]
since when set on say a <div>
21:25:18 [hyatt]
anyone who needed to vertically align to the baseline would align to that spot
21:25:49 [dbaron]
well, that wouldn't address your use case, I don't think, since it wouldn't affect the position of the first line when in the normal flow
21:26:07 [hyatt]
it wouldn't?
21:26:19 [hyatt]
when set on a block it would establish the baseline's position
21:26:23 [dbaron]
Blocks don't normally care about their block baseline
21:26:43 [hyatt]
well, it would establish the baseline of the "root line"
21:26:53 [hyatt]
e.g., in the sense that :first-line { font-size: 24px }
21:27:01 [dbaron]
so why shouldn't it do the same for inline-blocks?
21:27:25 [dbaron]
your baselinePosition method seems to mix the baseline as applied to the inside (block) with baseline as applied to the outside (inline-block)
21:27:26 [hyatt]
well, inline-blocks have a duality that makes it hard to decide what the property would mean
21:27:41 [hyatt]
our baselinePosition method in webkit takes a bool
21:27:42 [hyatt]
it's dual
21:27:48 [hyatt]
so good point.
21:27:58 [hyatt]
an inline-block has two baselinePositions in our code
21:28:00 [hyatt]
the outside one and the inside one
21:28:14 [dbaron]
one option is to make the property apply only to blocks, and set their baseline-inside
21:28:20 [dbaron]
and not be inherited
21:28:23 [hyatt]
yeah
21:28:25 [hyatt]
that would work
21:28:35 [hyatt]
i think having both props is useful
21:28:40 [hyatt]
one thing we had to hardcode for example
21:28:44 [hyatt]
is the baseline position of checkboxes
21:28:48 [dbaron]
this seems the least dangerous in terms of messing up the idea that the model generally tries to lead to things not overlapping
21:28:52 [hyatt]
somethign like a checkbox is an "image"
21:29:03 [hyatt]
but it needs to establish a baseline other than marginTOp + height + marginbottom
21:29:08 [dbaron]
but it would still lead to bad results if a large image or something in a bigger font ended up in the first line
21:29:13 [hyatt]
since you'd like to be able to set a baseline at an aesthetically pleasing point
21:29:30 [hyatt]
dbaron: yeah this property is all about pixel level control
21:29:34 [hyatt]
which is giving the author rope
21:29:39 [hyatt]
potentially to hang themselves
21:29:56 [hyatt]
it may be unsuitable for a standard
21:30:00 [hyatt]
i don't feel strongly about it
21:30:10 [hyatt]
we can always add in the hack just for the people who need it
21:30:15 [hyatt]
but i thought it was worth bringing up
21:30:46 [dbaron]
ok; I was scared you were trying to push it in and were already implementing
21:30:52 [hyatt]
(this is post-leopard too, so there's not any rush)
21:30:57 [hyatt]
nope
21:34:11 [Philip`]
Philip` has joined #html-wg
21:41:40 [hyatt]
dbaron: i'd point out that line-height is "dangerous" in that it can easily cause lines to overlap
21:46:06 [dbaron]
hyatt, though only when less than 1
21:46:16 [hyatt]
irght
21:46:45 [dbaron]
or less than 'normal', anyway
21:47:47 [hyatt]
i wish the css wg wasn't closed
21:47:55 [hyatt]
i hate that all these conversations take place on a private list
21:48:01 [hyatt]
i should send stuff to www-style instead
21:49:34 [hyatt]
Hixie: ping
21:50:44 [hyatt]
Hixie: regarding your comment in http://bugs.webkit.org/show_bug.cgi?id=13696
21:50:57 [hyatt]
Hixie: html5 was more restrictive when compared with firefox
21:51:02 [hyatt]
Hixie: i went ahead and matched firefox
21:51:07 [hyatt]
which reopened in more circumstances
21:55:11 [hyatt]
Hixie: i'd be curious to try my test case in WinIE
21:55:29 [hyatt]
Hixie: this is not about the set of tags that reopen
21:55:45 [hyatt]
Hixie: this is about the set of tags that (when closed) cause those tags to reopen when popping
21:55:57 [hyatt]
which should be nearly everything
21:56:11 [hyatt]
except for table stuff and selects basically
21:56:24 [hyatt]
(and for some odd reason ffx didn't reopen object but did for embed, applet, canvas)
21:56:27 [hyatt]
(so i matched ffx)
21:56:37 [hyatt]
reopen across </object>
22:07:00 [zcorpan_]
http://simon.html5.org/test/html/parsing/residual.html
22:07:21 [zcorpan_]
TABLE is red in ie7
22:07:32 [zcorpan_]
COL and COLGROUP too
22:07:55 [zcorpan_]
after TABLE, everything is green (also things that probably should be black)
22:13:24 [hyatt]
wow so winie reopens even more than ffx does
22:14:35 [zcorpan_]
at the TABLE, it nests the rest of the document inside that FONT
22:14:52 [hyatt]
oh i may have missed a close
22:14:56 [hyatt]
or it could have to do with how it recovered
22:15:20 [hyatt]
if you see red, it means WinIE reopened
22:15:23 [zcorpan_]
TABLE: <font><table><font class="fail"></table>All of this should be green.</font>
22:15:37 [hyatt]
oh i should just throw in an extra </font> in those examples
22:15:48 [hyatt]
</font></font>
22:15:56 [hyatt]
since it will be harmless in the other browsers
22:17:04 [zcorpan_]
in opera everything is green after the OBJECT
22:18:44 [hyatt]
hey if you cleaned that test up and noted what each browser does that would rule
22:18:58 [hyatt]
the test right now basically shows what ffx and safari d
22:18:59 [hyatt]
o
22:18:59 [zcorpan_]
sure
22:21:33 [zcorpan_]
test cleaned up, also uploaded to live dom viewer
22:22:13 [zcorpan_]
ok, ie7 again...
22:23:33 [zcorpan_]
these are black: APPLET BUTTON MARQUEE. these are red: COL COLGROUP TABLE
22:24:25 [hyatt]
fascinating
22:24:40 [hyatt]
wow i should add applet button marquee
22:24:44 [hyatt]
i should probably just add embed too
22:25:05 [hyatt]
i'm scared to reopen col/colgroup/table
22:25:08 [hyatt]
even if winie doe
22:30:01 [zcorpan_]
zcorpan_ has joined #html-wg
22:31:08 [zcorpan_]
ie6 is same as ie7
22:31:15 [zcorpan_]
all are green in Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9a5pre) Gecko/20070503 Minefield/3.0a5pre
22:32:59 [zcorpan_]
opera 9.20 build 8771: these don't include the test text at all: APPLET FIELDSET LEGEND. these are black: BUTTON CANVAS. these are red: OBJECT PARAM
22:33:20 [gavin_]
gavin_ has joined #html-wg
22:41:01 [zcorpan]
zcorpan has joined #html-wg
23:20:33 [Philip`]
zcorpan: In what way is it lying?
23:20:45 [Philip`]
(About non-tree DOM, I guess?)
23:21:15 [zcorpan]
Philip`: for the residual style test, it claims that everything is nested inside each FONT as a very deep tree
23:21:48 [Philip`]
Ah
23:58:21 [hyatt]
hyatt has joined #html-wg