| 08:15 | <hsivonen> | http://www.webstandards.org/2008/02/24/wasp-round-table-ie8s-default-version-targeting-behavior/ |
| 08:27 | <roc> | when is MS going to explain exactly what the mode switch does? Sounds like it affects the DOM APIs but I haven't seen anything explicitly saying that |
| 08:33 | <hsivonen> | I don't like the way ALA and WaSP are the main PR channels for communicating about this. |
| 08:57 | <roc> | yeah |
| 08:57 | <roc> | they're obviously being told things under NDA |
| 09:36 | <Dashiva> | roc: Non-disagreement agreement? :P |
| 09:38 | <annevk> | man, where are all the tough questions in those transcripts |
| 09:38 | <annevk> | yay, we all agree, lets have beer |
| 09:42 | <jruderman> | Dashiva++ |
| 10:59 | <hsivonen> | microformats have arrived. RDFa Syntax introduces itself by contrasting itself with microformats: http://www.w3.org/TR/2008/WD-rdfa-syntax-20080221/ |
| 11:03 | <annevk> | makes sense i guess, lots of people heard of microformats |
| 11:05 | <hsivonen> | reminds me of http://en.wikipedia.org/wiki/Second-system_effect |
| 11:07 | <hsivonen> | of course, RDFa is fatally flawed because it doesn't work with HTML |
| 11:08 | <annevk> | it sort of does with simple dom traversal |
| 11:08 | <Philip`> | Bananas don't work with HTML either, but they still seem pretty popular, just in different problem domains :-) |
| 11:08 | <annevk> | it's just qnames in attribute values and magic xmlns attributes after all |
| 11:09 | <hsivonen> | annevk: true |
| 11:09 | <hsivonen> | still, it's sad that Creative Commons is advancing this complexity when they should endeavor to make things simple |
| 11:11 | <hsivonen> | I still don't know if http://intertwingly.net/blog/2008/02/09/Mashups-Smashups#c1202810109 was meant as a joke or as a serious question |
| 11:12 | <annevk> | prolly serious... |
| 11:12 | <annevk> | given the second sentence |
| 11:14 | <hsivonen> | btw, when Lessig uses CC-licensed photos off the Web in his presentations, he doesn't include the license URI... |
| 11:16 | <hsivonen> | hmm. it's 2008 and XHTML+RDFa document conformance deals with DTDs |
| 11:16 | <hsivonen> | sad |
| 11:20 | <mpt> | That appears to be an example of the "Appeal to the current year" logical fallacy :-) |
| 11:21 | <hsivonen> | mpt: true. |
| 11:23 | <annevk> | it's still sad |
| 11:23 | <annevk> | :p |
| 12:15 | <Philip`> | http://www.glendathegood.com/wasp/transcript.html - "The version vector plan right now, have the version vector [for conditional comments] and the UA string reflect the real version of IE8 and see what compatibility that turn out to be." - hmm, that would break things with "lte IE7" like http://lists.w3.org/Archives/Public/public-html/2008Jan/0189.html |
| 12:16 | <annevk> | in effect they still require pages to update sniffing, etc. |
| 12:17 | <Philip`> | Only pages which currently do sniffing, which is a subset of the pages that depend on IE6/7 bugs/features |
| 12:18 | <Camaban> | there's a suggestion there to make strict doctypes trigger 'ie8 mode' isn't there? |
| 12:18 | <Camaban> | (near the end) |
| 12:19 | <Philip`> | There was, but they didn't have any data about how common that was in practice |
| 12:20 | <Camaban> | yeah, a suggestion to look into more, rather than a "lets do it" suggestion |
| 12:20 | <Philip`> | (http://philip.html5.org/data/doctypes-lc.txt has data) |
| 12:21 | <annevk> | it's not clear how large the data set is and how many sites did not have a doctype |
| 12:22 | <Philip`> | annevk: It has a non-hyper link to a version with more detail |
| 12:22 | <Philip`> | It doesn't say how many don't have a doctype, since I didn't measure that, but I'll say it's 50% based on other data from the same source |
