| 05:18 | <heycam> | huh. occasional W3C mailing list contributor http://www.w3.org/Search/Mail/Public/search?keywords=Slim+Amamou now a Minister in the new Tunisian Government after being arrested last month. http://torrentfreak.com/arrested-pirate-party-member-becomes-tunisian-minister-110117/ |
| 05:50 | <MikeSmith> | wow |
| 05:50 | <MikeSmith> | that's pretty cool |
| 05:51 | <MikeSmith> | cool name too |
| 08:32 | <hsivonen> | how does ES5 distinguish between illegal programs that work in a well-defined way and legal programs that work in a well-defined way? |
| 08:33 | <hsivonen> | Suppose one writes "\/" |
| 08:33 | <hsivonen> | I see that / doesn't match SingleEscapeCharacter |
| 08:33 | <hsivonen> | and I see that the above is therefore equivant with "/" |
| 08:34 | <hsivonen> | but what cues should I look for to know it is conforming? |
| 08:34 | <hsivonen> | is stuff that "works" always conforming in ES5? |
| 08:35 | <hsivonen> | clearly, I can't use JSLint, since it encodes Crockford's opinion--not the spec |
| 08:36 | <Hixie> | isn't anything that matches the grammar conforming? |
| 08:37 | <Hixie> | and vice-versa? |
| 08:37 | <Dashiva> | Working in practice is not a guarantee of being conforming, at least, since browsers aren't always 100% by the spec |
| 08:39 | <hsivonen> | Hixie: dunno |
| 08:39 | <zcorpan> | 2 Conformance just talks about conforming implementations |
| 08:41 | <hsivonen> | Hixie: isn't the whole semicolon insertion mechanism based on first hitting an error in the grammar and then magically correcting it? |
| 08:41 | <Hixie> | hsivonen: beats me |
| 08:42 | <Hixie> | haha, http://lists.w3.org/Archives/Member/w3c-ac-members/2011JanMar/0006.html |
| 08:42 | <Hixie> | w3c, ever on time |
| 08:42 | <Hixie> | just as we're moving on from html5... |
| 08:47 | <zcorpan> | time for s/HTML5 (including next generation additions still in development)/HTML/ ? |
| 08:48 | <hsivonen> | zcorpan: not yet, I think |
| 08:49 | <erlehmann> | zcorpan, but “HTML” is not buzzword tiem compatible. |
| 08:49 | <erlehmann> | “HTML5“, OTOH … ! |
| 08:50 | <Dashiva> | hsivonen: Semicolon insertion only works once, if there's still an error after insertion it's an invalid program |
| 08:53 | <hsivonen> | Dashiva: "invalid program" is not a term defined by ES5 |
| 08:57 | <Dashiva> | Well, it produces a syntax error |
| 09:04 | <zcorpan> | Hixie: is "Event handler content attributes, when specified, must contain valid JavaScript code matching the FunctionBody production. [ECMA262]" correct considering semicolon insertion? |
| 09:05 | <Hixie> | zcorpan: yeah, i want to do that again, but i guess the buzzword still has enough life in it... sometime this year, probably. |
| 09:06 | <annevk> | http://www.w3.org/Bugs/Public/show_bug.cgi?id=11777 is funny |
| 09:06 | annevk | starts a list of things that happened this week so he won't forget |
| 09:06 | <Hixie> | zcorpan: no idea |
| 09:06 | <Hixie> | zcorpan: actually, yes, since newlines get lost in attributes |
| 09:11 | <zcorpan> | Hixie: what? newlines don't get lost in attributes |
| 09:11 | <jgraham> | hsivonen: Surely the practical approach for you is to use the parser component from Rhino and see if it generates an AST? |
| 09:12 | <MikeSmith> | you can test it with the validator now |
| 09:12 | <MikeSmith> | if you check the value of an event-handler attribute |
| 09:12 | <hsivonen> | jgraham: this isn't for V.nu. this is for a person complaining about Mozilla Developer Network documentation about escaping <script> inside <script> |
| 09:13 | <jgraham> | hsivonen: pointer? |
| 09:13 | <Hixie> | zcorpan: in some UAs, I meant |
| 09:13 | <Hixie> | zcorpan: and in XML |
| 09:13 | <hsivonen> | jgraham: no public reference |
| 09:13 | <hsivonen> | jgraham: oh did you mean the document being complained about? |
| 09:13 | <jgraham> | hsivonen: To the MDC page? |
| 09:14 | <zcorpan> | Hixie: i don't care about some UAs, i'm talking about whether the spec is correct or has a bug :) for XML you can get newlines in attributes with |
| 09:14 | <hsivonen> | jgraham: https://developer.mozilla.org/en/HTML/HTML5/HTML5_Parser |
| 09:14 | <Hixie> | zcorpan: i guess |
| 09:14 | <Hixie> | zcorpan: dunno then |
| 09:14 | <Hixie> | zcorpan: i'm happy to change it if someone can tell me what it _should_ say |
| 09:16 | <jgraham> | hsivonen: I don't understand the complaint |
| 09:16 | <jgraham> | I note that '<' + '/script>' will also work for the usual case of a string literal |
| 09:18 | <brucel> | 'scuse me all. Quick Q about forms. Trying to understand if entering the letter "K" into input type=number should be an error. It's not out of range for a number eg min=1 max=10 but seems to me it should be a type mismatch. However, spec says only email, url can have type mismatch. http://dev.w3.org/html5/spec/Overview.html#suffering-from-a-type-mismatch or am i reading it wrong? |
| 09:19 | <zcorpan> | MikeSmith: onclick="(" validates in v.nu |
| 09:19 | <Peter`> | brucel, the rule specifically talks about controls which allow arbitrary user input, which <input number> does not |
| 09:20 | <MikeSmith> | zcorpan: well that's bad… guess we'll have to fix it |
| 09:20 | <MikeSmith> | all it's doing now is handing it straight off to the rhino compiler |
| 09:20 | <Peter`> | brucel, constrained input for <input type=number> could be interpreted as "just throw everything away that's not a number" I guess |
| 09:20 | <zcorpan> | MikeSmith: got an example where it gives an error? |
| 09:21 | <Peter`> | which is what various OS controls do as well |
| 09:22 | <MikeSmith> | zcorpan: yeah, that one |
| 09:22 | <MikeSmith> | if you try it at http://www.w3.org/html/check |
| 09:22 | <zcorpan> | ah |
| 09:22 | <MikeSmith> | I guess hsivonen didn't redeploy that code at validator.nu yet |
| 09:22 | <MikeSmith> | thought he had on Friday |
| 09:22 | <hsivonen> | MikeSmith: ooops. sorry. |
| 09:23 | <MikeSmith> | no biggie |
| 09:23 | <Hixie> | brucel: the browser will never set the value to "K" even if it displays it in the contrcol |
| 09:23 | <Hixie> | control |
| 09:24 | <brucel> | Hixie, but should it be a type mismatch? |
| 09:24 | <zcorpan> | MikeSmith: cool |
| 09:24 | <Hixie> | brucel: should what be a type mismatch? |
| 09:24 | <Hixie> | brucel: say you started with <input type=number value=0> and typed K in it. The value is still going to be "0", which is a number -- not a type mismatch. |
| 09:25 | <Hixie> | the display might say "K" but what is validated is the value, not the display. |
