| 00:41 | <Hixie> | oooh, chaals praised the existence of something i asked for |
| 00:51 | <Hixie> | first hit for "moma html opera": "New York Private MoMA and Opera Backstage Tour" |
| 00:51 | <Hixie> | not what i meant... |
| 00:51 | <gavin_> | isn't it mama? |
| 00:51 | <Hixie> | yes |
| 00:51 | <Hixie> | but usually google fixes that kind of thing for me! :-) |
| 01:33 | <deltab> | Google seems to be 'correcting' more aggressively now: for instance, try searching for elementmoveto |
| 01:36 | <deltab> | and I think it should correct hashchanged to hashchange, not "has changed" :-) |
| 02:02 | <Hixie> | holy crap zcorpan, you're amazing |
| 02:02 | <Hixie> | all these status markers! |
| 02:03 | <Hixie> | i wonder if we should ban <caption> when the table is in a <figure> |
| 02:18 | <kinetik> | Hixie: Is the video element intended to resume playback after seeking completes if it was playing back prior to the seek? |
| 02:18 | <kinetik> | I think that's what you want, but I'm having trouble convincing myself that the spec says that's what should happen. |
| 02:18 | <Hixie> | yes |
| 02:18 | <Hixie> | the video doesn't stop "potentially playing" while seeking |
| 02:19 | <Hixie> | er |
| 02:19 | <Hixie> | that's a lie |
| 02:19 | <Hixie> | let me rephrase |
| 02:19 | <Hixie> | the video is defined to play back while it is "potentially playing", which becomes true again once there is data to play, if paused is false |
| 02:19 | <Hixie> | http://www.whatwg.org/specs/web-apps/current-work/#potentially-playing |
| 02:20 | <Hixie> | (the spec says "When a media element is potentially playing and its Document is an active document, its current playback position must increase monotonically at playbackRate units of media time per unit time of wall clock time.") |
| 02:21 | <kinetik> | Hixie: ah, I see, thanks |
| 02:29 | <Hixie> | hsivonen: it would be interesting for the validator to report tables that, once you've applied the CSS, have no borders. |
| 02:38 | gsnedders | wonders how many of the statuses zcorpan didn't touch |
| 02:52 | <Hixie> | sam wants me to deal with profile="" next, but i've no idea what i am supposed to do |
| 02:52 | <Hixie> | so i shall go eat instead |
| 02:54 | <gsnedders> | Hixie: Comfort eating, again? :P |
| 03:14 | gsnedders | shakes head at the size of the to and cc lists of the latest email to public-html |
| 03:45 | <gsnedders> | Would I be right to say that <link xmlns="http://www.w3.org/2004/Atom"; href="foobar"> is non-conforming because the @href isn't an IRI? |
| 03:46 | <gsnedders> | Actually, the spec is probably referring to IRI-reference by "IRI reference" |
| 05:24 | <gsnedders> | http://stuff.gsnedders.com/atom-export.html |
| 05:41 | <annevk> | The Live DOM Viewer does not do JavaScript? |
| 06:00 | <annevk> | gsnedders, if we're going with xref rather than title, please just name it data-xref="" |
| 06:00 | <annevk> | unlikely specs will need another xref |
| 06:00 | <gsnedders> | annevk: Blame Hixie. He came up with data-anolis-xref :P |
| 06:00 | <annevk> | ok |
| 06:00 | annevk | blames Hixie |
| 06:00 | <annevk> | will you change it now? |
| 06:00 | <gsnedders> | Send email, and I'll change it sometime |
| 06:01 | <gsnedders> | (before 1.1) |
| 07:10 | <annevk> | "accessibility theater" I like it |
| 07:49 | <sayrer> | annevk: interoperation with validators? honestly! |
| 07:52 | <hsivonen> | "interoperation with validators"? |
| 07:52 | <annevk> | sayrer, glad I listed some others :) |
| 07:52 | <sayrer> | those were bad too! |
| 07:52 | <sayrer> | :) |
| 07:53 | hsivonen | wonders what is being discussed |
| 07:53 | <annevk> | www-archive |
| 07:53 | <sayrer> | hsivonen, http://lists.w3.org/Archives/Public/www-archive/2009Feb/0108.html |
| 07:53 | <sayrer> | and anne's reply |
| 07:53 | <hsivonen> | thanks |
| 07:53 | <gsnedders> | Is it bad that all the random poetry I have in my ringbinder for school is in the computing section? |
| 07:54 | <hsivonen> | sayrer: requiring authors not to use reserved element or attribute names is relevant to future interop, though |
| 07:54 | <sayrer> | maybe |
| 07:55 | <sayrer> | that is a very small part of the requirements placed on documents, aiui |
| 07:56 | <annevk> | the same goes for possible syntax extensions, etc. |
| 07:56 | <hsivonen> | sayrer: also, if one wants validators to be substitutable, then there's an uniformness (aka. interop in W3C speak) issue between validators |
| 07:57 | <sayrer> | hsivonen, I agree that validators have to "agree" on the content of the spec. are you saying something different? |
| 07:57 | <hsivonen> | sayrer: I'm saying that if validators have to agree on the validation target, then there's an interop-sensitive thing going between validators |
| 07:58 | <hsivonen> | which isn't exactly relevant to interop between browsers |
| 07:58 | <sayrer> | hsivonen, and thus the requirements placed in the document now are ok? |
| 07:58 | <hsivonen> | sayrer: many are ok, in my opinion. there are some that I disagree with. |
| 07:59 | <sayrer> | (I totally don't agree with that, but I want to understand) |
| 07:59 | <sayrer> | hsivonen, I mean the division highlighted by the note I quoted |
| 08:00 | <sayrer> | that rationale would seem to make the requirement in RFC2119 meaningless |
| 08:00 | <hsivonen> | sayrer: I agree that there should be no implied relationship as far as the spec reader does the implication. I think the spec writer needs to keep the relationship sane. |
| 08:01 | <sayrer> | I don't understand what you agree with or what you think, no malice intended |
| 08:01 | <hsivonen> | sayrer: we are still discussing the passage quoted in http://lists.w3.org/Archives/Public/www-archive/2009Feb/0108.html right? |
| 08:02 | <sayrer> | right |
| 08:02 | <sayrer> | my question is this |
| 08:02 | <sayrer> | how does the spec call a document non-conforming if it still interoperable? |
| 08:02 | <sayrer> | it's |
| 08:03 | <sayrer> | and it claims to use RFC2119 |
| 08:03 | <sayrer> | something has to give |
| 08:03 | <hsivonen> | sayrer: the RFC 2119 criteria for use of MUST includes another thing |
| 08:03 | <hsivonen> | "to limit behavior which has potential for causing harm |
| 08:03 | <hsivonen> | " |
| 08:03 | <hsivonen> | the MUST that apply to authors are supposed to limit "harm" |
| 08:04 | <hsivonen> | sayrer: I'm sure people disagree on what constitutes "harm" |
| 08:04 | <sayrer> | ok, where is the potential for harm? |
| 08:04 | <sayrer> | given a conformant HTML5 user agent |
| 08:04 | <sayrer> | user agents that don't conform to HTML5 are likely to do terrible things |
| 08:05 | <sayrer> | even with documents that match all the requirements that are in there today |
| 08:05 | <hsivonen> | sayrer: the authoring "harm" harms users or authors who maintain what has been authored |
| 08:05 | <sayrer> | hsivonen, I don't see a tenable interpretation here. I could be wrong. |
| 08:06 | <hsivonen> | sayrer: example: making @profile non-conforming protects authors from the harm of the epic waste of time that including no-op syntax is |
| 08:06 | <hsivonen> | sayrer: example: making <font> non-conforming protects users from the harm of media-dependent design |
| 08:06 | <hsivonen> | sayrer: I fully expect you to disagree that there is 'harm' |
| 08:06 | <sayrer> | well, reasonable people can disagree |