| 12:23 | <Camaban> | so a few hundred strict, standards mode triggering doctypes, compared to 6000 or so transitional, standards and quirks |
| 12:23 | <Camaban> | at a quick glance |
| 12:43 | <Philip`> | Would <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN"> be valid? (I don't see anybody using it without the SYSTEM URI) |
| 12:47 | <hsivonen> | Philip`: no |
| 12:48 | <hsivonen> | XML requires a system id |
| 12:48 | <Philip`> | Okay, thanks |
| 13:16 | <zcorpan_> | Philip`: do those doctypes get the same mode using the html5 algorithm? or rather, do you know (or can easily find out) which don't? |
| 13:17 | <zcorpan_> | just glancing i see html 4.0 transitional with SI |
| 13:18 | <zcorpan_> | <!doctype html public "-//"aol hometown//html 3.0 transitional//en"> |
| 13:18 | <Philip`> | zcorpan_: They won't all be the same, but I haven't tried to work out which ones will differ |
| 13:19 | <zcorpan_> | <!doctype htm public "-//w3c//dtd htm 4.0 transitional//en"> |
| 13:19 | <zcorpan_> | 1 - <!doctype html public "-//"aol hometown//html 3.0 transitional//de"> |
| 13:19 | <zcorpan_> | 1 - <!doctype html public "-//"aol hometown//html 3.0 transitional//fr"> |
| 13:19 | <Philip`> | <!doctype html public "-//w3c//dtd xhtml 1.0 strict//en" "http://www.w3.org/tr/xhtml2/dtd/xhtml1-strict.dtd">; |
| 13:20 | <zcorpan_> | that one would get standards mode per html5 too, right? |
| 13:21 | <zcorpan_> | i wonder if the "2" was a typo or deliberate |
| 13:22 | <Philip`> | I believe that'd be HTML5 standards |
| 13:22 | <zcorpan_> | <!doctype html public "-//ietf//dtd html strict level 2//en"> |
| 13:22 | <Philip`> | Perhaps the most significant difference is that IE treats ill-formed doctypes as standards, whereas HTML5 treats them as quirks |
| 13:22 | <Philip`> | (like in that AOL Hometown case) |
| 13:24 | <Philip`> | or maybe the most significant difference is that IE treats many typoed doctypes as quirks (if they still contain a blacklisted substring), whereas HTML5 treats them as standards |
| 13:27 | <zcorpan_> | yes, i think changing the last two characters to the locale of the page is not that uncommon |
| 13:48 | Philip` | tries doing a comparison against HTML5 |
| 15:00 | Philip` | wonders why Safari doesn't do compatMode |
| 16:03 | <Philip`> | Hmm, looks like someone deleted cnn.com from dmoz.org |
| 16:04 | <Philip`> | There used to be 223658 of it, but now there's only 3156 |
| 16:05 | <Philip`> | There's 4552771 URIs spread over 2982776 domains, and the top domain is www.geocities.com with 80118 |
| 16:14 | <Philip`> | (Top 1% of domains have 24% of URIs; top 10% have 36%) |
| 16:17 | <Philip`> | (I do like how I can do "sort -R four-point-five-million-line-text-file" and not need to even begin to worry about how much RAM it's using) |
| 18:11 | <annevk> | peoples: http://www.w3.org/TR/cssom-view/ & http://www.w3.org/TR/XMLHttpRequest2/ |
| 18:14 | <aroben> | annevk: neat |
| 18:16 | <aroben> | annevk: how many specs are you editing now? |
| 18:18 | <annevk> | apart from informal HTML drafts 5 I think (plus Selectors API which Lachlan is editing now) |
| 18:18 | <Dashiva> | He's probably aiming for 5 > 2 :) |
| 18:18 | <annevk> | heh |
| 18:18 | <annevk> | the CSSOM stuff isn't getting enough attention though as it requires a lot of research I haven't found time for |
| 19:18 | <Hixie> | annevk: pity someone edited the spec for you |
| 19:29 | <Hixie> | holy crap |
| 19:29 | <Hixie> | RDFa has gotten BIGGER since i last looked |
| 19:29 | <Hixie> | talk about serious second-system effect |
| 19:29 | <Hixie> | jesus |
| 19:33 | <takkaria> | I thought rdfa was meant to be small and compact |
| 19:34 | <Hixie> | i still don't understand how RDFa is supposed to interact with, e.g., <title> |
| 19:35 | <Hixie> | if an RDFa bit says the document title is X, and <title> says the document title is Y, then... what? |