| 09:25 | <brucel> | well, having a type mismatch is defined as "Suffering from a type mismatch: When a control that allows arbitrary user input has a value that is not in the correct syntax (E-mail, URL)." |
| 09:25 | <zcorpan> | MikeSmith: next up is checking <script> if the type is javascript? :) |
| 09:25 | <MikeSmith> | lord please no |
| 09:26 | <MikeSmith> | does the spec require that now? |
| 09:26 | <MikeSmith> | it didn't last time I checked |
| 09:26 | <Hixie> | brucel: right, so? |
| 09:26 | <Hixie> | brucel: type=number doesn't allow arbitrary user input |
| 09:27 | <jgraham> | MikeSmith: It would be quite inconsistent if it didn't |
| 09:27 | <jgraham> | (so it probably should in any case) |
| 09:27 | <MikeSmith> | well, that would be in line with the platform itself |
| 09:27 | <MikeSmith> | yeah |
| 09:27 | <MikeSmith> | true |
| 09:27 | <MikeSmith> | could warn about it at least |
| 09:27 | <Hixie> | it should check it against python if the type is python, too :-) |
| 09:27 | <Hixie> | and against 80286 assembler if the type is 80286 assembler :-) |
| 09:27 | <MikeSmith> | simple warning for that case |
| 09:28 | <Hixie> | (but the validator isn't required to support any languages) |
| 09:28 | <MikeSmith> | "warning: don't do that" |
| 09:28 | <Hixie> | (iirc) |
| 09:28 | <annevk> | brucel, what is somewhat difficult, but important here, is that the new form controls have several values |
| 09:28 | <MikeSmith> | Hixie: you made it require checking for event-handler attributes |
| 09:28 | <zcorpan> | MikeSmith: appears the spec doesn't require <script> to compile |
| 09:28 | <brucel> | Hixie anneck: so there can never be a type mismatch for type=number, so to check that the user has entered something numeric, yoiu have to use some regex with pattern attrib, right? |
| 09:28 | <annevk> | brucel, a "UI value" (what the user sees), an "actual value" (what the user agent has computed), and maybe things such as a "default value" |
| 09:29 | <MikeSmith> | zcorpan: w00t |
| 09:29 | <Hixie> | MikeSmith: yeah, well, there's no way to change the language for event handler attributes |
| 09:29 | <Hixie> | MikeSmith: so it's not open-ended |
| 09:29 | <annevk> | brucel, e.g. for URL the user might see "www.example.org" which is not valid, but the "actual value" might be "http://www.example.org"; which is |
| 09:29 | <Hixie> | MikeSmith: <script> is open-ended |
| 09:30 | <annevk> | brucel, for Number the idea is that you never change it to a non-number, only a number that is not within the range |
| 09:30 | <jgraham> | Hixie: So if you specify <script type="text/javascript"> the spec doesn't care if there is a SyntaxError? Why the difference? |
| 09:30 | <annevk> | brucel, so if the UI allowed entering K (and displaying it back) (the UI value), the "actual value" would still be a number |
| 09:31 | <Hixie> | holy cow, i really didn't require that the <script>'s contents be valid |
| 09:31 | <brucel> | annevk so if I have <input type=number> what actually is the number? |
| 09:31 | <Hixie> | that's totally bogus |
| 09:31 | Hixie | files a bug |
| 09:31 | <SlexAxton> | brucel: "" |
| 09:31 | <zcorpan> | Hixie: i already did |
| 09:31 | <Hixie> | ah |
| 09:31 | <Hixie> | well then |
| 09:31 | <MikeSmith> | heh |
| 09:31 | <Hixie> | dupes all around! |
| 09:32 | SlexAxton | submits as well |
| 09:32 | <annevk> | brucel, well okay, no value is also an option |
| 09:33 | <annevk> | brucel, another important thing is that the interaction between the UI value and actual value is up to the user agent, much like all UI decisions |
| 09:33 | <phrearch> | hey |
| 09:33 | <annevk> | brucel, so entering "K" in the UI could mean erasing the value or could mean not changing the current value, or changing it to zero |
| 09:33 | <phrearch> | i was wondering whether there are webkit branches that have some experimental websocket features, like the sony ericsson webcam-over-websocket demo |
| 09:33 | <annevk> | brucel, ideally though UI implementations are not that crazy and simply disallow that :) |
| 09:34 | <Hixie> | entering "K" could mean setting the value to 1000! :-) |
| 09:34 | <brucel> | annevk hixie thanks. Pretty unintuitive so glad I asked and thx for answering. SO my workaround has to be use a pattern attribute to not accept non-numerics right? |
| 09:34 | <annevk> | phrearch, #webkit might be a better channel |
| 09:34 | <phrearch> | annevk: ok thanks |
| 09:34 | <annevk> | brucel, no |
| 09:36 | <annevk> | brucel, pattern="" does not even apply, why would you need it? |
| 09:36 | <SlexAxton> | to prevent 'K' - regardless of meaning, I would assume |
| 09:36 | <brucel> | annevk want to check that user entered a number = numerics |
| 09:37 | <zcorpan> | brucel: the user can't enter anything but a number |
| 09:37 | <annevk> | brucel, that is all the user can do per the constraints of the specification (note that the UI value is never exposed to anyone but the user) |
| 09:39 | <SlexAxton> | Would it be a bug in chrome if your element starts as <input type=number value=15 /> and if the user deletes the characters in the input the elem.value is still '15' and not '' ? |
| 09:39 | <SlexAxton> | or would that conform to spec |
| 09:39 | <brucel> | annevk zcorpan OK. So I think I'm getting a soupcon of comprehension. So: the ability to type something non-numeric is a UA decision (it could require you to choose from a spinner, for example). What the user actually type has no relation to the contents of the field. But how to give sensible feedback to the user? |
| 09:39 | <annevk> | it could theoretically conform to the spec, but it sounds like a bug in the UI SlexAxton |
| 09:39 | <SlexAxton> | kk, will file. |
| 09:40 | <annevk> | brucel, the UA should give sensible feedback |
| 09:40 | <annevk> | brucel, if the UA has not implemented the control well, there are two options |
| 09:40 | <Hixie> | brucel: it's up to the UA to worry about all this |
| 09:40 | <Hixie> | brucel: that's the whole point of semantic markup -- you say what you want, and the browser takes care of it |
| 09:41 | <annevk> | brucel, a) get the UA fixed; b) wait for the UA to implement something XBL-like and fix it yourself |
| 09:42 | <brucel> | gotcha. thanks annvk and hixie |
| 09:45 | <MikeSmith> | hsivonen: you looking at section 16 of the EcmaScript spec? |
| 09:45 | annevk | will be back in an hour; is going to attempt working in an office for half a day |
| 09:46 | <MikeSmith> | hsivonen: it seems like the spec doesn't so much state the notion of a conforming program at all |
