00:00
gsnedders
wonders if it does anything
00:00
<zewt>
it should warn if you *don't* have typos
00:02
<gsnedders>
Okay, it doesn't do anything.
00:02
<zewt>
seems strange that browsers still don't allow copying out of password fields
00:02
<zewt>
even though every browser lets you get at it in inspector tools anyway
00:10
<Hixie>
WeirdAl: they keep fricking changing and keeping them up to date is of no value to me whatsoever
00:11
<WeirdAl>
I'll admit, it probably looks good on a thousand resumes.
00:11
<WeirdAl>
but other than that...
00:13
<Hixie>
not the acknowledgements, i mean the references section
00:13
<WeirdAl>
oh
00:13
<Hixie>
certainly not going to remove the names from the acknowledgements section, listing those names is the least i can do given how much help everyone has been
00:16
<Hixie>
annevk: btw i checked and i'm still sending your script the "typo"/"xref" thing, even when i'm sending the svn repo and e-mails a longer description
06:03
<hsivonen>
hober: The OAuth story really calls for "if you take your spec to the IETF, you're going to have a bad time", despite the frequency of the appearance of the super-cool ski instructor.
06:04
<hsivonen>
hober: I guess that the reason why the super-cool ski instructor appears so frequently is that it's the only advice animal that actually gives advice, and advice is needed when making standards.
06:14
<roc>
I believe the Opus IETF group wasn't terrible
06:20
<tantek>
that's good to hear
06:24
<annevk>
Hixie: cool
06:24
<annevk>
hober: my meme suggestion was a bad one
06:45
<annevk>
http://www.karlgroves.com/2012/07/26/selling-accessibility-negative-factors/ this is pretty how much how accessibility topics are handled in a F2F
06:48
<tantek>
annevk - wow
06:48
<tantek>
"The fact that accessibility is so hard and time consuming is partly because accessibility folks spend a lot of time chasing unicorns – seeking to pile on new requirements, check-points, and validation plans in hopes to make everything perfect. They separate instead of integrate and only show up when it is time to tell people they’ve messed everything up."
06:48
<annevk>
also the bit about hostility and FUD
06:49
<annevk>
every time I meet John Foliot he mentions lawsuits, rules of large corporations, new US legislation
06:50
<annevk>
not very compelling
07:16
<matjas>
zewt: re: astral symbols in <input type=password>, is it a bug to show them as two symbols or not? I’m not sure, with DOM strings being UTF-16 and all
07:30
<annevk>
matjas: depends on the UI requirements
07:30
<annevk>
matjas: it seems like suboptimal UI, but not necessarily a bug
07:31
<matjas>
annevk: is this stuff in a spec somewhere?
07:31
<annevk>
matjas: we don't spec UI
07:31
<hsivonen>
it seems like a bug but not one that's important to fix
07:31
<annevk>
matjas: you could take a look if the Rendering section says something, but I doubt it
07:32
<annevk>
"When the input-password binding applies, to an input element whose type attribute is in the Password state, the element is expected to render as an 'inline-block' box rendered as a text field whose contents are obscured."
07:32
<matjas>
that’s all I could find too
07:33
<kennyluck>
Hixie, you mentioned you were asking people about requirements on input mode control. Do you have a mail that summarizes your questions or something?
07:33
<hsivonen>
rendering two bullets for one character sure counts as obscured!
07:33
<annevk>
what I want fixed more than this is a "reveal password" thingie
07:34
<annevk>
kennyluck: we're look for use cases here: http://wiki.whatwg.org/wiki/Text_input_keyboard_mode_control
07:35
<kennyluck>
annevk, yeah, I know that page, though Hixie indicated that he sent some mails out, not sure if those were internal.
07:35
<kennyluck>
or I might have misinterpreted what he said.
07:51
<hsivonen>
MikeSmith: I think I should gather some stats about how validator.nu is used. Stuff like HTML5 versus HTML 4 usage numbers.