| 19:35 | SadEagle | keeps Hixie away from ES4 :-) |
| 19:36 | <Hixie> | ES4 has one gigantic problem right now |
| 19:36 | <Hixie> | and it's not clear to me anyone on the ES4 group is striving to fix it |
| 19:36 | <Hixie> | but since the RI is the only spec, it's not clear to me how to proceed |
| 19:36 | <SadEagle> | do you mean the lack of spec, or the hyper second-system syndrome? |
| 19:36 | <Hixie> | neither |
| 19:37 | <Hixie> | ES4 scripts have to explicitly opt-in to being processed as ES4 instead of ES3, but the opt-in can only be out-of-band (e.g. in <script type=""> attributes) which doesn't work for a whole series of use cases |
| 19:38 | <Hixie> | e.g. it doesn't work as a way to upgrade existing deployed <script> blocks doing third-party includes, it doesn't work for HTML event handler attributes, it doesn't work when you don't know what type your script will be, etc |
| 19:38 | <Hixie> | it also means you need separate scripts for ES4 and ES3 |
| 19:39 | <Hixie> | this is unlike, e.g., HTML and CSS, both of which have backwards-compatible mechanisms |
| 19:40 | <SadEagle> | the language has a lot of complexity to be backwards compatible when ES3 is handled as ES4, though. |
| 19:40 | <SadEagle> | Partly for migration, of course |
| 19:41 | <SadEagle> | One can do this sort of thing in-band, though, but it'll be ugly. |
| 19:42 | <Hixie> | ugly is better than fatally flawed :-) |
| 19:43 | <SadEagle> | things like cross-interpreter calls would be a total mess, though. |
| 19:43 | <hsivonen> | is there a reason for not going all the way so that a browser could run all ES3 scripts as ES4 scripts? |
| 19:43 | <Hixie> | so have one interpreter |
| 19:43 | <Philip`> | When JavaScript reserved all the Java keywords, was that intended for future extensibility, or was it just for compatibility between Java and JS (so you wouldn't accidentally make variables you couldn't talk about in the other language) or something else? |
| 19:44 | <Hixie> | Philip`: everyone ignored that requirement, so it's moot, sadly |
| 19:44 | hsivonen | notes that v.nu doesn't ignore the reserved word list |
| 19:45 | <SadEagle> | it's actually not entirely ignored, but semi-weird |
| 19:45 | <hsivonen> | (in callback name checking) |
| 19:47 | <Philip`> | On the subject of JS, what syntax feature allows <a onclick="javascript: alert('hello')"> (and equivalently <a onclick="vbscript: alert('hello')"> etc)? |
| 19:51 | <annevk> | Hixie, if you mean the HTML5 reference, it's still part of XMLHttpRequest Level 2... guess Bert removed it from the CSSOM View Module or something |
| 19:51 | <Hixie> | man, i'd be pissed if someone did that to my spec |
| 19:52 | <Hixie> | Philip`: goto labels, no? |
| 19:52 | <annevk> | I don't feel too strongly about references at this point |
| 19:52 | <SadEagle> | heh, good point Hixie |
| 19:53 | <Hixie> | annevk: for me it's not about references, it's about someone changing something i'd have explicitly done |
| 19:53 | <Hixie> | but anyway |
| 19:54 | <Philip`> | Hixie: JS doesn't have goto |
| 19:54 | <Hixie> | oh |
| 19:54 | <Philip`> | so goto labels don't seem an entirely sensible thing to have |
| 19:54 | <Hixie> | well |
| 19:54 | <Hixie> | i seem to recall there was some sort of label thing |
| 19:54 | <Philip`> | Then again, it does have the goto reserved word despite not having goto |
| 19:55 | <Philip`> | so it's not an entirely sensible language |
| 19:55 | <annevk> | Hixie, true |
| 19:56 | <SadEagle> | Philip`: they're not goto labels, they're breka/continue labels. |
| 19:58 | <Philip`> | SadEagle: Oh, I didn't know JS had that |
| 19:58 | <Philip`> | (but some testing indicates that it does) |
| 19:59 | <SadEagle> | you can actually break a non-loop, too. |
| 19:59 | <SadEagle> | (with a labeled break, that is) |