| 09:46 | <MikeSmith> | it just says what should be reported as errors |
| 09:47 | <MikeSmith> | and "All errors that are not early errors are runtime errors." |
| 09:47 | <Hixie> | i guess that's one way to define it |
| 09:47 | <Hixie> | ok i should go to bed |
| 09:47 | <Hixie> | nn |
| 09:48 | <jgraham> | You might have to special-case javascript and say that a program has to be valid after applying automatic semicolon insertion |
| 09:52 | <SlexAxton> | well, un semi-colon'd javascript is still valid, no? |
| 09:58 | <matjas> | unless you remove a semicolon that’s actually required, yes |
| 09:59 | <jgraham> | Not per spec, arguably |
| 10:01 | <jgraham> | I mean, if there is no error there is no automatic insertion |
| 10:02 | <hsivonen> | jgraham: the complaint is that \u003C!-- is longer than <\!-- |
| 10:03 | <hsivonen> | MikeSmith, zcorpan: the deployment script has finished now |
| 10:04 | <phrearch> | hm, maybe anyone here knows the answer to this? according to http://stackoverflow.com/questions/3806100/websocket-for-binary-transfer-of-data-decode , chrome7+ should support binary transfers over websockets? |
| 10:04 | <phrearch> | probably just a rumour, but.. |
| 10:06 | <hsivonen> | where is "syntax error" defined in ES5? |
| 10:07 | <hsivonen> | Hixie: speaking of <script> being open-ended, did you see the emails to public-html-xml over the last 12 hours or so? |
| 10:07 | <zcorpan> | phrearch: i'm pretty sure binary messages aren't supported yet |
| 10:07 | <Hixie> | hsivonen: i can happily report that i don't read that list |
| 10:08 | <phrearch> | aha ok. thanks |
| 10:08 | <hsivonen> | Hixie: I see |
| 10:08 | <Hixie> | hsivonen: should i read the recent thread? |
| 10:09 | <Hixie> | this thread doesn't start in a direction i would assume will lead to me being interested... |
| 10:09 | <Hixie> | apparently norm doesn't know about vbscript |
| 10:11 | <hsivonen> | Hixie: well, if you want to reply on the topic, then it makes sense to read the thread |
| 10:11 | <Hixie> | i can't imagine that i would |
| 10:11 | <Hixie> | this tf isn't going to do anything to change the web |
| 10:12 | <Hixie> | so it can't improve the web, which is a subset of changing the web |
| 10:12 | <Hixie> | and it's not a topic i find personally interesting |
| 10:12 | <Hixie> | so there's really not much reason to partake, as far as i can see |
| 10:14 | <Hixie> | having now read the whole thread, i can't really see what i would contribute anyway |
| 10:14 | <Hixie> | in any case, i really do have to sleep now :-) |
| 10:14 | <Hixie> | nn |
| 10:36 | <nielsle> | Does planet html5 have a feed where twitter messages are filtered out? |
| 12:25 | <annevk> | http://www.w3.org/html/logo/ ? |
| 12:25 | <annevk> | via Peter` |
| 12:26 | <Peter`> | Is the W3C now officially going widen the definition of "HTML5" to "any (partially) avaialble web technology in 2011"? |
| 12:27 | <annevk> | it's not like embrace and extend has not worked before |
| 12:28 | <annevk> | I guess it is better to do it yourself than let someone else do it to you |
| 12:28 | <Peter`> | Fair point |
| 12:28 | <hsivonen> | What's the deal with the military rank symbol for semantics? |
| 12:29 | <hsivonen> | The styling symbol looks like military insignia, too. |
| 12:30 | <hsivonen> | Sarge Semantics? |
| 12:30 | <Peter`> | "If, as W3C hopes, the community embraces the logo, W3C will adopt it as its own official logo for HTML5 in the first quarter of 2011." |
| 12:31 | <adactio> | Great. Now the W3C are encouraging people to think of CSS3 as being part of HTML5. I give up. |
| 12:31 | <Peter`> | Exactly my thought, telling people it's not true will be a lost cause if this catches on |
| 12:34 | <annevk> | Will be interesting to see what the Wikipedia consensus is going to be. Or maybe the flamewars will be more interesting. |
| 12:34 | <adactio> | I feel a blog post coming on. |
| 12:35 | jwalden | feels a sudden urge to do some trolling |
| 12:35 | <jwalden> | standards nerd catnip |
| 12:35 | <annevk> | I encourage laughing instead :) |
| 12:46 | <Lachy> | it's not that bad. They are providing separate logos to represent separate feature sets, and the way the logo builder works just seems to illustrate that the technologies all work together. |
| 12:47 | <Lachy> | also, the use of "HTML5" as the umbrella term to cover everything is already widespread. That's not a problem that can be solved without a replacement term, and many have been tried and eventually failed over the years |
| 12:48 | <annevk> | We already gave up on HTML5 some time ago for things other than marketing. It's called HTML these days. |
| 12:49 | <erlehmann> | Can't the HTML5 Super friends adopt the logo, getting spandex suits in the process? |
| 13:01 | karlcow | is discovering http://www.w3.org/html/logo/ |
| 13:01 | <karlcow> | I mean in the logs here |
| 13:01 | <annevk> | it's also plastered all over twitter by now |
| 13:01 | <karlcow> | annevk: what is that? that twitter thing :) |
| 13:02 | <annevk> | it's like the Internet and SMS had a baby |
| 13:02 | <Ms2ger> | http://twitter.com/karlcow ;) |
| 13:03 | <karlcow> | hehe :) Ms2ger this one is the private one. :) the public one is http://twitter.com/karlpro |
| 13:04 | <zcorpan> | private twitter? does not compute |
| 13:05 | <karlcow> | zcorpan: reserved to only a few people for comments to people who would be interested by these comments. |
| 13:07 | <karlcow> | but if you meant private in the sense not viewable by twitter dragons… I define my level of opacity |
| 13:07 | karlcow | sees http://www.w3.org/QA/2011/01/an_html5_logo |
| 13:09 | zcorpan | already made an html5 logo years ago http://simon.html5.org/valid-html5.png |
| 13:09 | jgraham | wonders who chose "zeldman orange" for the logo |
| 13:10 | <karlcow> | Michael Nieling |
| 13:10 | <jgraham> | So it's him that I will offend by not liking it? (Not just the orange) |
| 13:11 | <zcorpan> | wait, the logo is for selling t-shirts to fund the html5 testsuite? |
| 13:11 | <jgraham> | For real? |
| 13:12 | <Ms2ger> | As in, everyone who submits a test gets a beer? |
| 13:12 | jgraham | doesn't like that plan |
| 13:12 | <jgraham> | (the beer one) |
| 13:13 | <karlcow> | jgraham: too many tests? You fear for your liver? |
| 13:13 | <zcorpan> | "We are also selling T-shirts and part of the proceeds will support the HTML5 test suite effort." http://www.w3.org/QA/2011/01/an_html5_logo |
| 13:14 | <zcorpan> | wonder why they didn't make the logo say "5 > 2" |
| 13:14 | <jgraham> | :) |