| 08:06 | <hsivonen> | sayrer: I disagree with some assumed 'harms' myself |
| 08:07 | <sayrer> | hsivonen, the default wordpress template includes @profile. My blog got that for free. :) |
| 08:07 | <hsivonen> | sayrer: do you find the entire concept of document conformance flawed? |
| 08:08 | <hsivonen> | sayrer: it costs your host in bandwidth bills :-) |
| 08:08 | <sayrer> | so would lots of features in HTML5 |
| 08:09 | <hsivonen> | sayrer: if you add something that the profile doesn't cover as far as angels on a head of a pin are concerned, dealing with people emailing you about your incorrect profile would cost you in terms of the annoyance of dealing with the feedback |
| 08:10 | <hsivonen> | sayrer: if you hadn't gotten the wordpress profile for free, would adding the 'right' profile values be a useful activity for you? |
| 08:10 | <sayrer> | beats me |
| 08:10 | <sayrer> | (honestly) |
| 08:10 | <sayrer> | though I do doubt it |
| 08:11 | <sayrer> | hsivonen, I don't find the concept of validators flawed. they help people fix mistakes they didn't know about. |
| 08:11 | <hsivonen> | sayrer: if you wouldn't add a profile yourself but might use a microformat without profile, would you consider massive permathreads about the right profile URIs for use with those formats an epic waste of time for humanity? |
| 08:12 | <sayrer> | I might use a microformat without knowing I am using a microformat, because they squat short strings |
| 08:12 | <sayrer> | not that there is anything wrong with that |
| 08:13 | <sayrer> | and, yes, I would consider a massive permathread an epic waste of time |
| 08:13 | <hsivonen> | sayrer: well, there's 'harm' from profile |
| 08:14 | <sayrer> | hsivonen, you mean from spec mailing lists? some people don't want relief from the pain. |
| 08:15 | <hsivonen> | sayrer: well, discussing profile is a huge time sink for the HTML WG. However, if the HTML WG doesn't get rid of profile, over the years profile will sink much more human effort. |
| 08:15 | <sayrer> | maybe. less than em vs. b? |
| 08:16 | <sayrer> | less than "semantic" whatever? |
| 08:16 | <sayrer> | highly questionable |
| 08:16 | <hsivonen> | sayrer: discussing em vs. i is a total waste of time, too. |
| 08:16 | <sayrer> | er |
| 08:16 | <sayrer> | yeah :) |
| 08:16 | <hsivonen> | the right solution would have been to never introduce em, strong, var, and dfn |
| 08:17 | <hsivonen> | but now that both em and i are in use, taking away either would be too disruptive |
| 08:17 | <sayrer> | yeah, and it doesn't matter a whole lot |
| 08:17 | <hsivonen> | however, we might just get away with zapping profile still |
| 08:17 | <sayrer> | would you consider that a great victory? |
| 08:17 | <hsivonen> | no |
| 08:18 | <sayrer> | a small one? |
| 08:18 | <sayrer> | a victory against what? |
| 08:18 | <sayrer> | I honestly have no idea |
| 08:18 | <sayrer> | against wasted bits? I have several vendors you may wish to contact |
| 08:18 | <annevk> | it simplifies the language |
| 08:19 | <annevk> | you consider the Dreamweaver thing a serious implementation? |
| 08:19 | <sayrer> | really? I thought no one used it |
| 08:19 | <annevk> | well, some people do, and apparently a whole bunch of people thought it actually did something, including the guys from WordPress |
| 08:19 | <sayrer> | sure, I consider Mozilla a serious implementation as well. we do useless things. |
| 08:19 | <sayrer> | annevk, they did it because tantek said to. |
| 08:20 | <annevk> | no, I meant what Dreamweaver did for profile="" |
| 08:20 | <hsivonen> | sayrer: a small victory against permathreads |
| 08:20 | <hsivonen> | sayrer: huh? Mozilla does something with profile? |
| 08:21 | <sayrer> | hsivonen, I meant that Mozilla does useless things, but I still consider it a serious implementation |
| 08:21 | <sayrer> | since it mostly does non-useless things |
| 08:21 | <hsivonen> | sayrer: but those useless things aren't relevant to profile? |
| 08:22 | <sayrer> | hsivonen, annevk asked a question about whether I considered Dreamweaver a serious implementation in relation to @profile. I think the two issues are orthogonal. |
| 08:22 | <annevk> | no, I meant if you considered what Dreamweaver did for profile="" to be serious |
| 08:23 | <hsivonen> | sayrer: fwiw, when the microformats community instructs people to use profile, I think they violate the Priority of Constituencies principle by putting the political convenience of specifiers ahead of the convenience of implementors and authors |
| 08:23 | <sayrer> | annevk, sorry. and no, not at all. |
| 08:23 | <annevk> | because it looks like they treat it as an attribute that takes a URL without providing any UI that makes sense for the feature |
| 08:23 | <hsivonen> | it's not their principle, though |
| 08:23 | <annevk> | (besides the fact that profile="" takes a space-separated list of URLs; one might argue Dreamweaver is buggy) |
| 08:24 | <hsivonen> | yay, QA cost of profile |
| 08:24 | <sayrer> | hsivonen, I think arguments that run for the charter or design principles are only useful when something is really off the rails. like some XHTML stuff. |
| 08:24 | <sayrer> | I don't see how there is a QA cost if it doesn't do anything |
| 08:25 | <sayrer> | there is a real cost to <o:p> |
| 08:25 | <hsivonen> | sayrer: there's the cost of checking that you can enter the right kind of talismans |
| 08:25 | <hsivonen> | in general, I disagree with the argument that no-op talismans have no cost. |
| 08:25 | <sayrer> | ah |
| 08:26 | <hsivonen> | every piece of spec text has a cost |
| 08:26 | <sayrer> | yes |
| 08:26 | <sayrer> | do you buy the backward compatibility argument? |
| 08:26 | <hsivonen> | even if it is only the cost of people making the informed decision that the text is silly and needs to implmentation action |
| 08:26 | <hsivonen> | sayrer: which argument? |
| 08:26 | <sayrer> | that HTML5 must be backward compatible? |
| 08:27 | <annevk> | backwards compatible with what? |
| 08:27 | <sayrer> | with content that uses @profile, like my blog |
| 08:27 | <hsivonen> | sayrer: I buy it in the sense that a consumer agent implementing HTML5 should not continue to work with existing Web content |
| 08:28 | <sayrer> | should continue to work, you mean? |
| 08:28 | <annevk> | we're already incompatible with your blog |
| 08:28 | <annevk> | e.g. with the DOCTYPE |
| 08:28 | <hsivonen> | sayrer: well, if agents now do nothing with profile, backward compat requires them not to do any thing with it |
| 08:29 | <sayrer> | annevk. user agents would change their parsing of my blog? |
| 08:29 | <sayrer> | hsivonen, is @profile reserved by Ian's HTML5 document? |
| 08:30 | <annevk> | sayrer, no |
| 08:30 | <hsivonen> | sayrer: yes, it is reserved |
| 08:30 | <sayrer> | hsivonen, is the hope that we can some day unreserve it? |
| 08:30 | <hsivonen> | sayrer: no |
| 08:30 | <sayrer> | then I am a little lost |
| 08:30 | <hsivonen> | sayrer: it is forever tainted |
| 08:30 | <sayrer> | you say forever tainted, I say forever "defined" |
| 08:31 | <hsivonen> | sayrer: how would you define it? |
| 08:31 | <sayrer> | similarly to RSS2 <textinput> |
| 08:31 | <annevk> | a change to the DOCTYPE on your blog would change the layout model |
| 08:32 | <hsivonen> | sayrer: sorry, my knowledge of RSS2 details is so rusty that I don't get the analogy |
