00:00:17 kazuhito has joined #html-wg
00:23:44 mjs has joined #html-wg
00:25:15 so much scrollback to read
01:07:33 i'm really starting to think that john boyer isn't discussing the form issues in good faith
01:09:03 why do you say that?
01:09:39 because he keeps asking what's wrong with xforms, but always ignores the replies, and never response to questions about what's wrong with web forms 2
01:09:52 furthermore, he repeatedly has stated incorrect facts about web forms 2
01:10:04 despite having been corrected multiple times
01:10:18 marcos has joined #html-wg
01:10:49 he may not have understood
01:11:01 then he should say so
01:11:25 you don't always know when you've misunderstood something :)
01:11:51 but he may be looking at it with XForms-colored glasses
01:12:53 or, as you say, he may be operating from vested interests (as are the browser vendors)
01:13:56 i have no objection to him operating with vested interests. But there's no point having a discussion if one of the people in the discussion isn't actually interested in using logical, fact-based argumentation.
01:14:08 fair enough
01:25:51 gavin_ has joined #html-wg
02:22:31 mjs has joined #html-wg
02:30:03 mjs: when you do the line count stuff, you have to add up html4 + dom2 html + xhtml1
02:31:40 vs wf2 + wa1
02:32:06 html4 refers to sgml for parsing, so...
02:32:13 as it's important not to forget xhtml1 :-)
02:32:18 hm
02:32:26 i guess you could indeed argue that you should also include sgml
02:37:57 hmm, html4 has many normative references
02:38:25 esp [HTML40] surprised me for being normative
02:49:06 html4?
02:52:07 dbaron has joined #html-wg
02:56:59 yes. http://www.w3.org/TR/html401/references.html
03:17:46 billmason has joined #html-wg
03:20:28 weicd
03:20:29 weird even
03:27:04 Voluminous has joined #html-wg
03:32:23 gavin_ has joined #html-wg
03:40:55 myakura has joined #html-wg
04:15:47 Shunsuke has joined #html-wg
04:36:23 Shunsuke has joined #html-wg
04:48:35 myakura has joined #html-wg
04:55:58 Shunsuke has joined #html-wg
04:57:05 hyatt has joined #html-wg
05:19:01 Shunsuke has joined #html-wg
05:34:52 Zeros has joined #html-wg
05:39:03 gavin_ has joined #html-wg
05:57:52 Hixie: You can and should include SGML.
05:58:24 in the count of lines that html5 is replacing?
05:58:32 yes
05:58:42 that's what, 500 pages?
05:59:09 And this is the biggest fundamental problem with html4 IMO; the lack of public (web) availability of the ISO 8879 spec.
05:59:18 Summjat like that, yeah.
06:00:23 so html5 is like a third of the size of the specs its replacing
06:00:24 wow
06:00:51 Quite likely, yes.
06:01:14 But note that that's not /necessarily/ a compliment.
06:01:36 given how much more detailed html5 is, i'll take it as one :-)
06:01:59 SGML is quite a bit more complex than html5
06:02:39 If you bundled all the previous CSS and DOM specs and the HTML5 spec that'd be pretty dense too
06:03:20 hmm. call it "Web Platform Specification" and charge $99 per copy and you have ISO
06:05:16 that size comparison is not quite fair
06:05:59 the mode for those specs was to reference other specs, to build on technologies
06:06:27 you are simply taking all the relevant parts and putting them in one place
06:06:37 yeah, sgml is equivalent to xml
06:06:41 not xml+dom+css+...
06:06:42 not that that's bad
06:06:55 I sincerely hope HTML5 tries to stand on various proverbial shoulders here!
06:06:56 Hixie, sgml defines quite a lot more than xml
06:07:10 not really
06:07:17 I was talking size though
06:07:21 I'm just saying it's not a true comparison
06:08:58 schepers: No such comparison could be claimed to be "accurate". But I think, put in context, it can be an _apt_ comparison.
06:09:09 Hixie, if that were true then there'd be no need for xml, we'd just use sgml
06:09:18 apt to what?
06:09:39 I mean, what are you claiming it says?
06:09:41 Zeros: sgml defines a syntax. xml defines a syntax. it just so happens the sgml syntax is way more expressive.
06:09:52 Zeros: they're still equivalent technologies
06:10:18 Hixie, Um, equivalent in the way that csv is the same as xml I guess.