| 13:15 | <Peter`> | The term "W3C" actually is a trending topic on Twitter now |
| 13:15 | <Ms2ger> | (Or an equivalent non-alcoholic beverage) |
| 13:16 | <Workshiva> | That's a terrible logo |
| 13:16 | <jgraham> | Hmm. What non-alcoholic beverage is equivalent to beer? |
| 13:16 | <Workshiva> | non-alcoholic beer? |
| 13:16 | <jgraham> | Ugh |
| 13:16 | <karlcow> | jgraham: That would be budweiser |
| 13:17 | <jgraham> | I still don't like this plan |
| 13:18 | <Workshiva> | "This Google App Engine application is temporarily over its serving quota. Please try again later. " |
| 13:19 | <annevk> | I'll take jgraham's beer |
| 13:20 | <jgraham> | This is not getting better :p |
| 13:23 | <annevk> | Ms2ger, haha |
| 13:23 | <annevk> | (re 11777) |
| 13:23 | <Ms2ger> | :) |
| 13:26 | <hsivonen> | Philip`: your font subsetter tells me: Too late for "-CA" option at subset.pl line 1. |
| 13:26 | <hsivonen> | Philip`: what should I do? |
| 13:28 | <hsivonen> | Philip`: it runs if I zap -CA... |
| 13:32 | <annevk> | where does HTML5 define what happens with displaying <table><form><tr><td>x ? |
| 13:32 | <annevk> | especially the <form> |
| 13:32 | <annevk> | it becomes an empty child of <table> per the parser |
| 13:33 | <annevk> | it ought to be display:none somehow, right? |
| 13:34 | <zcorpan> | in the rendering section |
| 13:35 | <annevk> | ooh |
| 13:35 | <annevk> | irrespective of CSS rules |
| 13:35 | <Ms2ger> | Gecko removed its table>form rule recently, IIRC |
| 13:36 | <Philip`> | hsivonen: Run it like "perl -CA subsetter.pl" or "./subsetter.pl", I think |
| 13:36 | <Philip`> | (not like "perl subsetter.pl") |
| 13:36 | <annevk> | so even if I do form { display:block; outline:4em solid purple } nothing should show up? |
| 13:36 | <annevk> | okay... |
| 13:36 | <Philip`> | (The -CA option is to make it parse command-line arguments as UTF-8, I think) |
| 13:37 | <zcorpan> | Ms2ger: what? why? |
| 13:38 | <Ms2ger> | I didn't recall correctly, apparently |
| 13:41 | <annevk> | thanks for the help btw |
| 13:41 | <hsivonen> | Philip`: thanks |
| 13:42 | <hsivonen> | Philip`: any idea why browsers are OK with fonts that I subset from .ttf original but are unhappy with loading subsets made by subsetting a .ttf converted from .otf using fontforge |
| 13:43 | <Philip`> | What kind of unhappy? |
| 13:43 | <hsivonen> | (I changed em box from 1000 to 1024, added points at extrema and rounded to integer coordinates) |
| 13:43 | <hsivonen> | Philip`: Firefox and Chrome don't load the font |
| 13:45 | <Philip`> | Hmm, I think Fontforge OTF->TTF conversion has worked for me in the past (without bothering to fix any of those warnings about em boxes or whatever), except for being ugly |
| 13:45 | <Philip`> | Can you load the un-subsetted TTF-from-OTF in those browsers? |
| 13:48 | <hsivonen> | Philip`: yes |
| 13:50 | <Philip`> | hsivonen: Hmm |
| 13:50 | <Philip`> | Sounds somewhat like a subsetter bug, then :-( |
| 13:53 | <hsivonen> | Philip`: font file is http://webm.html5.org/fonts/BaskervaldADFStd-Heavy.ttf and http://webm.html5.org/fonts/heading-characters.txt is the character file |
| 13:54 | <hsivonen> | Philip`: http://webm.html5.org/fonts/Baskervald.ttf is the output that doesn't work |
| 13:54 | <zcorpan> | "What’s" |
| 13:55 | <hsivonen> | maybe I should give up on Web fonts anyway. fonts look ugly on Windows anyway |
| 13:56 | <Philip`> | ttx says |
| 13:56 | <Philip`> | An exception occurred during the decompilation of the 'OS/2' table |
| 13:56 | <Philip`> | which doesn't sound good |
| 14:08 | <Philip`> | hsivonen: https://bitbucket.org/philip/font-optimizer/changeset/3ec118f146c7 |
| 14:08 | <Philip`> | hsivonen: Does it work better with that patch? |
| 14:08 | Philip` | can't trivially test, but at least the output .ttf isn't blatantly invalid now |
| 14:09 | <hsivonen> | Philip`: Yes! Thank you! |
| 14:09 | <Philip`> | (I assume I never hit that problem myself because I was using older versions of Fontforge which didn't use the new table version) |
| 14:10 | <hsivonen> | in other news, I suck at choosing color schemes for Web pages |
| 14:11 | <Rik`> | hsivonen: you're not alone |
| 14:11 | Philip` | thinks the text shadows make it look blurry and painful to read |
| 14:13 | <hsivonen> | Philip`: better now? |
| 14:14 | <Philip`> | Thanks, yes (in my entirely non-expert opinion) |
| 14:15 | <annevk> | hsivonen, I think the your browser does/does not support should be at the top |
| 14:15 | <Philip`> | though why does the site benefit from using custom fonts rather than standard built-in ones? |
| 14:15 | <annevk> | hsivonen, ideally with a mark of approval/disapproval |
| 14:15 | <annevk> | Philip`, prettyness |
| 14:15 | <annevk> | hsivonen, as in "yay/nay" and then "why do I care" |
| 14:50 | jwalden | discovers the aforementioned logo |
| 14:50 | <jwalden> | egad |
| 14:50 | <annevk> | hsivonen, did you see my comments? |
| 14:51 | <hsivonen> | annevk: which comment where? |
| 14:52 | <hsivonen> | annevk: oh, in this log. OK. I can fix that after the telecon |
| 14:52 | <annevk> | k |
| 14:52 | <hsivonen> | there's a telecon in 3 minutes, right? |
| 14:52 | <annevk> | I think so |
| 14:52 | Philip` | thinks T-shirt proceeds that go to the Testing TF ought to be distributed to participants in proportion to the number of approved tests they have written |
| 14:52 | <annevk> | hsivonen, also, I think the browser recommendations can be simplified somewhat |
| 14:53 | <jgraham> | Philip`: That seems to encourage a non-helpful method of writing tests |
| 14:53 | <hsivonen> | the message that the testing effort is underfunded unless you buy t-shirts is sad |
| 14:53 | <annevk> | hsivonen, maybe simply "These browsers support WebM:" [large logo] [large logo] [large logo] |
| 14:53 | <jgraham> | Like looping over all unicode codepoints and calling each one a test |
| 14:53 | <annevk> | or "These browsers support WebM on Mac OS X:" if we want to keep that |
| 14:54 | <annevk> | no need to mention when they added support |
| 14:54 | <jgraham> | hsivonen: I'm not convinced about the heading font or the bg colour |
| 14:54 | <jgraham> | Also, I would lose the word "rival" |
| 14:54 | <hsivonen> | annevk: the point of saying when they added support is to convince the user of and old version that they really have an old version |
| 14:54 | <Philip`> | jgraham: We could restrict it to test cases that were written before the money got involved, to avoid perverse incentives |
| 14:54 | <hsivonen> | jgraham: the color totally sucks, I know |