| 08:32 | <sayrer> | "The purpose of the <textInput> element is something of a mystery. You can use it to specify a search engine box. Or to allow a reader to provide feedback. Most aggregators ignore it." |
| 08:32 | <sayrer> | http://cyber.law.harvard.edu/rss/rss.html#lttextinputgtSubelementOfLtchannelgt |
| 08:33 | <annevk> | anyway, my point was that todays content is not compatible with HTML5 requirements on content and whether or not profile= is conforming is orthogonal to that unless we change other things first |
| 08:33 | <sayrer> | annevk, I don't get your point at all |
| 08:33 | <annevk> | ok, not sure how else to put it |
| 08:33 | <sayrer> | this entire argument is about whether requirements on content matter |
| 08:33 | <sayrer> | or at least how much they matter |
| 08:34 | <annevk> | well, I asked "backwards compatible with what" |
| 08:34 | <annevk> | you said with your content |
| 08:34 | <hsivonen> | sayrer: so do you think "The purpose of profile is something of a mystery." would meet the level of spec quality we are after with HTML5? |
| 08:34 | <annevk> | I said that your content was incompatible regardless |
| 08:34 | <sayrer> | I see, sloppy |
| 08:34 | <sayrer> | backwards compatible with current parsing of my content |
| 08:34 | <annevk> | I don't see what that has to do with @profile |
| 08:35 | <sayrer> | hsivonen, it is a true statment, so we have that at least |
| 08:35 | <annevk> | and no, there's no parsing differences between limited quirks mode and no quirks mode |
| 08:35 | <hsivonen> | sayrer: it doesn't say what authors or implementors need to do with it, so it's not a very useful true statement |
| 08:36 | <sayrer> | it says it is a mystery |
| 08:36 | <sayrer> | and user agents might ignore it |
| 08:36 | <hsivonen> | sayrer: btw, would you consider getting profile into HTML5 a victory, and if so, against what? |
| 08:36 | <sayrer> | that would be useful to me |
| 08:37 | <hsivonen> | sayrer: how? |
| 08:37 | <sayrer> | hsivonen, a victory against declaring most extant Wordpress blogs invalid |
| 08:37 | <hsivonen> | sayrer: I see |
| 08:37 | <annevk> | sayrer, they are invalid regardless of @profile! |
| 08:37 | <hsivonen> | sayrer: HTML5 does declare them invalid for other reasons, though |
| 08:37 | <sayrer> | ok, let them be invalid for better reasons |
| 08:38 | <sayrer> | better reasons than /> and @profile |
| 08:38 | <hsivonen> | sayrer: would it address your concern if HTML5 allowed profile with the exact value "http://gmpg.org/xfn/11";? |
| 08:39 | <sayrer> | that seems weird, and doesn't increase interoperation vs. letting it be. |
| 08:40 | <hsivonen> | fwiw, I think http://www.gmpg.org/xmdp/ might have served a political purpose some time, but I don't see it serving a technical purpose today |
| 08:40 | <sayrer> | but if someone showed that was all that was needed |
| 08:41 | <sayrer> | hsivonen, I think it was a political excuse to squat short strings |
| 08:41 | <sayrer> | (again, I don't care about squatting) |
| 08:41 | <hsivonen> | aside: regarding the cost of profile, when it was in HTML5, developing validation for it required a datatype that wasn't used anywhere else |
| 08:41 | <hsivonen> | sayrer: right |
| 08:42 | <hsivonen> | however, since *I* already have that code, profile validation is now sunk cost for me |
| 08:42 | <sayrer> | hmm, I'm not sure I would validate an attribute that doesn't do anything |
| 08:42 | <sayrer> | ah, excellent compromise! |
| 08:43 | <sayrer> | couldn't validators alert users to the fact that they don't validate the value of @profile because it doesn't do anything? |
| 08:43 | <annevk> | should be the same data type as <a ping> |
| 08:43 | <hsivonen> | sayrer: I do that with RDF. It's even more code, but it's possible, sure. |
| 08:43 | <sayrer> | <a ping> accepts arbitrary numbers of URIs |
| 08:43 | <sayrer> | hahhaahaha |
| 08:43 | <sayrer> | yeah, that's going to work great! |
| 08:44 | <hsivonen> | annevk: didn't ping accept relative URL? |
| 08:44 | <hsivonen> | sayrer: oops. the message doesn't say that RDF doesn't do anything. |
| 08:44 | <sayrer> | hsivonen, should I file a bug? ;) |
| 08:45 | <hsivonen> | sayrer: no ;-) |
| 08:46 | <sayrer> | hmm, it occurs to me that using @profile was quite dishonest of the microformats people |
| 08:46 | <sayrer> | they claimed namespaces were overengineering |
| 08:46 | <sayrer> | fair |
| 08:46 | <sayrer> | but then they used one that has no effects |
| 08:47 | <hsivonen> | is it engineering if it has no effects? :-) |
| 08:47 | <sayrer> | might have been better to just say "we're squatting, so what" |
| 08:47 | <hsivonen> | yes |
| 08:47 | <sayrer> | hsivonen, well, one of their central claims concerns unobservable metadata |
| 09:01 | <zcorpan_> | Hixie: function showLoginForm(network if (network.username && network.password) { |
| 09:01 | <zcorpan_> | http://www.whatwg.org/specs/web-apps/current-work/status.js |
| 09:01 | <zcorpan_> | doesn't look right |
| 09:03 | <Hixie> | wip |
| 09:07 | <sayrer> | Hixie, btw there are two kinds of bofs at the ietf |
| 09:07 | <sayrer> | one is formal, on the agenda |
| 09:07 | <sayrer> | the other is not |
| 09:07 | <sayrer> | my understanding is that the html5 thing is the latter |
| 09:07 | <sayrer> | avoiding all the fees etc of the ietf |
| 09:08 | <zcorpan_> | Hixie: while you're working on that, it would be nice if the edit popup said which 'id' it was bound to |
| 09:08 | <annevk> | good |
| 09:09 | <Hixie> | yeah i'm going to the html5 one |
| 09:09 | <sayrer> | the IETF process is pretty much the process of avoiding process |
| 09:09 | <Hixie> | i was talking about the web socket thing at the bof itself |
| 09:09 | <Hixie> | s/bof/official meeting/ |
| 09:09 | <sayrer> | oh, the HTTP meeting? |
| 09:09 | <annevk> | Hixie, oh, that's separate? |
| 09:09 | <Hixie> | zcorpan_: k |
| 09:09 | <Hixie> | you've both now exceeded the bounds of my knowledge |
| 09:10 | <sayrer> | they want to discuss web socket behind the paywall, so to speak? |
| 09:10 | <Hixie> | apparently |
| 09:10 | <Hixie> | you were cc'ed on the e-mail mentioning this |
| 09:10 | <sayrer> | some of them |
| 09:10 | <annevk> | how high is this wall in dollars? |
| 09:10 | <sayrer> | whatever the meeting fee is |
| 09:11 | hsivonen | didn't know the IETF was pay-to-play |
| 09:11 | <sayrer> | $700 US or so last time I went |
| 09:11 | <Hixie> | sayrer: i'm talking about the e-mail that says "Ian, are you attending the IETF officially" |
| 09:11 | <Hixie> | hsivonen: it's double pay-to-play -- some company sponsors the event, and then attendees pay some exhorbitant fee |
| 09:11 | <Hixie> | i don't mind company sponsorship |
| 09:12 | <Hixie> | i do mind restricting participation to those who can pay |
| 09:12 | <sayrer> | there are also ghost fees |
| 09:12 | <sayrer> | like group rates in the hotel hosting |
| 09:12 | <sayrer> | it is insidious |
| 09:12 | <Hixie> | i understand the ietf has rising costs and so forth, and respect that they have to make choices and so forth |
| 09:12 | <sayrer> | it is a little better than the w3c, though |
| 09:13 | <annevk> | 700 US? geez |
| 09:13 | <Hixie> | the w3c varis on a per-wg basis |
| 09:13 | <Hixie> | varies |
| 09:13 | <Hixie> | e.g. the htmlwg meeting is free (excluding personal costs) |
| 09:13 | <sayrer> | I was talking about membership |