| 20:00 | <Philip`> | By "non-loop", do you mean something like "javascript: { break javascript; }"? |
| 20:00 | <SadEagle> | yep. |
| 20:01 | Philip` | wonders why he had never noticed this feature |
| 20:01 | <Philip`> | (I knew Java did that, but not JS) |
| 20:03 | <hsivonen> | does java have non-loop break labels? |
| 20:05 | <hsivonen> | whoa! it indeed does. |
| 20:05 | <Philip`> | hsivonen: Eclipse doesn't give squiggly lines when I try that, so I guess so |
| 20:05 | <hsivonen> | I learned something new |
| 20:05 | <hsivonen> | thanks |
| 22:27 | <jruderman> | the phone meeting about cross-site xmlhttprequest in firefox is starting soon |
| 22:41 | <annevk> | Hixie, the <title> argument is bogus |
| 22:41 | <annevk> | Hixie, the spec should not be moving <title> around |
| 22:41 | <Hixie> | hm? |
| 22:41 | <Hixie> | even if it shouldn't, <title> is always gonna be PCDATA |
| 22:41 | <Hixie> | or CDATA |
| 22:41 | <annevk> | true |
| 22:41 | <Hixie> | or whatever its parse mode is called |
| 22:44 | <annevk> | also, what's the case where Opera does better than void? |
| 22:45 | <annevk> | i'm also curious what the use case is for a form control in a header/title |
| 22:45 | <Hixie> | templates |
| 22:47 | <annevk> | hmm, i don't quite get templates either :) |
| 22:48 | <Hixie> | opera seems to do better at least in the case where a <fieldset> is involved |
| 22:48 | <Hixie> | and re your second case, imagine |
| 22:49 | <annevk> | oh, with fieldset... hmm, but that's not the common case as you indicate |
| 22:49 | <Hixie> | <figure> <textarea name="poem"> </textarea> <legend> <input name="author">, <input type=number name="year"> </legend> </figure> |
| 22:49 | <annevk> | not that it matters much either way |
| 22:49 | <Hixie> | yeah |
| 22:49 | <Hixie> | i didn't mean to say opera was fine :-) |
| 22:50 | <Lachy> | Hixie, you wrote in the figure email: "<label> unfortunately would preclude putting more than one form control in a legend, which would be a weird restriction." - did you mean "control in a *figure*"? |
| 22:51 | <Hixie> | i meant in the legend of the figure |
| 22:51 | <Hixie> | as per my example above |
| 22:53 | <Lachy> | oh, ok. I should have read the IRC log before posting. |
| 22:53 | <Hixie> | heh |
| 22:55 | jgraham_ | still believes that making language features dependent on all major browsers changing their parsing in a largely untested way is not the greatest idea. |
| 22:56 | <jgraham_> | But I've said that before and I don't have any new arguments or anything |
| 22:57 | <annevk> | we can always revise our evil plans for world domination in two years |
| 22:57 | <Lachy> | jgraham_, we can always revisit the issue again if browsers come back after trying it and say it causes too many provlems |
| 23:01 | <gsnedders> | jgraham_: I'm currently leaving towards having supper together on Fri, FWIW |
| 23:11 | <Hixie> | jgraham_: i agree |
| 23:15 | mpt | wonders whether there should be an element for captions of form elements |
| 23:16 | <mpt> | If someone tabs to an input field, the browser automatically scrolls to show as much of the field as possible, but doesn't scroll to show that field's caption, because it doesn't know what a caption is |
| 23:18 | <mpt> | If there was a caption element that was tied to the input element, the browser would know to scroll to show it |
| 23:18 | <Hixie> | isn't that <label>? |
| 23:18 | <mpt> | no |
| 23:19 | <mpt> | <label> appears above or (in LTR scripts) to the left of an element |
| 23:19 | <mpt> | the caption appears underneath, and is usually in smaller print |
| 23:19 | <Hixie> | can you show me a page with the difference you are indicating? |
| 23:20 | <mpt> | <https://launchpad.net/projects/+new>;, but it requires a Launchpad account |