07:51
<MikeSmith>
yeah would be useful
07:51
<hsivonen>
MikeSmith: Any ideas what in particular I should measure?
07:51
<MikeSmith>
no particular ideas at the moment
07:51
<MikeSmith>
will try to think of some
07:51
<hsivonen>
MikeSmith: I'm not planning any fancy back end: just an object in RAM that can dump its data to a HTTP, so no disk I/O is needed.
07:52
<MikeSmith>
OK
07:53
<hsivonen>
My connectivity to Mozilla's important infrastructure is broken today, so I better take advantage of the validator hacking opportunity.
07:53
<MikeSmith>
hah
07:53
<MikeSmith>
yeah please do
07:55
<payman>
kennyluck: https://plus.google.com/107429617152575897589/posts/WwVmKNjcZ4p
07:55
<kennyluck>
payman, thanks!
07:57
<hsivonen>
MikeSmith: I think I should try to find out if people actually use custom Schematron
08:01
<annevk>
hsivonen: just measure all the settings?
08:03
<hsivonen>
annevk: it would be inappropriate to give a public dump of that
08:03
<hsivonen>
and I don't want to develop access-control
08:03
<annevk>
which would be wrong?
08:03
<hsivonen>
annevk: dumping the URLs of custom schemas
08:05
<annevk>
maybe you should hash them so you can at least count frequency?
08:05
<annevk>
but measuring first whether they're popular at all seems fine
08:06
tantek
just read that OAuth 2.0 and road to hell blog post and came upon Hixie's comment as well: http://hueniverse.com/2012/07/oauth-2-0-and-the-road-to-hell/#comment-42017
08:09
<annevk>
uhuh
08:09
<annevk>
WebSocket :(
08:09
<annevk>
although it didn't turn out all bad
08:12
<tantek>
I knew things about IETF were quirky and odd (e.g. the humming), but I didn't realize it was that broken (and susceptible to enterprise hijacking).
08:16
<ashemedai>
So with MathML in HTML5, am I supposed to use &Integral; or just use Unicode and put ∫ ?
08:24
<annevk>
up to you
08:35
<ashemedai>
annevk: Shouldn't cause any problems that you know of at least? I mean, semantically &Integral; might be passing along some additional info, not sure how UAs deal with this.
08:39
<hsivonen>
ashemedai: &Integral; doesn't pass any additional semantics
08:41
<ashemedai>
hsivonen: Brilliant, cheers.
08:44
<ashemedai>
Wonder if I should try compiling https://github.com/w3c/tidy-html5 under Windows one of these days.
08:50
<annevk>
ashemedai: the way it works is that &Integral; ends up as ∫ in the tree
08:51
<annevk>
ashemedai: so it's purely a source text affair and the source text is not exposed
08:56
<Stevef_>
hsivonen: your focus outline proposal is a crowd pleaser - thanks for filing and ccing me
08:56
<ashemedai>
annevk: gotcha
08:59
<ashemedai>
Microsoft's Math Input Panel loves to add a lot of extra <mrow>s in places it need not to.
09:06
<ashemedai>
Heh, even in a function specification like f(u,v) it made , a <mo>
09:08
<ashemedai>
Quite interesting how MathML can show the same rendering with quite varied MathML constructs.
09:21
<odinho>
I like that accessibility blog post.
09:23
<jgraham>
I would like it more if my experience wasn't that certain people it praises do all the things it criticises
09:24
<odinho>
jgraham: lol, you may replace them with "foo" and "bar" whilst reading.
09:28
<ashemedai>
Oh hey
09:28
<ashemedai>
The Wikipedia code has been updated. That MathJax change is now live on the main servers.
09:28
<ashemedai>
When logged in, under Appearance in Preferences
09:50
<Stevef_>
jgraham: I think you mean a certain person and I think it is important to draw a distinction between how people act in particular contexts (work) or when ineteracting with particular communities
09:57
<jgraham>
Stevef_: Yes, one person.
