03:59
<JonathanNeal>
hello
08:37
<tantek>
thanks MikeSmith for the #microformats edit - I've been unsure about the rel-publisher value since I saw it show up.
08:43
<MikeSmith>
tantek: somebody asked about it on the www-validator list (asked why the validator doesn't support it despite it being on the wiki page)
11:52
<annevk>
oh, seems some of the IETF found the URL Standard again
11:52
<annevk>
too bad their comments are not really actionable
12:04
<odinho>
What are they saying?
12:06
<annevk>
https://twitter.com/bagder/status/286544650284699648
12:09
<annevk>
I like this one too https://twitter.com/dret/status/286625563496574977
12:09
<annevk>
Especially combined with the fact that their hero Roy has been stating that URI == URL for some time now...
12:10
<annevk>
Anyway, back to creating more states for more sensible error reporting...
12:40
<annevk>
If you resolve http:example.org against http://example/ should that give a parse error?
12:43
<annevk>
I'm going with yes...
12:55
<odinho>
Hehe. One thing they do say though, is that it would indeed be nice to have a non-normative overview over what's supported. You're doing lots of findings, and then encode them into this spec, but it's not quickly referencable. Although I fear that would be too big to use in the end anyway...
12:56
<annevk>
Given that it all changes over time I don't really see how that's useful. Although you could "generate" such a thing if we completed that test suite...
12:56
<annevk>
Sorry, I can see how it's useful, but also how it'd be time consuming :-)
12:58
<odinho>
Yeah.
13:56
<SimonSapin>
annevk: maybe diagrams like http://dev.w3.org/csswg/css3-syntax/#token-diagrams would help explaining the syntax of URLs?
14:00
<annevk>
SimonSapin: yeah, the plan is to add those at some point
14:00
<annevk>
SimonSapin: though the comments were about parsing
14:38
<annevk>
SimonSapin: those syntax diagrams seem incomplete btw
14:38
<SimonSapin>
annevk: how so?
14:38
<annevk>
SimonSapin: e.g. for url() you can also escape the u / r / l
14:39
<SimonSapin>
indeed
14:44
<SimonSapin>
annevk: or maybe not. http://www.w3.org/TR/CSS21/syndata.html#tokenization uses "url" litteraly in the token definition, not something like {u}{r}{l}
14:44
<SimonSapin>
Green in Firefox and Opera but Chromium: data:text/html,<style>@import \75rl("data:text/css,body{background:green}")
14:50
<annevk>
SimonSapin: yeah, the old specs are messy that way
14:51
<Ms2ger>
Sounds like something needs a test
14:51
<annevk>
SimonSapin: there was a problem of sorts with writing it as {u}{r}{l} if I remember correctly
14:51
<SimonSapin>
annevk: http://dev.w3.org/csswg/css3-syntax/#data-state still does not allow escapes
14:52
<annevk>
SimonSapin: oh Tab forgot the hard part?
14:53
<annevk>
oh I see
14:53
<annevk>
hmm
14:57
<SimonSapin>
seemed deliberate to me last time I looked, but maybe
14:57
<annevk>
yeah, seems weird though to accept escapes everywhere but url()
14:57
<annevk>
well and @charset, but @charset is weird
14:58
<Ms2ger>
Ah, darobin is back?
14:59
<SimonSapin>
I don’t think this is really a problem in practice. Have you ever seen escapes used on CSS-defined keywords, other than in test cases?