| 09:14 | <sayrer> | something the IETF doesn't have |
| 09:14 | <Lachy> | Hixie, the solution to the table summary problem is this: |
| 09:14 | <Lachy> | <p id="summary">Useful summary for everyone...</p><table aria-describedby="summary">...</table> |
| 09:15 | <annevk> | I hope not |
| 09:15 | annevk | hates writing IDREFs |
| 09:15 | <sayrer> | Hixie, at any rate, I would be willing to discuss Web Sockets. I sincerely think it is a great idea. |
| 09:16 | <sayrer> | I might not be the advocate you are looking for, though |
| 09:16 | <Lachy> | or <figure><legend>summary</legend> <table>...</table></figure> |
| 09:16 | <sayrer> | I will probably defend it no matter what |
| 09:16 | <sayrer> | so maybe it doesn't matter |
| 09:16 | <Lachy> | or just don't use any explicit association, and rely on their proximity and context |
| 09:17 | <annevk> | I like it too, but I'm not gonna get Opera to pay 700 US just to say that; that's insane |
| 09:18 | <sayrer> | I contribute to the HTTP specs too |
| 09:18 | <sayrer> | and the HTTP group doesn't meet at every meeting |
| 09:19 | <annevk> | it seems weird for it to be part of the HTTP WG though since it would be out of charter for them |
| 09:19 | <sayrer> | it's on port 80 |
| 09:19 | <annevk> | and the IETF is apparently very peculiar about those |
| 09:19 | <annevk> | sayrer, it's on any port |
| 09:20 | <annevk> | sayrer, defaulting to 81 actually, iirc |
| 09:20 | <sayrer> | oh really |
| 09:20 | <sayrer> | isn't that the same as 80 for some legacy reason? |
| 09:20 | <sayrer> | I forget |
| 09:20 | <annevk> | no, you can do it over 80 if you want |
| 09:20 | <sayrer> | no, I mean 81 is a synonym for 80 |
| 09:20 | <sayrer> | just proxy busting |
| 09:20 | <sayrer> | iirc |
| 09:21 | <annevk> | ah, that could be true |
| 09:21 | <Hixie> | sayrer: for most groups membership is free if you can convince someone to invite you -- with the htmlwg, the main group whose creation i was involved in, anyone can join |
| 09:21 | <Hixie> | sayrer: (wg membership at ietf is free too, as i understand it) |
| 09:21 | <Hixie> | sayrer: (and meetings are optional once you've started the group) |
| 09:21 | <annevk> | neh, http://www.opera.com:81/ doesn't work |
| 09:21 | <Hixie> | sayrer: if you're going, feel free to talk about it :-) |
| 09:22 | <Hixie> | 81 isn't registered currently |
| 09:22 | <sayrer> | Hixie, maybe no one needs to go. I can already hear the objections. |
| 09:22 | <Hixie> | (though i'll register it if we go ahead with websocket) |
| 09:24 | <sayrer> | gosh, which port did I get recommended for big downloads |
| 09:24 | <annevk> | it seems even harder to convince the IETF of browser related issues than the W3C |
| 09:24 | <zcorpan_> | http://www.isolani.co.uk/blog/standards/TheIe8BlacklistMinefield - as expected ie7 mode in ie8 doesn't match ie7 |
| 09:24 | <sayrer> | so busted caching proxies wouldn't break it |
| 09:24 | <sayrer> | I swear it was 81 |
| 09:24 | <annevk> | mostly because the IETF is even more convinced we're doing it all wrong :) |
| 09:24 | <Hixie> | sayrer: might well have been |
| 09:25 | <sayrer> | if it was only that one site, no big deal |
| 09:25 | <zcorpan_> | oh wait ie8 has two "ie7" modes? |
| 09:25 | <sayrer> | I thought it was proxy lingua franca. bad assumption. |
| 09:26 | <sayrer> | annevk, it turns out the IETF people are receptive to the reality argument, tho |
| 09:34 | zcorpan_ | doesn't follow isofarro's reasoning |
| 09:34 | <annevk> | sayrer, not last time I tried to argue over something with the HTTP WG |
| 09:34 | <annevk> | Content-Location |
| 09:35 | <sayrer> | hmm, doesn't sound reality based? |
| 09:35 | Hixie | considers using a <canvas> the height of the HTML5 document |
| 09:35 | <annevk> | or URLs |
| 09:35 | <annevk> | sayrer, check bugzilla |
| 09:35 | <sayrer> | which |
| 09:35 | <sayrer> | URLs they are gonna lose |
| 09:35 | <annevk> | the one from Mozilla |
| 09:35 | <zcorpan_> | Hixie: are you trying to make it crash more browsers? |
| 09:35 | <sayrer> | annevk, which bug? |
| 09:36 | <annevk> | several, bz was involved |
| 09:36 | annevk | doesn't know numbers |
| 09:36 | <sayrer> | zcorpan_: interoperability is good |
| 09:36 | <sayrer> | annevk, bugs on Content-Location? |
| 09:36 | <sayrer> | with real websites? |
| 09:36 | <sayrer> | surprising |
| 09:36 | <annevk> | yeah, real websites broke because of Mozilla supporting it |
| 09:36 | <Hixie> | zcorpan_: :-P |
| 09:36 | <sayrer> | delicious |
| 09:37 | <annevk> | for some time bz even created whitelists/blacklists |
| 09:37 | <annevk> | I believe at some point he concluded the spec didn't make sense either, but this has been a while |
| 09:38 | <annevk> | most of my interactions with the IETF fail for some reason or another |
| 09:38 | <roc> | zcorpan: hmm, an even more impressive train-wreck than I thought |
| 09:38 | <annevk> | could just be me of course, but elsewhere similar interactions work out much better |
| 09:39 | <Hixie> | my interactions with actual ietf people (primarily lisa) have been very productive and friendly |
| 09:40 | <sayrer> | my feeling is that the IETF desperately wants to be relevant |
| 09:40 | <roc> | zcorpan: but I don't think "Compatibility View" is different from "IE7 standards mode" |
| 09:40 | <sayrer> | and they don't have a coordination requirement like the W3C does |
| 09:41 | <sayrer> | but they do have fuzzy consensus processes |
| 09:41 | <sayrer> | that defend against process hacking, but also transparency |
| 09:42 | <sayrer> | also a chickenshit patent policy |
| 09:42 | <Hixie> | personally my experience hasn't been that the problem is with the processes themselves, since as far as i can tell nobody really follows them whatever the standards organisation |
| 09:42 | <sayrer> | I have seen you rail against W3C process violations |
| 09:43 | <Hixie> | oh i'm not saying that violating documented process is a good thing |
| 09:43 | <Hixie> | i'm just saying that i don't think that's the real problem |
| 09:43 | <sayrer> | IETF processes are written such that only loonies go off like that |
| 09:43 | <Hixie> | the main problem that i've found afflict people in all standards organisations is a reliance on expertise rather than logical argument and research |
| 09:43 | <Hixie> | and i've seen that everywhere |
| 09:44 | <Hixie> | ietf, ecma, w3c, iso, you name it |
| 09:44 | <sayrer> | I don't really like "logical argument", but I do like "research" |
| 09:44 | <Hixie> | yeah, research certainly is the more important of the two |
| 09:44 | <Hixie> | a lot of people in the spec world will just go with "well i think it should be foo" and that's the end of their argument |
| 09:44 | <sayrer> | rationalization can usually pass for logical argument |
| 09:44 | <Hixie> | they never say "because..." something |
| 09:45 | <Hixie> | or they say "because i know best", or "because i've had 20 years' experience", or some other argument from authority |
| 09:45 | <sayrer> | I've seen counterexamples |
| 09:46 | <Hixie> | how do you mean? |
| 09:46 | <Hixie> | there are certainly many people doing it the right way |
| 09:46 | <Philip`> | Hixie: I think Opera limits canvases to 32000x32000 pixels, and I'd guess the HTML5 spec is taller than that, and it would be nice if the spec still actually worked in most browsers |
| 09:46 | <Hixie> | with research, etc |