| 14:55 | <annevk> | hsivonen, maybe we should deal with that specifically then |
| 14:55 | <Philip`> | though preferably wait until after the rest of the canvas tests get approved |
| 14:55 | <annevk> | hsivonen, it does not make much sense for people using Safari/IE |
| 14:55 | <hsivonen> | jgraham: also, feel free to suggest a font |
| 14:55 | <hsivonen> | annevk: I didn't do the [large logo] thing, because I didn't want to think about trademark policies |
| 14:56 | <annevk> | the other "pick a browser" sites get away with it right? hmm |
| 14:56 | <jgraham> | hsivonen: You could, um, hop on the HTML5 logo bandwagon and use the same fonts as http://www.w3.org/html/logo/ Although I don't really like the body text there |
| 14:56 | <jgraham> | Dunno about the licensing either |
| 14:56 | <jgraham> | One might hope W3C chose things under a RF license, but who knows |
| 14:57 | <annevk> | My main concern is that the UI is too complex |
| 15:15 | <erlehmann> | Is it Geolocation? Is it CSS3? No, its SuperHTML5!! |
| 15:17 | <miketaylr> | :D |
| 15:21 | <karlcow> | http://www.w3.org/html/logo/faq.html#how-licenced |
| 15:21 | <karlcow> | How is this work licensed? |
| 15:21 | <karlcow> | The logo, icons, and website are licensed under Creative Commons Attribution 3.0 Unported |
| 15:22 | <karlcow> | http://creativecommons.org/licenses/by/3.0/ |
| 15:23 | <jgraham> | karlcow: So if you just copy the example code, you are breaking the license? |
| 15:24 | <karlcow> | jgraham: and? lawyers with black helicopters will come after you and put you in guatanamo? |
| 15:25 | <zcorpan> | maybe the fonts have their own licenses |
| 15:26 | <jgraham> | karlcow: Well I'm pretty sure that there are lobbyists that would describe copyright infringment as cultral terrorism. So, uh, maybe. |
| 15:26 | <Philip`> | Why attempt to require attribution? |
| 15:26 | <Philip`> | It's not like the link between HTML5 and W3C is particularly subtle and in need of promotion |
| 15:27 | <Philip`> | and it wastes the time of people who care about following licence terms accurately |
| 15:28 | karlcow | is trying to imagine Ian Jacobs with a black suit and black glasses… hmm no that would be Mike |
| 15:28 | <jgraham> | Not just attribution, but a link back to the license (or the complete text of the license) |
| 15:29 | <zcorpan> | adactio: indeed, the right term is just "HTML" |
| 15:30 | <adactio> | zcorpan: Yes, comrade. |
| 15:34 | <annevk> | adactio, I should have read your article completely |
| 15:34 | <annevk> | lol |
| 15:34 | <annevk> | <-- idiot |
| 15:35 | <jgraham> | adactio: So we think that W3C are subtly pushing the Open Standards === Communism line? |
| 15:36 | jgraham | remembers back in 2000 or so when Mozilla got a lot of "your artwork promotes Stalinism" flames |
| 15:49 | <hsivonen> | annevk: I reordered the WebM page |
| 15:49 | <hsivonen> | jgraham: struck "rival" |
| 15:52 | <hsivonen> | annevk: simplified the browser list for users who aren't already running one of the browsers listed |
| 15:53 | <hsivonen> | a non-crappy color scheme would be a most welcome contribution |
| 15:54 | <hsivonen> | maybe there should be a distinct box around the content |
| 15:54 | <hsivonen> | and a more proper test clip is still needed |
| 15:54 | <hsivonen> | jgraham: do you have suggestions on what to do about fonts |
| 15:55 | <jgraham> | hsivonen: Get a contribution from someone design minded who comes up with the non-crappy colour scheme |
| 15:55 | <jgraham> | You probably need to make a good video first though |
| 15:55 | <jgraham> | Since the video colours will need to work with the site colours |
| 15:56 | <hsivonen> | jgraham: I was thinking of making the video color depend on the site design |
| 15:56 | hsivonen | was unable to find any kind of WebM logo usage policy |
| 15:59 | <jgraham> | hsivonen: Well that would work too |
| 15:59 | <hsivonen> | annevk: AFAICT, the default Mozilla policy doesn't approve of making download links using the icon (as opposed to using a set of ready-made banners) |
| 16:00 | <hsivonen> | and I *really* don't want to make a jumble of browser banners |
| 16:14 | <hsivonen> | YSoD: http://media.opera.com/media/images/icon/opera_icon.svgz |
| 16:18 | <jgraham> | Hmm, works in Opera |
| 16:18 | <jgraham> | Presumably a problem with svgz served as image/svg+xml |
| 16:21 | <Lachy> | hsivonen, jgraham, there's no Content-Encoding header sent with that image. That would explain why Firefox gives a YSoD. But I can't explain Opera's behaviour. |
| 16:22 | <annevk> | hsivonen, thanks |
| 16:23 | <gsnedders> | Lachy: http://www.w3.org/TR/SVG/intro.html#MIMEType states the MIME type is image/svg+xml for SVG files, making no distrinction between gzip-compressed and non-gzip-compressed files (as is made below) |
| 16:23 | <hsivonen> | gsnedders: I believe that's one of the reasons why registering the type took so long |
| 16:25 | <annevk> | I hope they are not recommending file extensions over HTTP |
| 16:25 | <gsnedders> | hsivonen: It is indeed |
| 16:29 | <Lachy> | gsnedders, surely they don't mean gzipped files can be sent over HTTP without declaring as such in the headers?! |
| 16:29 | <gsnedders> | Lachy: That is the intention, I believe. |
| 16:29 | <Lachy> | WTF? |
| 16:30 | <Lachy> | I don't see where it states that in the spec though, nor where you're implying that from |
| 16:33 | <annevk> | hsivonen, could post on blog.whatwg.org asking for designer input on http://webm.html5.org/ |
| 16:34 | <hsivonen> | annevk: I guess that's an option. gotta run now, though |
| 16:34 | <hsivonen> | (I got rid of the bg color, since I suck at choosing one) |
| 16:34 | <hsivonen> | got rid of @font-face, too |
| 16:35 | <mpilgrim> | ooh, webm is part of html5 now? |
| 16:37 | <annevk> | mpilgrim, you could read it as WebM for HTML5 video :) |
| 16:59 | <jwalden> | hsivonen: where are the OS/2 and BeOS suggestions? |
| 16:59 | <jwalden> | </troll> |
| 16:59 | <jwalden> | also: that's a pretty impressive bit of enumeration there |
| 17:00 | <jwalden> | "Apple’s policies effectively block other browser engines from the platform" fightin' words |
| 17:02 | <jwalden> | can't be sure just from source-reading, does that do anything special for non-mobile Safari right now, or is it just hitting the OS X generic fallback case? |
| 17:10 | <jwalden> | also suggest can i use webm dot com perhaps (riffing on caniuse.com) |
| 17:21 | <a-ja> | TabAtkins: ping (re css3-images / webkit implementation. suggest somewhere else to discuss?) |