14:59
<SimonSapin>
But yes, it should be clarified
15:00
<annevk>
well I'd prefer if we killed CSS escapes
15:00
<annevk>
especially for keywords/identifiers
15:00
<annevk>
but that seems unrealistic
15:00
<darobin>
Ms2ger: ayup
15:01
<Ms2ger>
darobin, go review some tests :)
15:01
<darobin>
thanks for the pull requests wrangling you did over the holidays
15:01
<darobin>
haha
15:01
<darobin>
will do, but I'm still catching up on email
15:01
<SimonSapin>
annevk: probably unrealistic. But I usually pretend they’re not there. Just like I pretend that everything is case-sensitive ;)
15:02
<Ms2ger>
Good :)
15:02
<Ms2ger>
I've got some more pull requests to create, but branches hurt my head, so it'll be one at a time
15:17
<annevk>
SimonSapin: if only we could go back in time and simplify some of this stuff
15:17
<SimonSapin>
annevk: @charset is case-sensitive too :)
15:17
<annevk>
SimonSapin: @charset is beautiful
15:18
<annevk>
SimonSapin: requires double quotes as well
15:18
<annevk>
SimonSapin: and a single space preceding the double quotes
15:18
<SimonSapin>
and a single space, no comment, no space before the ;
15:18
<annevk>
yeah, once you have to write out the details people suddenly realize they should keep it simple...
15:19
<SimonSapin>
well, if we go back in time let’s support UTF-8 only, @charset doesn’t even exist
15:19
<annevk>
fair point
15:19
<annevk>
and kill utf-16 before hsivonen et al get a chance to be upset by it
15:19
<SimonSapin>
hehe
15:23
<annevk>
how many days left in this TAG election thingie darobin?
15:23
<darobin>
6
15:23
<SimonSapin>
wrote to www-style about escaping url()
15:23
<annevk>
so it's closed after Jan 9?
15:24
<annevk>
SimonSapin: might want to search the archives
15:24
<annevk>
SimonSapin: Bert and I went over it before...
15:25
<annevk>
SimonSapin: http://lists.w3.org/Archives/Public/www-style/2007Dec/0215.html
15:26
<annevk>
SimonSapin: http://wiki.csswg.org/spec/css2.1#issue-23
15:27
<annevk>
SimonSapin: it was closed, but no idea how it was fixed (assumed editorial suggests they agreed with {u}{r}{l}...)
15:27
<darobin>
annevk: yes, correct
15:32
<SimonSapin>
annevk: http://test.csswg.org/suites/css2.1/20110323/html4/uri-015.htm
16:15
<annevk>
SimonSapin: the assert there is wrong
16:15
<annevk>
metadata fail #nobodyiscounting
16:19
<annevk>
SimonSapin: see also e.g. http://www.w3.org/TR/2010/WD-CSS2-20101207/grammar.html#scanner
16:20
<annevk>
SimonSapin: stuff changed there, I forgot if there was an email explaining why
16:20
<SimonSapin>
annevk: lots referenced from http://lists.w3.org/Archives/Public/www-style/2012May/0329.html
16:20
<SimonSapin>
I’m writing a patch for css3-syntax
16:22
<annevk>
SimonSapin: except the edit was made and then reverted, and I believe Bert had a reason for that
16:22
<annevk>
SimonSapin: see the above draft for where the change was actually in
16:31
<annevk>
Ah, the problem is http://lists.w3.org/Archives/Public/www-style/2010Jul/0499.html I suppose
18:38
<SimonSapin>
annevk: well, in level 3 tree construction actually uses the tokenization
18:38
<SimonSapin>
having a full CSS 2.1 separated from the core grammar was a mistake IMO
18:46
<annevk>
uhuh
19:26
<Ms2ger>
w3fools is down?
19:55
<Hixie>
annevk: https://twitter.com/dret/status/286625563496574977 is funny because that's what people from the w3c said back when we started with Web Forms 2 :-)
19:55
<yroc>
Hi, happy new year all -- wondering about rationale for <pre>
19:56
<Hixie>
anything specific about it? or just why it's in the spec?
19:56
<yroc>
why not just use white-space: pre;
19:56
<yroc>
and display: block;
19:56
<tantek>
because the white space has meaning in the content sometimes
19:56
<Hixie>
the idea is that CSS is optional
19:56
<yroc>
on any particular element that you want preformatting?
19:57
<Hixie>
you should be able to get everything important even if the CSS doesn't get applied
19:57
<tantek>
e.g. code, python, etc.