| 09:46 | <Hixie> | Philip`: don't worry, i didn't seriously consider it |
| 09:46 | <Philip`> | and I think Firefox (2, maybe 3?) has a coordinate system that runs out at +/- 32K or something |
| 09:47 | <Philip`> | and of course you'd be incompatible with IE :-( |
| 09:47 | <sayrer> | Hixie, I mean that the value propositions for each party are obvious for people who've been around the block once |
| 09:48 | <Hixie> | value propositions? |
| 09:48 | <sayrer> | Hixie, I am thinking of the classic "firewalls will block it" |
| 09:49 | <sayrer> | turns out they probably will |
| 09:49 | <sayrer> | so people who want to standardize something have a burden of proof that kind of sucks |
| 09:50 | <sayrer> | but they might be prone to argue for evidence that the default is true |
| 09:51 | <hsivonen> | the wish to tweak Referer instead of minting a new header seems more principled than evidence-based, though. at least to a casual observer. |
| 09:51 | <sayrer> | yeah, I agree |
| 09:51 | <sayrer> | the belief that Origin won't be blocked is a different subject |
| 09:55 | <annevk> | whether or not it's blocked, it's near impossible to change right now |
| 09:55 | <sayrer> | Origin? |
| 09:55 | <annevk> | yeah |
| 09:56 | <sayrer> | why? |
| 09:56 | <annevk> | because in a few months it will be shipping in Safari 4, Firefox 3.1, and Internet Explorer 8 |
| 09:56 | <jgraham> | FWIW it's still not clear to me that using <caption><details> or <figure><details> doesn't meet the WCAG-recommended use case for <table @summary>. However some control over the styling would be needed for authors to get it right. |
| 09:57 | <sayrer> | annevk, which other header will ship in all of those? |
| 09:57 | <sayrer> | starts with "Ref" |
| 09:57 | Philip` | guesses that if authors can't trivially work out how to style it 'right' (for whatever their definition of 'right' is, which might include being invisible) then they'll not bother including any summary text anywhere at all |
| 09:57 | <annevk> | servers doing anything with the style of cross-origin requests those implementations allow will use Origin |
| 09:58 | <jgraham> | Philip`: Which would, roughly speaking, be the status quo :) |
| 09:58 | <sayrer> | annevk, I agree. What value does origin deliver to proxies. |
| 09:59 | <annevk> | is that question? (I don't know the answer.) |
| 10:00 | <sayrer> | annevk, yes a question. sorry for the poor punctuation. |
| 10:00 | <sayrer> | annevk, the answer is: nothing. |
| 10:00 | <sayrer> | so it goes in the same circular file |
| 10:00 | <sayrer> | since it does have a risk |
| 10:01 | <Philip`> | Does a lack of complaining users count as value? |
| 10:01 | <sayrer> | no |
| 10:01 | <sayrer> | you have obviously never sat behind an energetic proxy |
| 10:01 | <annevk> | I heard that most proxies are not updated often |
| 10:02 | <annevk> | but ok |
| 10:02 | <sayrer> | not such a good argument |
| 10:02 | <annevk> | it's not an argument |
| 10:02 | <sayrer> | not such a good mitigating factor? |
| 10:02 | <annevk> | I don't really care either way |
| 10:02 | <sayrer> | what matters is how many users sit behind a proxy updated for Origin |
| 10:03 | <sayrer> | or how many of a given site's users sit behind that proxy |
| 10:03 | <sayrer> | tough! |
| 10:03 | <annevk> | yeah |
| 10:04 | <Philip`> | sayrer: Do you mean a proxy updated to allow Origin, or a proxy updated to block Origin? |
| 10:05 | <sayrer> | proxies generally allow unknown headers in my experience, and abarth had some data that showed that too |
| 10:05 | <sayrer> | he misapplied that data to claim that they would continue to allow Origin |
| 10:05 | <sayrer> | but quibble quibble |
| 10:06 | <Philip`> | "In our experiment, the X-Requested-By header is correctly delivered to servers approximately 99.90–99.99% of the time" |
| 10:07 | Philip` | supposes it would also be nice to know whether X-* headers are handled any differently to non-prefixed unknown headers |
| 10:07 | <sayrer> | I tried arbitrary headers on 4 different US cell networks once |
| 10:07 | <sayrer> | they worked |
| 10:08 | <sayrer> | but at least two of them were using Squid |
| 10:08 | <sayrer> | and that was years ago nwo |
| 10:08 | <sayrer> | now |
| 10:09 | <roc> | I think Origin has a chance if we can deploy applications that use it faster than people can deploy proxies that strip it |
| 10:09 | <sayrer> | roc, Origin doesn't necessarily need to disclose the origin domain though |
| 10:09 | <sayrer> | unless you want to whitelist domains regardless of hierarchy |
| 10:10 | <sayrer> | it depends whether the goal is to allow cross site requests or prevent CSRF |
| 10:11 | <annevk> | my main use case is the former |
| 10:12 | <sayrer> | I've seen both prioritized |
| 10:12 | <annevk> | at some point it was hijacked to also do the latter; I'm ambivalent to that |
| 10:12 | <sayrer> | enough to question whether they need to be the same header |
| 10:12 | <sayrer> | actually |
| 10:12 | <annevk> | see public-webapps archive |
| 10:12 | <Philip`> | roc: That deployment approach would require applications that fail without the header (e.g. use it for allowing cross site requests), because if it merely prevents CSRF then nobody will notice when proxies start stripping it and so proxies would have no reason to stop stripping it |
| 10:13 | <sayrer> | I suppose it's ok to try it for both at first |
| 10:13 | <Philip`> | but the former case doesn't seem to be backward-compatible enough for anyone to rely on it for the next several years |
| 10:13 | <sayrer> | then when Origin fails, a sensible URL-proximity header can be minted |
| 10:13 | <sayrer> | the no stop energy approach |
| 10:14 | <roc> | Philip`: true about CSRF, false about CORS ... applications might have optional features that require CORS |
| 10:15 | <Philip`> | roc: Ah, I suppose it would work better if it was used as a bonus rather than as a core requirement |
| 10:16 | <sayrer> | "it doesn't work from my computer" |
| 10:16 | <sayrer> | "ok, I'll proxy the request" |
| 10:16 | <sayrer> | problem solved |
| 10:16 | <sayrer> | just saying |
| 10:17 | <sayrer> | I could be wrong, unless some website operator is willing to fall on their sword for CORS |
| 10:17 | <annevk> | i guess we'll see, right? |
| 10:17 | <sayrer> | er, I could be wrong if |
| 10:17 | <sayrer> | annevk, exactly. That's why the no stop energy appproach is best. |
| 10:20 | <annevk> | URL proximity does not really allow for the case where you only want to share with specific URLs |
| 10:20 | <annevk> | that are on completely separate domains |
| 10:24 | <annevk> | hmm, how does Referer work in face of redirects? |
| 10:28 | <annevk> | I don't think it does, but the HTTP spec seems very vague |
| 10:29 | <hsivonen> | hashing origin wouldn't solve the issue of a cross-site widget provider seeing what public sites a user reads but it would address the issue of leaking sensitive intranet host names like wiki.supersentiviteproduct.example.com |
| 10:32 | <Philip`> | But it'd be quite vulnerable to dictionary attacks, because domain names rarely have much entropy |
| 10:32 | <Hixie> | not only that, but typically you're looking for a yes/no answer against a small set of domains |
| 10:34 | <hsivonen> | I can see how a dictionary attack could tell who has been using a cross-site widget at www.playboy.com, but how would one use a dictionary attack to figure out iphonekiller.corp.google.com? |