06:10:24 one is just more expressive
06:10:32 right
06:11:16 That wasn't the parallel I was trying to make though
06:11:17 schepers: It illustrates, for instance, the point that HTML5 — whatever else one may think about it — does go to great lengths to specify parsing rules that HTML4 handwaves somewhat to SGML.
06:11:49 Which, btw, is an argument both pro and con the HTML5 approach in my book.
06:12:04 huh? how does the length of a spec have any bearing on that, xover?
06:12:45 Strictly speaking it doesn't; it "illustrates" it, not "proves" it.
06:14:26 Of course, the mere weight of lines is one of the reasons I'm still on the fence.
06:14:56 I didn't say anything about "proof"... anyway, nm, it just struck me as a silly exercise... carry on
06:15:08 Then we're in agreement. :-)
06:15:20 I doubt that :)
06:15:55 Or perhaps "instructuve" would be more appropriate.
06:19:51
tH has joined #html-wg
06:20:14 hmm, html5 is 211 pages
06:20:20 not including web forms
06:20:36 and its not done yet. Doesn't seem that much smaller
06:20:43 it's not far from done
06:20:58 other than the rendering section, i'd be surprised if we added more than 50 pages
06:21:12 the rendering section might be 50 to 100 pages of its own
06:21:21 we have to add wf2, though
06:21:43 but don't forget that it replaces sgml, html4, dom2 html, and xhtml1, all at once
06:21:54 which together probably add up to 2000 pages or more
06:25:42 part of that is related to whitespace
06:25:50 HTML5 is very very dense
06:27:12 I think the HTML4 spec is much easier to read too honestly
06:28:47 it would be easy to make the html5 spec easier to read if it failed to say anything like html4 does...
06:30:08 that's no reason for it to be so dense
06:30:09 http://www.w3.org/TR/html401/struct/links.html#h-12.2 => http://www.whatwg.org/specs/web-apps/current-work/#the-a
06:30:40 The definition of the attributes is clearly defined in a nice list there. Separated and easy to approach. You have it buried in paragraphs of text.
06:31:29 the list of attributes in the html4 spec says almost nothing normative
06:31:54 we can certainly add more (informative) text to the spec at some point when it's stable
06:31:59 but that's not spec material
06:32:37 and the html5 spec has a little contents list for jumping to the attribute definitions, if you actually want that
06:32:44 It's not just that. The organization of the spec is dense
06:32:57 If you don't know what you're looking for the HTML5 spec requires a lot more reading to find anything useful.
06:33:25 that's because there simply isn't anything useful _in_ the html4 spec
06:33:37 so if you want to find something useful in teh html4 spec, you can immediately know it's not there without looking
06:34:45 hah
06:34:49 you're way too biased
06:35:13 and much too close to the HTML5 spec to see how dense it really is. The person writing it understands it clearly. Of course it makes sense to you.
06:35:51 well, i'd be happy to rearrange the html5 spec to be easier to read, if you have any concrete suggestions
06:36:33 Hixie, break out the attributes into lists with nice headers. Starting the definition inline and making it red requires more scanning.
06:37:05 hmm, the Media Query [MQ] links don't work in Safari
06:37:28 the [...] links don't work at all, i haven't done the references section yet
06:37:33 the attributes wouldn't work that way
06:37:46 e.g. the link attributes are defined in an entirely different section
06:38:13 and some of the attributes have multiple definitions depending on the other attributes
06:38:33 generally i don't like keying things from attributes, because the attributes aren't what matter
06:40:10 Right now trying to find information about an attribute is a lot of work
06:40:27 Where do you define that an attribute is optional unless it says "required" next to it?
06:41:37 attributes are optional or required based on what it says next to their definition
06:41:45 the (required)/(optional) stuff is old and will all be removed
06:41:49 because it's often wrong
06:41:57 usually an attribute is required if something else applies
06:42:06 e.g. "one of these attributes must be present"
06:43:25 That's not very approachable to have to dig through a paragraph of prose to figure out if something is required or not
06:43:36 i don't really know how else to do it
06:43:46 attributes aren't just optional or required
06:45:53 Some are defined that way. It says the src attribute is required
06:46:24 You also jump around in the prose which makes it hard to find what you're looking for. It goes from talking about the src attribute, to talking about alt, to talking about the src attribute again
06:46:29 requiring src doesn't always make sense for