19:57
<tantek>
if the whitespace has meaning, use <pre>, if the white-space is presentational, use white-space property
19:57
<Hixie>
tantek: yroc is writing our rationale document, so he needs the "why" more than the "what" :-)
19:58
<tantek>
Hixie, s/important/has meaning in the content
19:58
<tantek>
"important" is too fuzzy / not actionable
19:59
<Hixie>
yeah not saying what i said was good, just that "do X" isn't helpful if the question is "why do we have X" :-)
20:00
<tantek>
my answer was for why
20:00
<yroc>
But really, where do we stop adding elements then. What makes preformatted text special enough to be given (and retained) a semantic tag?
20:00
<tantek>
we have <pre> because there are cases where whitespace has meaning in the content.
20:00
<tantek>
yroc - elements have to meet a certain high bar for inclusion. I believe W3Cmemes has an explanatory image for this.
20:00
<yroc>
Or is it just historical?
20:00
<Hixie>
yroc: realistically, we have <pre> because it was implemented already when i wrote hte spec
20:01
<tantek>
(new) elements
20:01
<yroc>
Hixie: right then, it's basically historical...
20:01
<Hixie>
yroc: and for things that were implemented already, the choice is either (a) drop it in the "obsolete" section or (b) go all-in and spec it properly
20:01
<tantek>
Hixie, more stuff should have gone into the obsolete section
20:01
<yroc>
@So what made you decide the latter for pre?
20:02
<Hixie>
yroc: for <pre>, (b) seemed more useful since there are times, e.g. source code, e-mails, etc, where it's impossible without CSS
20:02
<yroc>
@Hixie, sorry
20:02
<Hixie>
tantek: file bugs
20:02
<Hixie>
s/etc/etc)/
20:02
<tantek>
Am now of the opinion that the back-semantic definitions of <i>, <b>, <s>, <u> are nearly useless and confusing.
20:02
<Hixie>
wait forget my last regexp
20:02
<tantek>
Hixie, I'd expect you to close the bugs as "already discussed"
20:02
<Hixie>
tantek: not if there's new data :-)
20:03
<Hixie>
tantek: but if you want to get rid of <i>, make sure to say what element you'd use instead for its six zillion use cases
20:03
<tantek>
well, I don't really consider "changing my opinion" to be sufficiently new/objective "data" for that purpose. however, it's sufficiently anecdotally interesting IMO that I'm going to teach that way.
20:03
<Hixie>
tantek: same for the others (but not six zillion, more like 5, 3, and 2 respectively)
20:03
<tantek>
the point is not that there are no use-cases. the point is that the use-cases are not relevant enough to justify inclusion in the core language.
20:04
<Hixie>
yroc: see also http://wiki.whatwg.org/wiki/FAQ#Where.27s_the_harm_in_adding.E2.80.94 for rationale behind not adding features btw
20:04
<Hixie>
tantek: <i> has a ton of use cases that are pretty critical, imho.
20:04
<yroc>
@Hixie: will do
20:05
<tantek>
Hixie, that's a good link.
20:05
<Hixie>
tantek: the others i agree are weaker, and i wouldn't have added them if we didn't have a way to get them for free
20:05
<tantek>
Thanks, will be using that.
20:05
<tantek>
They're not free in terms of teaching cost.
20:05
<tantek>
<i> does stand out more than the others, I'll admit.
20:05
<tantek>
I have found more often uses for it in prose, posts etc.
20:05
<Hixie>
the teaching cost i think is gonna have to be paid for anyway
20:06
<tantek>
(per the "instance of a term" rough definition)
20:06
<tantek>
Hixie, nah, we can simply ignore old/obsolete elements when authoring.