| 23:21 | mpt | looks for another |
| 23:22 | <Hixie> | annevk: would be cool if one could bookmark the settings on http://html5.org/tools/web-apps-tracker |
| 23:22 | <Hixie> | would also be cool to have an atom feed or something for that |
| 23:23 | <Hixie> | and finally, it would also be cool to adda Google Gears icon to the list of vendors, maybe with the letter 'r' as the way for me to trigger it |
| 23:23 | <annevk> | is anyone actually using that interface? |
| 23:23 | <Hixie> | they have apparently asked if they can be added since they are about to start doing more html5-y stuff |
| 23:23 | <Hixie> | i dunno if anyone else is, but i certainly use it |
| 23:23 | <annevk> | i guess i'm not representative |
| 23:24 | <Hixie> | from #webkit: |
| 23:24 | <Hixie> | 00:28 < aroben> Hixie: I find the tracker very handy, though |
| 23:24 | <Hixie> | (referring to the page) |
| 23:24 | <annevk> | sorry, i'm using the tracker *a lot* |
| 23:24 | <annevk> | i meant the UA-specific crap |
| 23:24 | <annevk> | all those options |
| 23:24 | <annevk> | (the colors and icons in the changelog lines are useful though, to me) |
| 23:25 | <Hixie> | i like the options, i would probably use it more to point to other people if i could bookmark them though |
| 23:25 | <Hixie> | if you want an icon for gears, there's one at http://google-gears.googlecode.com/svn/trunk/gears/ui/common/icon_16x16.png |
| 23:27 | <mpt> | Hixie, in <http://login.live.com/login.srf?wa=wsignin1.0&rpsnv=10&ct=1203982238&rver=4.5.2130.0&wp=MBI&wreply=http:%2F%2Fmail.live.com%2Fdefault.aspx&id=64855>; both the text fields have a label and a caption |
| 23:27 | <mpt> | (though in that page the scrolling isn't important, because you can probably see both without scrolling already) |
| 23:27 | <Hixie> | i disagree with calling that a caption |
| 23:27 | <Hixie> | but i agree they are related to the field |
| 23:27 | <Hixie> | and are not labels |
| 23:28 | <Hixie> | they're like the (?) part of the checkbox labels |
| 23:28 | <Hixie> | interesting cases |
| 23:28 | <mpt> | I don't know what to call them other than captions |
| 23:29 | <mpt> | <https://www.google.com/accounts/NewAccount?service=mail&continue=http%3A%2F%2Fmail.google.com%2Fmail%2Fe-11-10b567b3542c73d051854a3694480bd-27a8b709838b6e5bc593de69344785f3ae19edf1&type=2>; has examples too |
| 23:30 | <Hixie> | yeah it's a common idionm |
| 23:30 | <Hixie> | idiom |
| 23:30 | <Hixie> | most are basically help strings |
| 23:30 | <mpt> | Tab to the "Secondary email" field, and you won't see its caption, because the browser doesn't know any better. |
| 23:30 | <Hixie> | in fact i think all of these are really help strings |
| 23:31 | <Hixie> | might make sense to have a <label> ... <input> <help>...</help> </label> or <label for="">...</label> <input id=""> <help for="">...</help> construct |
| 23:33 | <mpt> | But then, see that Gmail has "Learn More" links |
| 23:33 | <mpt> | *That* is help |
| 23:34 | <Hixie> | sure, you can have inline short help and external longer help |
| 23:34 | <Hixie> | i've noted this for wf3. |
| 23:34 | <mpt> | Right, it's a continuum |
| 23:34 | <mpt> | thanks |
| 23:34 | <Hixie> | thank _you_! |
| 23:35 | <annevk> | man, web-apps-tracker is a hack |
| 23:35 | <annevk> | i've to add google-gears to like five different places |
| 23:35 | <annevk> | scary |
| 23:35 | <Hixie> | hehe |
| 23:35 | <Hixie> | feel free to make a new one that's better :-) |
| 23:36 | <Hixie> | afk. i really do have to actually go to work... or i'll starve. |
| 23:45 | <annevk> | ok, four different places |
| 23:45 | <annevk> | gears is now supported, all the fancy features will have to wait |
| 23:48 | <annevk> | Hixie, "r" triggers Google Gears as I couldn't think of something better either |
| 23:50 | annevk | -> bed |