| 17:25 | <TabAtkins> | a-ja: Either here, or in a private message. |
| 17:45 | <hsivonen> | jwalden: I believe what I wrote about iOS policies is accurate, but in the interest of the non-partisan design pinciple, I'll change it to the generic mobile thing |
| 17:45 | <hsivonen> | jwalden: OS/2, BeOS, Solaris, AIX, FreeBSD, etc. are pending contributions from some who knows the situation with those platforms properly |
| 17:46 | <hsivonen> | jwalden: Safari on OS X hit the generic Mac OS X case |
| 17:46 | <hsivonen> | *hits |
| 17:46 | <TabAtkins> | jwalden: I forget - did you have a question for me? |
| 17:47 | <hsivonen> | jwalden: once there's a QuickTime component, I'll point to it |
| 17:47 | <hsivonen> | same for IE9 |
| 19:05 | <zcorpan> | hsivonen: you could conclude success when you get the 'playing' event |
| 19:25 | <Hixie> | hober: you have got to love plh's response to you about community support |
| 19:25 | <Hixie> | hober: "it was developed with community support! meaning, please support it, even though we didn't do any of its development in public" |
| 19:26 | <Hixie> | how can the "communication team" of the consortium that builds itself as "leading the web" be so utterly out of touch with the web |
| 19:26 | <Hixie> | it baffles my mind |
| 19:27 | <karlcow> | Hixie, you are living in your own world too. |
| 19:29 | <jwalden> | TabAtkins: aside from one offhand reference in css3-images that seemed to address an orthogonal point, I couldn't find anything that said whether gradients have intrinsic image ratios or not -- do they? or is that not spec'd yet? |
| 19:38 | <zcorpan> | what's with the brokenness of SSE that http://html5doctor.com/methods-of-communication/ talks about? |
| 19:39 | <Rik`> | zcorpan: last time I checked, only Opera and WebKit had implemented it |
| 19:39 | <Rik`> | and it wasn't really working the same way |
| 19:39 | <zcorpan> | opera 11 should be spec compliant |
| 19:40 | <miketaylr> | hmm http://node.remysharp.com:8001/sse-client.html seems down |
| 19:40 | <zcorpan> | it sounds like the server is closing the connection and the browser reconnects, or something |
| 19:41 | <zcorpan> | where's annevk when you need him |
| 19:42 | <Hixie> | so apparently browsers don't implement a11y by mapping elements to roles and then roles to native aapis |
| 19:42 | <Hixie> | they just map elements to native aapis except if roles are present in which case they let the roles override the default mapping |
| 19:45 | <miketaylr> | zcorpan: yes, in this demo the server is closing connections http://googlecodesamples.com/html5/sse/sse.html |
| 19:45 | <miketaylr> | (linked from http://www.html5rocks.com/tutorials/eventsource/basics/) |
| 19:46 | <othermaciej> | Hixie: WebKit has a role concept internally |
| 19:47 | <othermaciej> | Hixie: some directly correspond to ARIA roles, others do not |
| 19:47 | <othermaciej> | in general, we assign a role to every element, whether or not it has a role attribute |
| 19:47 | <othermaciej> | this is because A11Y APIs expect their own concept of role as one of the pieces of data about a UI element |
| 19:47 | <othermaciej> | we map our internal ARIA-like roles to the respective A11Y API's role |
| 19:48 | <othermaciej> | Hixie: so at least as to WebKit, I don't think your description of how browsers do things is totally accurate |
| 19:51 | <annevk> | why did I not know |
| 19:51 | <annevk> | http://html5homi.es/ |
| 19:54 | <zcorpan> | ah there he is |
| 19:54 | <karlcow> | annevk: yet another case of wwic ;) |
| 19:54 | <zcorpan> | annevk: see logs |
| 19:54 | <zcorpan> | annevk: do we have a bug or is it just that the server closes the connection? |
| 19:55 | <Hixie> | othermaciej: interesting |
| 19:55 | <Hixie> | othermaciej: does that mean webkit doesn't find steven's table helpful? |
| 19:55 | <annevk> | wwic? |
| 19:55 | annevk | goes to read logs |
| 19:56 | <Hixie> | othermaciej: specifically, see http://lists.w3.org/Archives/Public/public-html-a11y/2011Jan/0159.html |
| 19:56 | <Hixie> | annevk: for you, see http://lists.w3.org/Archives/Public/public-html-a11y/2011Jan/0173.html |
| 19:57 | <annevk> | zcorpan, the client should reconnect |
| 19:57 | <annevk> | zcorpan, unless the server specifically kills it with a 4xx response |
| 19:58 | <karlcow> | wwic - "Why Wasn't I Consulted?" http://www.ftrain.com/wwic.html |
| 19:58 | <annevk> | zcorpan, http://www.whatwg.org/specs/web-apps/current-work/complete/comms.html#processing-model-5 |
| 19:58 | <othermaciej> | Hixie: in its current form where it is 90% ?, definitely not |
| 19:58 | <othermaciej> | interesting, the table is too wide to fit in my browser window but I don't get a scrollbar |
| 19:58 | <annevk> | zcorpan, oops, anything but 2xx kills the connection |
| 19:58 | <zcorpan> | annevk: rem says browsers are broken because it keeps reconnecting instead of keeping the connection open. but then what he's seeing is that the server is closing the connection and the browser is reconnecting |
| 19:59 | <Hixie> | othermaciej: well obviously the quality of the table leaves a lot to be desired, but i just meant in general |
| 19:59 | <annevk> | zcorpan, what is his script on the server? |
| 19:59 | <Hixie> | othermaciej: (btw, this table is part of what steven is proposaing in issue-129 to add to the spec, though you have to read his CP and proposed edits pretty carefully to realise that) |
| 20:00 | <othermaciej> | Hixie: well, to actually make use of it, we'd have to add a level of indirection, since we have a cross-platform concept of role that we use before mapping to the API's native role |
| 20:00 | <zcorpan> | annevk: article points to http://node.remysharp.com:8001/sse-client.html but that seems down from here |
| 20:01 | <annevk> | zcorpan, it very much sounds like he closes the connection indeed |
| 20:01 | <othermaciej> | Hixie: I'd have to ask our accessibility experts to determine whether it is useful in general |
| 20:01 | <Hixie> | othermaciej: k |
| 20:01 | <othermaciej> | Hixie: in cases where an element maps to an ARIA role in an obvious way, and the ARIA role's mapping to system APIs is defined, it's definitely not useful |
| 20:01 | <annevk> | zcorpan, but unfortunate that he did not check with us first, especially given that he found all impls "broken" -- i.e. likely indicator of a problem on his end |
| 20:01 | <annevk> | s/but/bit/ |
| 20:02 | <othermaciej> | but there are some elements with AX behavior that can't be expressed purely in ARIA terms |