09:58
<Stevef_>
jgraham: John foliot can be a very positive force in web standards and bringing together the different communities open web camp for example http://openwebcamp.org/
09:59
<Stevef_>
jgraham: and like myself and many other people in the web standards community he can be disruptive at times
10:01
<hsivonen>
Stevef_: I filed the focus thing out of my personal need--not to please a crowd. I CCed you, because I mentioned it to you earlier.
10:02
<Stevef_>
hsivonen: fine, what i meant was it is a good idea and was welcomed
10:02
<hsivonen>
Stevef_: ok
10:03
<Stevef_>
hsivonen: my choice of words may not have been the best
10:03
<jgraham>
Stevef_: He must have done something worthwhile because a number of people seem to respect him. That doesn't change the fact that the attitudes and behaviours that he is quoted as advocatin in the article are not those that I have experienced him employing himself
10:07
<Stevef_>
jgraham: yes I understand that, which is why i mentioned the work distinction as I believe he does exhibit the qualities he is quoted as advocating in his professional accessibility work.
10:13
<hsivonen>
Stevef_: PARSER_XML_EXTERNAL ("Parser set to XML with external entities");
10:13
<hsivonen>
doh
10:13
<hsivonen>
Stevef_: is HTML working group work nonprofessional?
10:20
<Stevef_>
hsivonen: involvement is on a voluntary basis for many of us I think not affiliated with work. is what I meant. Also in my job (professional capacity) the clients I work with are on the whole not from the standards communities, many have no idea it exists, adobe and google being the exceptions
10:20
<Stevef_>
maybe microsoft, but who knows what they know
10:21
<hsivonen>
ok
10:25
<jgraham>
(I think it seems very strange to say that one behaviour is needed to get positive action for accessibility in a "professional" context and then display quite different behaviour when trying to achieve similar things in a standards context)
10:25
<Stevef_>
hsivonen: i generally work with web app developers who build interfaces with tables and buttons using links and have little clue as to how to provide a label for a form control
10:25
<jgraham>
But what I actually wanted to ask is whether there is a good reason to fire beforescriptexecute after getting the source of the script rather than before
10:26
<jgraham>
In particular for inline scripts where beforescriptexecute changes the script text
10:27
<jgraham>
(Gecko gets thins "wrong")
10:27
<jgraham>
hsivonen: ^
10:28
<Stevef_>
jgraham: its a 2 way street the different behaviour is often exhibited due to different responses/reactions recieved
10:30
<Stevef_>
also note that it is not binary standards/professional , its interaction with particular people or particular subgroups that makes a difference for many people
10:33
<jgraham>
Stevef_: I still don't understand why you would believe that acting in a certian way will be best at effecting change in one context and acting in a very different way will be the best way to effect the same kinds of change in a very similar context
10:34
<hsivonen>
jgraham: sicking's design. He would be a better person to answer.
10:35
<hsivonen>
jgraham: If Gecko and the specs disagree, then Hixie disagreed with Gecko.
10:36
<Stevef_>
jgraham: I don't, i am not talking about what is the best strategy for reaching one's goals I am talking about how other peoples behaviour can effects how one behaves, not in a good way
10:37
<jgraham>
hsivonen: It seems that Gecko complies with the spec in the src case
10:37
<jgraham>
Which makes sense, sort of
11:20
<Stevef_>
on another note, anyone interested in an MSAA/IA2/UIA accessibility object inspection tool for windows, developed specifically for testing web pages, can try out http://www.paciellogroup.com/blog/2012/07/latest-aviewer-alpha-2012/ feedback welcome
11:35
<smaug____>
Stevef_: make sure to file browser bugs about the issues
11:36
<Stevef_>
smaug____: yeah am in contact with various browser acc engineers on the bugs
11:38
<Stevef_>
smaug____: main issue is that chrome does not currently expose API interface to enable the tool to get at innerhtml
11:38
smaug____
knows nothing about chrome
11:39
<smaug____>
but you mentioned few Firefox issues
11:39
<smaug____>
and would be good to have bugs filed
11:39
<Stevef_>
smaug____: i think the issues are on our/NVDAs side
11:39
<smaug____>
Stevef_: just curious, how does Opera behave ?