| 10:35 | <Philip`> | By having a dictionary that contains terms like "iphone", "killer" and "corp", then trying all the combinations of terms and dots until you find one with the right hash |
| 10:35 | <hsivonen> | ok |
| 10:37 | <beowulf> | FWIW I think "Tables must not be used as layout aids" is too strong, tables layout data afterall |
| 10:37 | <sayrer> | Hixie/annevk, URl proximity does not allow an arbitrary whitelist of domains. it does allow stanford.facebook.com vs. www.facebook.com |
| 10:38 | <sayrer> | it seems like a reasonable trade-off for CSRF defense, and maybe useless for cross-origin request stuff |
| 10:38 | <annevk> | yeah, though if Origin works it won't be needed |
| 10:39 | <sayrer> | but that's why I say they don't need to be the same header, and also that I don't care if it is tried |
| 10:39 | <annevk> | i.e. if proxies are just scared of the path |
| 10:39 | <sayrer> | plausible, but I think they are scared of preventable breaches |
| 10:40 | <sayrer> | time will tell |
| 10:40 | Philip` | wonders if Opera has stats on how many people enable its disable-referrer-sending feature |
| 10:41 | hsivonen | wonders if intranet hosts are actually named iphonekiller or if they use kewl code names |
| 10:41 | <sayrer> | there are lots of interesting variations on the theme |
| 10:42 | <sayrer> | like scanning for hostname patterns that reveal a particular vulnerability might be present |
| 10:43 | <sayrer> | ooh, how about a wordpress vulnerability that made a cross-site request |
| 10:43 | <sayrer> | delicious |
| 10:49 | <Philip`> | http://www.google.com/search?q=referer+%22%2A.corp.google.com |
| 10:51 | <sayrer> | I once found that a googler was mooching my wifi that way :) |
| 10:51 | <sayrer> | close quarters here in nyc |
| 10:53 | <sayrer> | I think the person was a sysadmin too. access to lots of juicy sounding host names |
| 10:53 | <sayrer> | :) |
| 11:05 | Philip` | wonders if someone broke Gmail |
| 11:06 | <annevk> | http://twitter.com/mikeyk/statuses/1244442800 |
| 11:47 | <Hixie> | zcorpan_: ok, added the id |
| 11:47 | <Hixie> | to the little boxes |
| 11:51 | <Philip`> | Argh, now even the multipage version is incredibly slow to load in Opera |
| 11:51 | <hsivonen> | Hixie: HTML5 has changed <!doctype html></body><title>X</title> to parse in a way that's different from Gecko, Safari and Opera |
| 11:51 | <hsivonen> | Hixie: is that intentional? |
| 11:51 | <Hixie> | oh? |
| 11:51 | <Hixie> | oh |
| 11:51 | <Hixie> | hm |
| 11:51 | <Hixie> | i guess </body> shouldn't autoclose before head huh |
| 11:52 | <hsivonen> | seem so |
| 11:52 | <hsivonen> | +s |
| 11:52 | <Hixie> | do you know which diff that was? |
| 11:52 | <Hixie> | actually if you could file a bug that'd be great |
| 11:53 | <Philip`> | Hixie: There still seems to be an off-by-one error in the month of the dates in the status annotations |
| 11:53 | <Hixie> | i don't want to be half-way through this edit when i leave here, which could happen any moment |
| 11:53 | <hsivonen> | Hixie: I'll try to figure that out and file a bug |
| 11:53 | <Philip`> | e.g. all zcorpan's annotations are marked as 2009-01-23 |
| 11:53 | <Hixie> | hsivonen: thanks, that's awesome |
| 11:53 | <Hixie> | Philip`: oops |
| 11:54 | <Hixie> | try now? |
| 11:54 | <Philip`> | Hixie: Looks better now |
| 11:55 | <Hixie> | cool |
| 11:56 | <Philip`> | Hixie: I quite dislike how it uses 100% CPU when I view any spec page in Opera |
| 11:56 | <Hixie> | if you can work out why it does it, i'm open to improving the code |
| 11:57 | <Philip`> | Hixie: Hmm, I don't care enough to actually debug it |
| 11:57 | <Hixie> | heh |
| 11:58 | Philip` | adds http://www.whatwg.org/*/status.js into Opera's blocked content list |
| 12:27 | <hsivonen> | http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!doctype%20html%3E%3Chtml%3E%3C%2Fhtml%3E%3C!--%20foo%20--%3E is interesting in WebKit |
| 12:36 | <zcorpan_> | hsivonen: why is it interesting? |
| 12:37 | <hsivonen> | zcorpan_: there's no head node |
| 12:38 | <zcorpan_> | hsivonen: there's never a head in webkit unless there's a <head> tag or some element that implies <head> |
| 12:39 | <hsivonen> | ah |
| 12:48 | <Hixie> | anyone have a good 16x16 icon that could represent JS libraries / shims / plugins? |
| 12:48 | <Hixie> | oooh |
| 12:48 | <Hixie> | wikipedia plugin icon |
| 12:49 | <zcorpan_> | Hixie: the spec seems to freeze firefox upon scrolling, too |
| 12:51 | Philip` | wonders if Hixie is getting confused about the differences between specification documents and Acid tests |
| 12:53 | <zcorpan_> | Acid5 is the html5 spec |
| 12:53 | <Hixie> | heh |
| 12:53 | <jgraham> | Does the spec load at all now? I froze two browsers this morning tring to load it. This did not make me happy. |
| 12:54 | <jgraham> | Where by "load at all" I mean "load in a sensible amount of time i.e. < 60 seconds" |
| 12:54 | <Philip`> | I can use the multipage version in Firefox 2 with no problems |
| 12:54 | <jgraham> | The multipage version sucks on account of the multipageness |
| 12:54 | <Philip`> | Not sure about what other possibilities work |
| 12:55 | <Philip`> | jgraham: On the other hand, at least it doesn't suffer from singlepageness |
| 12:56 | <zcorpan_> | jgraham: it works in safari |
| 12:57 | <jgraham> | zcorpan_: If you could only convince Apple to release Safari for Linux that would be a little more helpful :) |
| 12:58 | jgraham | could use epiphany-webkit but that kind of sucks |
| 12:58 | <Philip`> | jgraham: Use Wine |
| 12:58 | zcorpan_ | thinks Hixie does a "if (!safari) { while(true); }" |
| 12:58 | <Hixie> | i really can't work out what is causing firefox's slowness during scrolling |
| 12:58 | <Hixie> | unless it's the offsetTop calls |
| 12:58 | <Hixie> | is offsetTop expensive? |
| 12:59 | <Philip`> | Is it moving stuff and causing a re-layout? |
| 12:59 | <Hixie> | no, only position:fixed stuff is changing |
| 12:59 | <Hixie> | at least in theory |
| 12:59 | <Philip`> | In theory, the code works perfectly with no slowdown at all |
| 13:00 | <Philip`> | so I wouldn't put too much faith in theory |
| 13:04 | <Hixie> | looks like firefox's problem is offsetTop is ass-slow |
| 13:06 | jgraham | onders how many ways there are to tell the difference between 0 and -0 in ES3 |
| 13:06 | <jgraham> | *wonders |
| 13:06 | <Philip`> | Infinite |
| 13:07 | <jgraham> | Philip`: That doesn't seem like a vey helpful answer |
| 13:07 | <Philip`> | function is_minus_zero(n) { return n == 0 && 1/n < 0 && ...an infinite number of expressions that evaluate to true... } |
| 13:08 | <Philip`> | jgraham: Would it be more helpful to lie and give you any other answer? |
| 13:08 | <jgraham> | Philip`: Yes. |
| 13:09 | <zcorpan_> | jgraham: 42 |
| 13:09 | <Philip`> | jgraham: There's about seven, then |
| 13:09 | <jgraham> | I could try to make the problem more well defined, if you like |
| 13:09 | <jgraham> | But I would probablu leave some hole |
| 13:09 | <Philip`> | Since I have no idea what the problem is, that might be helpful |
| 13:10 | <jgraham> | What I actually mean is something like "howw many different functions are there such that f(0) != f(-0) in an observable way" |
| 13:10 | <jgraham> | So "division" would be one function |