| 20:02 | <Hixie> | othermaciej: well the table in general just repeats what the HTML spec and the ARIAIMPL specs say, it's not original in any way as far as i'm aware |
| 20:03 | <othermaciej> | Hixie: if it said something for elements with no defined ARIA role, that might be useful |
| 20:03 | <annevk> | Hixie, he's certainly entitled to his opinion -- but Chaals and I are known to disagree every now and then :) |
| 20:03 | <Hixie> | othermaciej: fair enough |
| 20:08 | <heycam> | re the logos, I like the the Offline & Storage one has clouds as a background, even though it's basically the opposite of cloud storage |
| 20:08 | <heycam> | *that the |
| 20:08 | <heycam> | also Device Access seems like it's in a firestorm |
| 20:10 | <TabAtkins> | jwalden: gradients don't have an intrinsic aspect ratio. I say they have "no intrinsic dimensions", and I'm counting aspect-ratio as a dimension. |
| 20:10 | <jwalden> | ah |
| 20:10 | <othermaciej> | Hixie: replied to the email |
| 20:11 | <jwalden> | TabAtkins: it would seem to me best to be very explicit about that, given that (as I recall) the SVG spec didn't talk about this stuff in quite the same way |
| 20:12 | <annevk> | submitted a comment to html5doctor zcorpan |
| 20:14 | <annevk> | and I fixed my obvious typo in a follow up comment awaiting moderation |
| 20:14 | <TabAtkins> | jwalden: Okay, I'll change the statement to be explicit. |
| 20:15 | <jwalden> | thanks muchly |
| 20:16 | <annevk> | TabAtkins, pm regarding CSS meeting |
| 20:17 | <zcorpan> | annevk: oh, i added a comment too. seems you beat me to it |
| 20:18 | <annevk> | hopefully people read the comments |
| 20:19 | <zcorpan> | annevk: you aware of a demo that keeps the connection alive? |
| 20:19 | <annevk> | not infinite |
| 20:20 | <annevk> | zcorpan, but e.g. http://tc.labs.opera.com/apis/EventSource/format-field-event.htm sends two events |
| 20:23 | <annevk> | I guess I could look into parsing of Content-Type tomorrow |
| 20:23 | <annevk> | well, I'd need some kind of server that is not Apache |
| 20:24 | <annevk> | I wonder if I still have that Python script which just allows you to write some bytes to a socket |
| 20:24 | <annevk> | I really dislike that Apache has no fine-tuned HTTP control |
| 20:25 | <zcorpan> | mod_asis isn't good enough? |
| 20:25 | <annevk> | maybe it is for Content-Type, but in general Apache still fiddles with things |
| 20:26 | <zcorpan> | i thought asis would escape all fiddling, though i haven't tested it myself |
| 20:26 | <annevk> | the Apache on DreamHost still outputs some crazy header to aid an old Netscape browser |
| 20:27 | <annevk> | zcorpan, no, you still get HTTP/1.1 and some Date stuff if I remember correctly |
| 20:27 | <zcorpan> | ah. boo |
| 20:28 | <zcorpan> | annevk: as it happens pywebsocket allows you to write out whatever bytes you want |
| 20:29 | <annevk> | "X-Pad: avoid browser bug" |
| 20:29 | <annevk> | is that header |
| 20:29 | <annevk> | crazy |
| 20:29 | <zcorpan> | lol |
| 20:30 | <annevk> | yeah, chrisk wrote a small Python script as well once that makes it really easy to serve raw bytes |
| 20:30 | <annevk> | and I still have it in my inbox :) |
| 20:31 | <annevk> | I should probably put that on bitbucket.org when I play with it again |
| 20:31 | <TabAtkins> | Yes please! |
| 20:40 | <annevk> | If anyone can find the commercial described here and upload it to YouTube or some such it would be much appreciated: http://tbar.vernon18.org/cgi-bin/yabb/YaBB.pl?board=General;action=display;num=1035662382 |
| 20:40 | <annevk> | That thread is one of the only two references I can find via Google |
| 20:41 | <annevk> | I think I only saw it twice and it was long time ago, but I remember laughing a lot |
| 20:48 | <zcorpan> | annevk: try bing? |
| 20:50 | <annevk> | seems worse |
| 20:53 | <karlcow> | duck something |
| 20:54 | <karlcow> | http://duckduckgo.com/ |
| 20:54 | <annevk> | maybe they have it in Hilversum somewhere |
| 20:54 | <annevk> | in the media archives |
| 20:54 | <annevk> | but it seems somewhat unlikely |
| 20:57 | <zcorpan> | hmm, haven't we discussed http://www.w3.org/Bugs/Public/show_bug.cgi?id=11794 before? |
| 20:58 | <Hixie> | probably |
| 20:58 | <Hixie> | what was our solution? |
| 21:00 | <zcorpan> | http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2010-August/027624.html |
| 21:01 | <annevk> | making hidden= bar validation makes sense though |
| 21:01 | <Hixie> | well there's no use case there... we have one now |
| 21:01 | <Hixie> | that changes matters :-) |
| 21:02 | <zcorpan> | Hixie: can you elaborate on the use case? |
| 21:02 | <Hixie> | a check box that hides or shows a bunch of controls that are required if shown |
| 21:02 | <zcorpan> | <fieldset disabled> is no good? |
| 21:03 | <Hixie> | it's in the context of a nested list, i believe |
| 21:03 | <Hixie> | so fieldset would be a bit awkward |
| 21:03 | <Hixie> | that was my first response too though |
| 21:03 | <Hixie> | (this was discussed in moznet #developers just now) |
| 21:04 | <zcorpan> | maybe disabled should be a global attribute |
| 21:04 | <TabAtkins> | Hm, that would be interesting. |
| 21:05 | <Hixie> | interesting idea |
| 21:05 | <Hixie> | bit radical! |
| 21:05 | <TabAtkins> | Radical? Seems straightforward and simple. |
| 21:06 | <TabAtkins> | It works better than hidden, because then you can still do the "gray out disabled sections" thing. |
| 21:07 | <Hixie> | radical only in the sense that it would mean changes to :disabled, to constraint validation, to the form controls processing models, etc |
| 21:07 | <Hixie> | not that it's a bad idea |
| 21:07 | <TabAtkins> | Ah, kk. Yeah, there's a lot of fiddly bits attached to it. But it's a clean idea. |
| 21:08 | <Hixie> | probably won't be popular with bz |
| 21:08 | <Hixie> | he doesn't like moving functionality up to HTMLElement, on perf grounds |
| 21:09 | <annevk> | I think it matches IE |
| 21:09 | <Hixie> | matching IE perf is a non-goal :-P |
| 21:09 | <annevk> | I discovered and mentioned that back when I played with contenteditable |
| 21:09 | <Hixie> | oh? |
| 21:09 | <annevk> | heh |
| 21:10 | <annevk> | http://annevankesteren.nl/projects/whatwg/spec.html |
| 21:10 | <annevk> | "Web Applications 1.0 — contentEditable and disabled" |
| 21:10 | <annevk> | long time ago that was :) |
| 21:10 | <zcorpan> | it greys stuff in ie, but doesn't seem to prevent input in <input> |
| 21:10 | <annevk> | bbl |
| 21:11 | <Hixie> | did i know about this document? |
| 21:11 | <zcorpan> | but it disables submit buttons and form submission |