11:39
<smaug____>
Stevef_: ah
11:39
<odinho>
Prolly misbehaves :P
11:40
<Stevef_>
smaug____:last i checked, does not expose anything useful, unfortunately
11:41
<smaug____>
odinho was right then :)
11:43
<Stevef_>
bits of the browser chrome, none of the HTML
11:51
<odinho>
Stevef_: So, if Opera should be better, -- is there a progressive way to get there? I mean, like small feature A making it lots better, and then many more small fixes to make it progressively better. -- Or is it all requiring a big up-front investment?
11:51
<odinho>
I have no idea over Opera's accesibility, or even what tools expect to extract.
11:52
<Stevef_>
odinho: i wrote on the subject here: http://www.paciellogroup.com/blog/2012/06/open-letter-to-opera-re-accessibility-support/
11:53
<Stevef_>
odinho: to be honest, Opera should look at implementing accessibility in its mobile browsers first (where they have marketshare)
11:55
<Stevef_>
odinho: there was an aborted attempt 5 or so years ago: http://my.opera.com/desktopteam/blog/2007/07/27/screen-reader-support
11:55
<odinho>
Stevef_: Yes, I did see that post actually. But it seems like real tasks to do stuff like that. I am doing spec stuff, tests etc, but I like doing small bug fixes on the side, so I'm always looking for really easy stuff to fix :P
11:55
<Stevef_>
adinhO; and belive there was some work on mac accessibility
11:57
<Stevef_>
odinho: right, so if you want something small scale (relaitively) that would improve access for keyboard users see discussion with hallvord: http://www.paciellogroup.com/blog/2012/07/w3c-html-and-whatwg-html-split-affect-on-html-accessibility/comment-page-1/#comment-19628
12:07
<odinho>
Stevef_: Yeah, that one seems a bit chewable. At least after a while :-) Found the bug about it, seems like it's a very old bug (created 2001).
12:07
<odinho>
Stevef_: chaals made an extension for it as a stop-gap, https://addons.opera.com/en/extensions/details/key-title/ -- has been downloaded 905 times.
12:08
<Stevef_>
odinho: yeah I saw chaals extension, its been an issue since forever in all browsers, i have been banging on about it for at least the mid 2000's...
12:09
Stevef_
hope to see issues fixed before retirement...
12:10
<odinho>
Stevef_: Hehe, when is that?
12:10
<Stevef_>
odinho: 20 years or so :-)
12:12
<odinho>
Stevef_: Hehe, if we haven't fixed at least the tooltip thing in that time, I will sit down and do it.
12:12
<Stevef_>
odinho: lol will hold you to that
12:13
<odinho>
^_^
12:19
<Stevef_>
odinho: also Opera could look at improving support for screen magnifiers on the desktop and also look at providing further features for low vision users who don't use AT, there are many of them, opera could corner the market
12:50
<sedovsek>
I wrote about CSS Flexbox Layout model, if anyone is interested > http://galjot.si/css-flexbox
12:58
<jgraham>
sedovsek: *dynamically
12:59
<sedovsek>
jgraham: Thanks. I fixed it.