| 13:11 | <Philip`> | I don't know of any ways that don't involve dividing a number by n and seeing if it's +Infinity or -Infinity |
| 13:11 | <jgraham> | Well atan2 does something different |
| 13:11 | <jgraham> | Though that is kindof division |
| 13:13 | <Philip`> | Ah, I forgot about atan2 |
| 13:13 | <hsivonen> | Hixie: do you have an opinion on implementing frameset-ok as an insertion mode? |
| 13:17 | <Hixie> | how does it interact with other modes that currently say "act as if it was in the in-body mode"? |
| 13:17 | <hsivonen> | Hixie: they'd presumably go to the real in-body |
| 13:18 | <hsivonen> | Hixie: to get to any of the table or select stuff, you end up setting the flag to "not ok" first |
| 13:18 | <Hixie> | can any of those cases be triggered when frameset-ok is true, on elements that would make it false? |
| 13:19 | <Hixie> | if you are 100% sure that it is equivalent, then i'm fine with removing state variables |
| 13:19 | <Hixie> | but i wasn't sure when i wrote it |
| 13:20 | <hsivonen> | Hixie: as far as I can tell, any Process the token using the rules for the "in body" insertion mode. happens in cases where you must have set the flag to "not ok" |
| 13:21 | <Hixie> | and all of the cases that currently set the mode to in-body either always happen when frameset-ok or !framset-ok? |
| 13:21 | <Hixie> | i don't understand why gecko is so slow at this |
| 13:23 | <hsivonen> | Hixie: search for "Switch the insertion mode to "in body"." doesn't work for me |
| 13:24 | <hsivonen> | Hixie: but it seems to be that if you are switching towards "in body" and you aren't in one of the head states, frameset-ok must already be "not ok" |
| 13:24 | <Hixie> | i guess i'd have to make <style> and other things that switch to CDATA use the current state |
| 13:25 | <Hixie> | instead of assuming in-body... though that might already be the case |
| 13:25 | <hsivonen> | Hixie: why? they already have to set the original mode. and coming out of CDATA switches to the orginal mode |
| 13:26 | <Hixie> | do any of them explicitly set the original mode to in body instead of to whatever the current mode is? |
| 13:26 | hsivonen | tries to understand his own character foster parenting code |
| 13:26 | <Hixie> | i guess not, if they did they'd fail in table |
| 13:27 | <hsivonen> | Hixie: none of them set explicitly to "in body" |
| 13:27 | <Hixie> | k |
| 13:32 | <hsivonen> | hmm. I think my text foster parenting may look crazy because it has bogus dead code... |
| 13:32 | <hsivonen> | I wonder why |
| 13:37 | <Hixie> | man i hate c-like languages that don't have the full complement of assignment operators |
| 13:37 | <hsivonen> | Hixie: I think it would be reasonable to do what WebKit does and foster-parent whitespace that comes before non-whitespace in a run of text tokens |
| 13:39 | <annevk> | how would you do that spec-wise? |
| 13:39 | <Hixie> | ok if the people who were complaining of new perf issues could recheck their browsers and see if the perf issues are back to where they were before, that'd be good |
| 13:39 | <annevk> | i guess you'd not special case whitespace |
| 13:40 | <hsivonen> | annevk: making the text coalescing buffer exist in the spec |
| 13:40 | <Hixie> | hsivonen: that would make the case of <body> text...... [long pause in network traffic] ...more text </body> not render the text received before the pause |
| 13:40 | <hsivonen> | annevk: it's black-box obvious that WebKit has one |
| 13:40 | <hsivonen> | I've got one, too, for the non-SAX cases |
| 13:40 | <hsivonen> | Hixie: correct |
| 13:41 | <hsivonen> | Hixie: but in practice, Web pages don't tend to have huge text nodes |
| 13:41 | <Hixie> | text/plain files do |
| 13:41 | <hsivonen> | instead there's plenty of intervening markup |
| 13:42 | <Hixie> | seems easy enough to me to just cut the buffer in half where necessary to match the current spec |
| 13:42 | <jgraham> | Doesn't it only matter inside tables? |
| 13:42 | jgraham | isn't paying much attention |
| 13:43 | <hsivonen> | Hixie: text/plain files don't need to use the HTML5 tree builder |
| 13:44 | <hsivonen> | jgraham: I never append to text nodes that are in the tree |
| 13:45 | <Hixie> | hsivonen: they're specced to right now |
| 13:45 | <Hixie> | anyway |
| 13:45 | <Hixie> | if you want me to look at this, file a bug |
| 13:45 | <Hixie> | bed time now |
| 13:45 | <Hixie> | nn |
| 13:45 | hsivonen | looks |
| 13:45 | <hsivonen> | nn |
| 13:48 | <annevk> | nn Hixie! |
| 13:48 | <hsivonen> | sigh. http://www.whatwg.org/specs/web-apps/current-work/#read-text will indeed be a problem |
| 13:48 | annevk | is tired/jetlagged too |
| 13:49 | <hsivonen> | does interop require text/plain documents to have one huge text node instead of a lot of tiny ones? |
| 13:50 | <annevk> | it would be good if there was interop |
| 13:50 | <annevk> | and one large text node makes more sense... |
| 13:50 | <hsivonen> | I'm tempted to write a one-off custom tree builder for text/plain instead of hacking the HTML one do more tricks |
| 13:51 | <annevk> | but you'll have the same issue for HTML then if people insert large text nodes... |
| 13:52 | <hsivonen> | unless a top site does it, I think I can get away with incremental rendering of HTML not happening in the middle of a text node |
| 13:56 | <hsivonen> | I guess the easy way out of my current problem is to make every text node flush check for potential foster parenting... |
| 14:00 | <hsivonen> | now I need an effient way to manage table taint so that I don't need to search the stack per non-WS char |
| 14:01 | <hsivonen> | how annoying |
| 14:03 | <hsivonen> | well, the stack doesn't need to be searched very deep anyway... |
| 14:24 | <annevk> | rubys, maybe interoperation doesn't matter too much indeed; see also archives of this channel for a discussion regarding "harm" |
| 14:26 | <rubys> | search today? |
| 14:26 | <rubys> | I do believe that interop is important, vital even. |
| 14:30 | <annevk> | interop for content in terms of RFC 2119 |
| 14:30 | <annevk> | (i.e. conformance requirements) |
| 14:30 | <annevk> | might be yesterday; bit confused with timezones :) |
| 14:30 | <rubys> | almost done reading... |
| 14:32 | <Lachy> | Hixie, the new floating status box on the left is really annoying |
| 14:33 | <Lachy> | Hixie, having the links to the section from within each status box is very useful, though |
| 14:34 | <rubys> | re: "the harm of the epic waste of time that including no-op syntax", that most emphatically is NOT what RFC 2119 is about. |
| 14:34 | <rubys> | re: "making <font> non-conforming protects users from the harm of media-dependent design", that may rise to the SHOULD NOT level at most. |
| 14:35 | <zcorpan> | hsivonen: "Error: The list attribute of the input element must refer to a datalist element or to a select element." - the spec changed here a while ago, only datalist is allowed |
| 14:36 | <jgraham> | rubys: So do you agree with the position that, in the face of a specification with defined error handling, there should be no MUST-level document conformance critera? |
| 14:38 | rubys | thinking... |
| 14:38 | <Lachy> | jgraham, I don't, since that basically means authors can do whatever they want, which isn't really a useful goal to aim for |
| 14:39 | <rubys> | Lachy: it isn't as binary as that. SHOULDs still can exist, and Validators can still complain, for example. |