20:06
<yroc>
I've interpreted <i>, for example, to be a "catch all" for semantic phrases that are conventionally marked up with italics. and
20:06
tantek
jokingly thinks of writing a summary of HTML: The Good Parts ;)
20:06
<Hixie>
there's so much material and software out there saying to use <b> and <u> and to a lesser extent <s> that you;ll have to field questions either way; we might as well, imho, head it off at the pass.
20:07
<Hixie>
by defining them
20:07
<tantek>
nah, easier to say, "old docs, ignore"
20:07
<Hixie>
since we have uses for them
20:07
<Ms2ger>
tantek, good luck finding those
20:07
<tantek>
just as we say for that-website-that-shall-not-be-named-but-shows-up-in-google-results
20:07
<tantek>
good luck finding what?
20:07
<tantek>
people love simplified summaries
20:07
<tantek>
that dispense with trivia, esoterica, obsolete "old people" stuff
20:07
<Ms2ger>
The good parts of HTML
20:07
<yroc>
You use <em> for stress emphasis, but you can't have tag for every possible case where italics are conventionally used. Therefore you defined <i> as a "catch-all".
20:08
<yroc>
@Hixie: is that a correct interpretation of <i>?
20:08
<Hixie>
tantek: "old docs, ignore" is not much cheaper than "b is for keywords, s is for intentionally inaccurate text (think price cuts), and u is for annotations (like spelling errors)"
20:08
<Hixie>
tantek: see http://www.whatwg.org/specs/web-apps/current-work/#usage-summary
20:09
<Hixie>
yroc: yeah, it's pretty close
20:09
<tantek>
Hixie, how is "keyword" different from "instance of a term"?
20:09
<Hixie>
yroc: (though i wouldn't use the word "italics")
20:09
<yroc>
@Hixie: why? and what word would you use?
20:10
<Hixie>
tantek: see the spec, it has some pretty good examples. the short answer is when you read text that has <i>, you use a different voice, whereas you don't read <b> differently at all, instead it's meant to draw your attention.
20:10
<Hixie>
tantek: e.g. gossip magazines will <b> a person's name so you can see the articles talks about them, but <i> wouldn't be appropriate for that.
20:11
<Hixie>
yroc: "instance of term" or "alternate mood or voice", like the spec :-)
20:12
<yroc>
@Hixie: right, but why are those particular (seemingly disparate) semantic elements lumped together in the first place?
20:12
<yroc>
Is it not because they are conventionally *italicized*?
20:12
<tantek>
Hixie, how do you "draw your attention" to <b> in a screen reader if "you don't read <b> differently at all" ?
20:12
<yroc>
in prose.
20:13
<yroc>
What does "instance of term", "alternate mood", and "ship name" have in common?
20:13
<yroc>
If not that they are conventionally italicized in prose?
20:15
<Hixie>
yroc: see the spec. The core definition is "alternate mood or voice", the others are all just examples of that. For example, when you say "I sailed on the /Enterprise/ for a couple of days", you tend to have a slight pause before the ship name and pronounce it with a slightly different stress. Same as "I like /felis domesticus/".
20:15
<Hixie>
tantek: a list or other UI, same as e.g. headings, tables, etc.
20:16
<yroc>
@Hixie not so sure about that one :-)
20:17
<yroc>
But you're the editor. I'm here to write what you're rationale is, so that's what I'll do.
20:19
<yroc>
Just circling back to <pre> for a moment, you said its continued inclusion is partially do to CSS being required.
20:19
<yroc>
Are you saying it's a goal of specing HTML to minimize authors having to use CSS?
20:20
<Hixie>
the core design of HTML is that it describes the page's structure/meaning/semantics/logic/content/other-buzz-words, while CSS describes the page's layout/presentational/formatting/style/other-buzz-words. The latter part is medium-specific and should never be necessary to understand the page.
20:21
<Hixie>
it should always be possible to just drop CSS entirely and still interact with / use / understand the page
20:21
<Hixie>
without CSS it'll likely be much uglier, but that's a separate issue
20:22
<yroc>
That's a good explanation -- makes me understand pre better, thanks.