| 21:14 | <zcorpan> | disabled also disables contenteditable in ie |
| 21:14 | <zcorpan> | Hixie: you made him write it iirc :) |
| 22:13 | <asmodai> | lol @ http://i.imgur.com/6U9fU.jpg |
| 22:15 | <jacobolus> | asmodai: it's alive! |
| 22:15 | <asmodai> | Where? Who? What? |
| 22:16 | <jacobolus> | HTML5 |
| 22:16 | <asmodai> | Yes it is. :) |
| 22:16 | <jacobolus> | I'm imagining him (her?) sitting with his pipe, looking condescendingly down at poor Flash |
| 22:16 | <asmodai> | "And what have you done for me lately? |
| 22:16 | <asmodai> | " |
| 22:17 | <zcorpan> | so where's teh flash version of the logo? |
| 22:25 | <Hixie> | oh hey, adactio's here |
| 22:25 | <adactio> | howdy |
| 22:25 | <Hixie> | just got your e-mail |
| 22:26 | <Hixie> | anyone object to changing the spec's name to just "HTML"? |
| 22:26 | <adactio> | Not I. |
| 22:26 | <adactio> | And I hereby apologise for freaking out about last time the change was made. |
| 22:26 | <adactio> | I was wrong. |
| 22:26 | <Hixie> | oh no worries |
| 22:27 | <Hixie> | i think you may have been right that delaying for a bit was ok |
| 22:27 | <Hixie> | it didn't cause any harm, anyway |
| 22:27 | <Hixie> | i'll mention it every few hours to get people from various time zones, but assuming nobody complains, i'll rename the spec tomorrow or so |
| 22:28 | <adactio> | Excellent. |
| 22:28 | <Hixie> | and then post a blog post or something |
| 22:28 | <TabAtkins> | I OBJECT, SIR. |
| 22:28 | <Hixie> | and then put this image into the spec in the "is this html5" section :-) |
| 22:29 | <Hixie> | TabAtkins: it's ok, you can still rename your spec the "HTML5 CSS3 Advanced Layout" spec :-P |
| 22:30 | <TabAtkins> | Oh, well that's all right then. |
| 22:30 | TabAtkins | is calling all of his specs that, actually. |
| 22:30 | <TabAtkins> | Flexbox is "HTML5 CSS3 Advanced Layout Chapter 1", Lists is "HTML5 CSS3 Advanced Layout Chapter 2", etc. |
| 22:31 | <Hixie> | heh |
| 22:32 | <Hixie> | not "HTML5 CSS3 Web 2.0 DHTML Advanced RESTful Layout"? |
| 22:32 | <Hixie> | oh oh and AJAX! |
| 22:32 | <TabAtkins> | REST is a fad, silly. |
| 22:32 | <Hixie> | i forgot AJAX! |
| 22:33 | <TabAtkins> | I should publish my "HTML5 CSS3 Advanced RDFa 2 (for Web 2.0)" spec. |
| 22:33 | <Hixie> | i recommend "HTML5 CSS3 Web 2.0 DHTML Advanced RESTful AJAX Layout Recommendation 1.0 Accessibility Guidelines Module" |
| 22:34 | <Hixie> | though you can probably drop the "Layout" part |
| 22:35 | <Rik`> | could you also please use a construction animated gif on the specs until they go to CR ? :) |
| 22:35 | <Hixie> | seems reasonable |
| 22:35 | <TabAtkins> | Yay! We can stop talking about CSS2.1 testing at the telcons for a few weeks! |
| 22:35 | <Hixie> | you still attend the telecons? |
| 22:35 | <Hixie> | man |
| 22:36 | <TabAtkins> | Decisions are made there. I have to ensure they're not bad decisions. |
| 22:36 | <crash\> | I think "WebM" is missing :) |
| 22:36 | <Hixie> | TabAtkins: dude the csswg is the wg where it is the easiest to revert decisions |
| 22:37 | <TabAtkins> | True, but still, better to not have to revert it at all. Plus, I like talking things out. |
| 22:37 | <crash\> | you can replace DHTML with it, since it's a pretty outdated buzzword |
| 22:37 | <Hixie> | TabAtkins: oh i'm not saying you shouldn't attend them :-) |
| 22:37 | <Hixie> | TabAtkins: it took me a like five years to burn out from them |
| 22:39 | <TabAtkins> | I've only been in the group a little over a year, dude. |
| 22:41 | <Hixie> | yeah but i have high hopes for you! you're far better than i am, you should burn out much faster than me. ;-) |
| 22:41 | <TabAtkins> | That's, um. Thanks? |
| 22:44 | <Hixie> | :-) |
| 22:45 | <Hixie> | i guess "burn out" is probably the wrong term |
| 22:45 | <Hixie> | "get disillusioned" maybe? |
| 22:45 | <Hixie> | i dunno |
| 22:45 | <Hixie> | i was young and naive, took me a while to get over that :-) |
| 22:46 | <Hixie> | hober: when do you start at apple btw? |
| 22:46 | <Hixie> | how is the latest editor's draft older than the latest public draft http://www.w3.org/2010/02/rdfa/drafts/#rdfa-core |
| 22:47 | <Hixie> | that makes no sense |
| 22:47 | <TabAtkins> | I guess they're using the generation date? |
| 22:48 | <TabAtkins> | Which should mean that the public and editor's draft are identical right now. |
| 22:49 | <Hixie> | i guess |
| 22:50 | <Hixie> | does that mean they're no longer maintaining it? or is it perfect...? |
| 22:52 | <TabAtkins> | It means it hasn't been touched for 3 months. |
| 22:53 | <Hixie> | oh i misread the date |
| 22:53 | <Hixie> | my bad |
| 22:53 | <Hixie> | i thought it was early last year |
| 22:57 | <Rik`> | TabAtkins: about image(), if I want to use webp today (cause it offers better compression) I'd like to do it without having an extra HTTP request |
| 22:58 | <Rik`> | cause with an extra HTTP request, it kinds of the defeat the purpose of using a lighter image format |
| 22:59 | <Rik`> | also, if the feature was already implemented, it would greatly simplify using SVG files on the web right now |
| 23:00 | <TabAtkins> | Rik`: I agree that dealing with the format stuff is a useful feature. I don't know if it's worth complicating up the syntax that much. |
| 23:01 | <TabAtkins> | Like, you'll get "background-image: image( "foo.webp" "image/webp", "foo.svg" "image/svg+xml", blue );" |
| 23:01 | <TabAtkins> | Plus maybe that weird resolution-handling stuff. |
| 23:01 | <Rik`> | why image/webp and not only webp ? |
| 23:01 | <TabAtkins> | Argh, I don't like the resolution stuff either. Again, useful functionality, but I don't think it fits in the image() function. |
| 23:02 | <TabAtkins> | I assumed it would be by mimetype? |
| 23:02 | <Rik`> | font-face doesn't use mimetype |
| 23:03 | <TabAtkins> | Yeah, it defines a small list of types. |
| 23:03 | <Rik`> | but I could live with mimetype |
| 23:03 | <TabAtkins> | Actually... going with that, and inverting the functionality (only download the resource is the type isn't specified, or if it is specified and it's something you think can render) might be more useful. |
| 23:04 | <TabAtkins> | That way new formats like webp automatically work as expected - they get ignored until they're supported. |
| 23:06 | <karlcow> | http://www.ietf.org/rfc/rfc2295.txt never implemented I think :/ |
| 23:06 | <karlcow> | Transparent Content Negotiation in HTTP |
| 23:07 | <karlcow> | was proposing an Alternates: header |
| 23:07 | <karlcow> | giving all the type of resources available |