13:13
<david_carlisle>
ashemedai: The comma should be in an <mo> (if it is explicit at all) or you could use mfenced and make the comma implict. All distinct character tokens in MathML are in some element or other
13:13
<jgraham>
In other news MathML is rather verbose :p
13:18
<david_carlisle>
jgraham: <word><letter>w</letter><letter>e</letter></word><space> </space><word><letter>l</letter><letter>i</letter><letter>k</letter><letter>e</letter></word><space> </space><word><letter>o</letter><letter>u</letter><letter>r</letter></word><space> </space><word><letter>t</letter><letter>a</letter><letter>g</letter><letter>s</letter></word>
13:19
<jgraham>
:)
13:46
<payman>
sedovsek: "back in in"
13:46
<sedovsek>
:P
13:53
<payman>
sedovsek: s/Another two/Other two
13:55
<sedovsek>
payman: Thanks! Fixed it.
13:55
<payman>
didn't fix the first issue though :)
13:56
<sedovsek>
Oh, yea.
13:56
<sedovsek>
Thanks again.
13:56
<payman>
np
14:18
<zewt>
matjas/annevk: definitely a bug--it doesn't matter that the character is represented as a surrogate pair; it's still one character/one codepoint
14:19
<matjas>
implementation details exposed to the user etc etc
14:19
<matjas>
either way, not really a very important issue indeed
14:19
<zewt>
and being able to backspace over half of the surrogate pair (which you can in chrome and FF) is definitely a bug--you should never be able to create a broken surrogate pair (at least not with a simple UI like that--not sure what should happen in more obscure cases, like pasting one in)
14:20
<matjas>
so who’s gonna file the bugs? /me ducks
14:20
<zewt>
the fact that non-BMP characters are so rarely used is the only reason the utf-16 fustercluck is even tolerable :)
14:20
<matjas>
haha
14:21
<matjas>
i bet no one actually uses astral symbols in passwords
14:21
<zewt>
personally I wish we'd just change JS to ucs-4, under the logic that although we'd break some sites ... it'd probably fix far more :)
14:21
<matjas>
heh
14:21
<zewt>
(unfortunately, the sites that it would break would be the few that it actually matters for, since after all they spent the time making it work)
14:25
<smus>
hey, would like to polyfill a CSS function, but it appears that there's no consistent CSS OM API so the only way this can be done is by parsing CSS via JS (yuck!). can someone point to a relevant spec?
14:26
<beverloo>
a spec for what?
14:26
<paul_irish>
(because CSSOM wont reveal anything it cant succesfully parse..)
14:26
<beverloo>
the CSS OM is specified here: http://dev.w3.org/csswg/cssom/
14:26
<paul_irish>
well.. ideally CSSOM had a UnknownRule definition
14:27
<smus>
right. thx paul
14:27
<paul_irish>
so when it failed we could at least hack up that part with regex
14:27
<smus>
missing piece!
14:27
<paul_irish>
aside from that, the browser could expose essentially the stylesheet AST
14:27
<SimonSapin>
parsing CSS from JS? https://github.com/tabatkins/css-parser
14:28
<smus>
so if ur like background: -webkit-image("foo.webp", "foo.jpeg"); the parser just ignores that rule entirely. would be nice if it parsed as unknown
14:28
<beverloo>
the css syntax specifically defines that unknown properties/at rules need to be ignored
14:28
<paul_irish>
SimonSapin: the goal is a polyfill you'd use in production... and you can't ask folks to parse css in js clientside for real.
14:29
<paul_irish>
beverloo: sure, but you can see the problem. we can't polyfill to extend css when unknowns are discarded.
14:29
<beverloo>
the css syntax (including the rules for ignoring it) is specced here: http://www.w3.org/TR/css3-syntax/
14:29
<beverloo>
well, you'd have to polyfill the hypothetical CSSUnknownRule too :p
14:30
<paul_irish>
hey now
14:30
<paul_irish>
i dont need to polyfill try/catch
14:30
<smus>
http://www.highdots.com/forums/javascript/retrieving-unknown-css-rules-48513.html
14:30
<smus>
specifically, http://lists.w3.org/Archives/Public/www-style/2003Oct/0347.html
14:31
<smus>
why drop unknowns?
14:31
<paul_irish>
omg. bert.