20:23
<yroc>
In the spec, regarding links, it says: media, hreflang, and type attributes are "purely advisory", what exactly does that mean?
20:23
<yroc>
"Purely advisory", that is...
20:23
<Hixie>
"media" should be gone
20:23
<Hixie>
unless you mean <link>
20:24
<yroc>
It's a general link section... I think it includes <a> and <link>
20:24
<Hixie>
"purely advisory" means "just information, not something anyone is expecting you to take as fact or do anything interesting with, if you get more authoritative information, ignore the purely informative one"
20:24
<Hixie>
try reloading, i removed "media" a few days ago
20:25
<yroc>
K.
20:28
<yroc>
Another <pre> question (sorry!) Why are <br> elements allowed in pre? Doesn't that contradict the semantics?
20:31
<Hixie>
<br> in <pre> means nothing different than a newline
20:33
<yroc>
Yes, but isn't the whole point of pre, that the text is formatted "by typographic conventions rather than by elements"? (where I interpret <br> to be an element).
20:33
<Ms2ger>
Not really
20:33
<Ms2ger>
You can have <strong>s too
20:33
<Ms2ger>
Or whatever you like
20:34
<Hixie>
yroc: that's the intent, but it's harmless to allow <br> in practice
20:34
<Hixie>
yroc: in general we try to make things be invalid if that would help catch likely errors
20:34
<yroc>
Aha, got ya.
20:34
<Hixie>
yroc: but if something is just a bit silly, not really an error, and solving it wouldn't save anyone any time, then we allow it, typically
20:35
<yroc>
Or perhaps could you call it "non-conforming"
20:35
<Hixie>
yroc: see also http://www.whatwg.org/specs/web-apps/current-work/#conformance-requirements-for-authors
20:35
<Hixie>
and its subsections
20:35
<yroc>
I know I'm being very (too) analytical, but that's me!
20:36
<yroc>
Yes, I'll check it out.
22:52
<Hixie>
not sure how i feel about @whatwg tweeting out w3c bureaucracy :-P
22:56
<tantek>
unless WHATWG is planning a takeover from within of W3C bureaucracy...
22:56
<Hixie>
"takeover"?
22:56
<Hixie>
of what?
22:57
<Hixie>
we already write all the specs we want to write, what more is there to take over
23:09
<gavin>
the w3 brand
23:09
<zewt>
who would want that
23:10
<Hixie>
i'm certainly ok with them keeping that...
23:10
<zewt>
that'd be like buying out RealPlayer for the brand recognition
23:10
<Hixie>
hah
23:11
<gavin>
it's not valuable for their reputation amongst people who actually know what they do
23:11
<gavin>
it's valuable for all the other people in the world, of which we are but a small minority
23:11
<zewt>
nobody else cares
23:12
<gavin>
I wish that that that were true, but it's a horribly naive claim
23:13
<Hixie>
it's not clear to me what difference it makes. authors use what's implemented. the implementors are those "who actually know what they do".
23:13
<gavin>
plenty of people still think the w3 is highly relevant to web standardization
23:13
<gavin>
that reputation has value
23:13
<Hixie>
but what does their thinking that actually do, concretely?
23:14
<Hixie>
i mean, if it's just a matter of who gets credit, i don't care
23:14
<Hixie>
i'm happy for the mainstream media to pester the w3c PR team and let us get on with the work :-)
23:14
<zewt>
can't say i've ever known anyone who isn't actually implementing (or complaining about something not implementing) who cared at all about the w3c, or the ietf for that matter
23:16
<gavin>
it's not just about credit. reputation is an important part of being able to influence people and things.
23:17
<zewt>
and reputation is something the w3c has been losing rapidly of late
23:20
<Hixie>
gavin: what is it the w3c can influence that is valuable?
23:20
<Hixie>
we already write all the specs we want (or rather, hand the bandwidth) to write
23:21
<Hixie>
s/hand/have/
23:21
<Hixie>
the browser vendors have influence on the spec
23:21
<Hixie>
the w3c has never had that influence
23:21
<gavin>
they can influence people's thoughts about what web standards are important
23:22
<gavin>
you might say that that's "not important"
23:22
<zewt>
i'd say that's a bit too vague to even have an opinion on, can you be more concrete
23:22
<Hixie>
do you have any examples of them doing that?
23:22
<Hixie>
i'm not sure what you mean
23:23
<Hixie>
like how they influenced people to think XHTML2, XForms, and RDF were important?
23:24
<zewt>
if browser vendors wants to implement something, authors want to use it and whatwg has the bandwidth to spec it, no amount of handwaving or ranting from the w3c would stop it from happening (if they even chose to try, which I'm not aware of them doing)
23:25
<Hixie>
well they tried to stop HTML
23:25
<gavin>
many people think that HTML5 CR is an important milestone, for example
23:25
<Hixie>
(and to a lesser extent canvas and microdata)
23:25
<gavin>
if the w3 tomorrow released a statement saying that HTML5 development hit a wall and that they'd need to delay its readiness as an official spec, many people would believe them
23:26
<Hixie>
and so?
23:26
<zewt>
meanwhile stuff would keep getting implemented and they'd look silly
23:26
<Hixie>
and people would still use it
23:26
<gavin>
no, people would not use it as much
23:26
<Hixie>
just like they have been using it for years while the w3c kept saying it wasn't done
23:26
<zewt>
that's quite a claim
23:26
<gavin>
I'm not talking about the existing uses. I'm talking about new uses.
23:27
<Hixie>
people have used new stuff that i've specced and people have implemented so quickly that we've actually been forced into compataibility constraints before i was done speccing them
23:27
<Hixie>
pushState() being the best example
23:27
<gavin>
getting people to buy that HTML is a suitable platform for mobile development is just as much about PR as it is about what gets implemented
23:27
<zewt>
not at all, it's only about getting apple and google to fix their crappy-to-useless mobile browsers
23:28
<Hixie>
getting people to buy that HTML is a suitable platform for mobile development is exclusively about whether HTML actually ir, or is not, a suitable platform.
23:28
<Hixie>
w3c has sod-all to do with that.
23:28
<Hixie>
(whatwg has very little to do with it either)
23:28
<gavin>
no, it's not at all exclusively about that
23:28
<gavin>
it would be nice if it were
23:28
<Hixie>
people are _trying_ to use the web, and can't because of browser limitations, and that has nothing to do with the w3c
23:28
<Hixie>
which would have them using XForms and RDF
23:28
<gavin>
but we don't live in a world where the only thing that matters is bits-shipped, nice as that would be
23:29
<gavin>
anyhow, I'm clearly not going to get anywhere arguing with you guys about this, so I'll drop it
23:29
<Hixie>
i think facebook's comments (which were based entirely on their experience, as far as i can tell, and not at all on the w3c) had more impact on people's view of the web as a platform for mobile dev than anything the w3c ever did or said
23:29
<zewt>
(and I suspect both apple and google half-ass their mobile browsers intentionally because they want people to write nonportable apps instead)
23:30
<gavin>
I gave a very theoretical example of something that the w3 could impact using its influence - they haven't done that, so it's kind of moot
23:30
<Hixie>
have they ever impacted anything with their influence?
23:30
<gavin>
I agree that they're losing their influence, obviously, but it's naive to claim that they don't have any
23:31
<gavin>
yes, plenty
23:31
<Hixie>
i'm happy to believe they have some if there's any evidence that they've had some, but so far the only impact i've seen is pretty much irrelevant. e.g. they had an impact on what doctypes people use, but that's just an opaque stirng.
23:31
<gavin>
companies have "must support w3 standards" on all of their checklists
23:31
<Hixie>
those companies write web pages that don't validate using flash and html5 features the w3c hasn't included in their specs.
23:32
<Hixie>
so i'm not sure that means anything.