14:31
<beverloo>
it sounds like you should take this to www-style
14:32
<smus>
k, any known reason to drop unknowns?
14:32
<smus>
(so i can present a balanced case)
14:33
<paul_irish>
smus: its probably somewhere in this 2003 era archives of www style
14:33
<paul_irish>
beverloo: yeah certainly. it'd be fun to get UnknownRule back
15:30
<smus>
cool, so <img srcset> doesn't seem to work in any browser. is there an existing reference implementation? http://smus.com/m/srcset/
15:31
<jreading>
iOS 6 mobile safari
15:33
<dglazkov>
good morning, Whatwg!
15:34
<SimonSapin>
good afternoon, dglazkov
15:43
<Wilto>
jreading: Any chance you have a link to the commit/some details handy? Curious as to whether it just covers resolution or the whole extended microsyntax.
16:14
<Velmont>
smus: ... You have an error in your syntax as well, it's with comma.
16:15
<smus>
Velmont: thx!
16:30
<jreading>
Wilto: http://trac.webkit.org/changeset/111637
16:31
<jreading>
and a polyfill: https://github.com/SunboX/-webkit-image-set-polyfill
16:31
<jreading>
:/
16:37
<miketaylr>
sweet README
16:37
<miketaylr>
accidentally the code
16:38
<jreading>
polyfills are the new domain squatting
16:45
<jreading>
i misspoke about srcset, it's "image-set", sorry for the confusion. *ducks*
16:48
<Hixie>
hsivonen: hey sometimes gecko and the spec disagree just because i made a mistake, or because mozilla people told me they disagree with gecko :-)
18:43
<Hixie>
hmmm
18:43
<Hixie>
i wonder how to spec the autocomplete="" field's grammar and parsing
18:44
<Hixie>
<empty> | on | off | [section] [subsection] ( generic-field | [contact-type] contact-field )
18:44
<Hixie>
i guess to parse it i split on spaces then go backwards through the tokens
18:47
<Hixie>
and to define the conformance criteria for authors, maybe an ordered list of tokens, and i just say what each token has to be or something
18:47
<Hixie>
ok
19:06
<Wilto>
jreading: Man, that’s gonna be a tough one to polyfill.
19:38
<tantek>
Hixie, sounds horribly complicated
19:39
<Hixie>
i'm certainly open to alternatives :-)
20:02
<zewt>
okay, the chrome thing where no form of refresh will make it reload css background: url() images is pretty much horrible
20:05
<TabAtkins>
I think Inspector is gaining some Refresh options soonish, so you can do hard refreshes or even full cache flushes.
20:06
<zewt>
that's what control-f5 is supposed to do, it just appears to not propagate to css resources
20:07
<TabAtkins>
That actually sounds like a bug. File on us?
20:26
<annevk>
paul_irish: TabAtkins and I at some point came up with data-* CSS properties that would be exposed in the CSSOM
20:26
<annevk>
paul_irish: as opposed to exposing every property which would kinda ruin forward compatibility
20:35
<TabAtkins>
annevk: That's the CSS Variables spec.
20:36
<annevk>
ah yeah, it morphed into that
20:36
<TabAtkins>
Which we'll finalize for WD and soonafter LC next month at the meeting.
20:36
<TabAtkins>
But yeah, exposing every property is a non-starter.
20:36
<TabAtkins>
Like custom elements without an x- prefix or something.
21:08
<jgraham>
TimBL!
22:41
<Hixie>
TabAtkins: thanks for answering that intern's question
22:41
<Hixie>
TabAtkins: i was like "wtf are you talking about"
22:44
<TabAtkins>
Hixie: Haha, np
22:44
<TabAtkins>
I want to see the feature implemented!
22:44
<TabAtkins>
And now I have some questions to post to the list about it.
22:50
<Hixie>
i like your "just implement the spec" thing
22:50
<Hixie>
i always feel weird saying that :-)
22:54
<TabAtkins>
Why?