| 14:41 | <Lachy> | rubys, so are you saying, you would reduce a MUST-level conformance requirement like this to a SHOULD? "The datetime attribute, if present, must contain a valid date or time string that identifies the date or time being specified." |
| 14:41 | rubys | repeats: THINKING. :-) |
| 14:42 | <Lachy> | so, that would mean authors SHOULD use a valid date, but they don't have to if they have a good reason?? |
| 14:42 | Philip` | wonders why everyone writing specs claims to use RFC 2119 terminology, even though they're usually not using the terms in the way defined there |
| 14:42 | <Philip`> | (rather than defining their own terms that are more applicable to their particular spec) |
| 14:42 | <Lachy> | Philip`, specs that don't use them as defined there should be fixd |
| 14:42 | <Lachy> | *fixed |
| 14:42 | <jgraham> | Philip`: Examples? |
| 14:44 | <rubys> | Lachy: I'll consult with some of my IETF friends. Datetime is a wonderful example, thanks! |
| 14:44 | <rubys> | The one I plan to start with (Paul Hoffman) is west coast US, so not likely to be up yet. |
| 14:46 | annevk | votes for RFC2119-5 |
| 14:46 | <annevk> | weinig, congrats! |
| 14:46 | <Philip`> | As far as I can tell, RFC 2119 was intended to apply to implementors of protocols, hence talking about "vendors" and "interoperability" (which I presume is in the sense of two pieces of software communicating with each other) and "e.g. limiting retransmissions", and was never intended to apply to human authors of documents |
| 14:47 | <weinig> | annevk: ? |
| 14:47 | <annevk> | on shipping and all |
| 14:48 | <weinig> | ah, we shipped! |
| 14:48 | <annevk> | :p |
| 14:48 | <weinig> | I forgot :) |
| 14:48 | <annevk> | did cross-origin requests ship? |
| 14:48 | <jgraham> | (FWIW I tend to think that such a position would be putting theoretical purity ahead of the actual needs of authors. It is much easier to write a useful QA tool based on MUST level conditions than SHOULD and, although it is true that defined error handling may lead to interoperable browser behaviour, authors should not be expected to understand the details of the error handling and therefore should be warned whenver they encounter it) |
| 14:48 | <weinig> | annevk: yes |
| 14:48 | <annevk> | yay |
| 14:48 | annevk | changes the spec :p |
| 14:48 | <jgraham> | Also: what Philip` said |
| 14:48 | <weinig> | but it is just a beta |
| 14:48 | <Philip`> | Given a definition like "MAY: ... One vendor may choose to include the item because a particular marketplace requires it or because the vendor feels that it enhances the product while another vendor may omit the same item." it makes no sense to say "Authors MAY use some particular element or whatever" |
| 14:48 | <weinig> | so anything goes |
| 14:48 | <annevk> | i see |
| 14:48 | <zcorpan> | hsivonen: i think opera emits several text nodes for text/plain |
| 14:48 | <Philip`> | because authors aren't vendors, and documents aren't products, and they're not offering it to a marketplace |
| 14:49 | <annevk> | well, bits in the spec have changed, but nothing that should preclude it from shipping (unless it's really out of date) |
| 14:49 | annevk | has been working on mostly editorial issues |
| 14:49 | <Philip`> | so any use of RFC 2119 for documents seems contrary to its intention, which results in the terms being effectively undefined and nobody quite agrees on what they mean |
| 14:50 | <zcorpan> | hsivonen: in fact, i think we emit several text nodes in text/html, too |
| 14:50 | <rubys> | jgraham: it is all too easy to recast positions one disagrees with in terms of the "other side" pursuing "theoretical purity". In general, when I see people do that, I see us as only being a step or two away from Godwin's law. |
| 14:51 | <rubys> | 2119 has provided useful and pragmatic advice for quite some time, sullying it (like many have done with summary, for example) has pragmatic implications too |
| 14:53 | <jgraham> | rubys: I am emphatically not suggesting that RFC 2119 is not a useful document. I am suggesting that using it as the basis for arguing that we should have weaker document conformance constraints ignores the practical value of those constraints |
| 14:54 | <jgraham> | I am very intersted to hear your considered opinion once you have formulated it |
| 14:55 | <rubys> | A document with a BOM that is actually encoded as win-1252, or with a datetime of mañana are both clearly wrong, even if all software in the world handles such documents identially. What we need to find is the most useful way to express that. Others here might differ, but I think that a document that uses profile, summary, or property as originally intended and documented in other (prior or elsewhere) specs are considerably less "wrong" |
| 15:28 | <zcorpan> | cufón seems to miss the concept of canvas fallback |
| 15:29 | <zcorpan> | <h3><span class="cufon cufon-canvas" style="width: 68px; height: 24px"><canvas width="86" height="25" style="top: -1px; left: -1px"></canvas><span class="cufon-alt">Select </span></span>... |
| 15:29 | <rubys> | this will likely end up being a blog entry. The cases aren't quite so cut and dry. |
| 15:30 | <annevk> | it depends on who has a say over the vocabulary |
| 15:31 | <rubys> | annevk: was that @me? |
| 15:31 | <annevk> | yeah |
| 15:31 | <annevk> | sorry |
| 15:32 | <yecril71> | There must be a compelling use case for date/time intervals out there, |
| 15:32 | <yecril71> | otherwise database engines would not create a separate data type for them |
| 15:32 | <yecril71> | Although, of course, many small database engines ignore the issue |
| 15:33 | <yecril71> | Since HTML5 started supporting DATE and TIME, |
| 15:33 | <yecril71> | it seems logical that it would ultimately support intervals as well, |
| 15:33 | <yecril71> | for database compatibility, if for no other purpose. |
| 15:34 | <yecril71> | That was @Lachy. |
| 15:34 | <yecril71> | There are different forms out there, and their layout is different. |
| 15:35 | <Lachy> | yecril71, HTML5 isn't solving the same problems as databases, so the use cases are not necessarily the same |
| 15:35 | <yecril71> | But database access is a strong application of HTML. |
| 15:36 | <Lachy> | so what problem would be solved by marking up time intervals in the HTML, even if it were output from a database? |
| 15:36 | <yecril71> | There are bureaucratic forms that require the applicant to fill a host of disconnected entries; |
| 15:37 | <yecril71> | The resulting HTML could be easily processed afterwards, without losing information. |
| 15:37 | <Lachy> | yecril71, just because someone might use HTML to publish some particular data type, doesn't necessarily mean there needs to be dedicated markup for it. |
| 15:37 | <yecril71> | Because the data types would be preserved. |
| 15:37 | <Lachy> | processed in what way? For what purpose? |
| 15:37 | <zcorpan> | why is the summary discussion cc-ed to every list i'm subscribed to (and a few others)? |
| 15:38 | <zcorpan> | and directly to me no less |
| 15:38 | <Lachy> | zcorpan, because the accessibility advocates feel the need to CC every list whenever an accessibility-related issue is discussed |
| 15:38 | <jgraham> | How many copies of each mail do you get? |
| 15:38 | <yecril71> | The report could be processed by a HTML grabber to enter information to a local database. |
| 15:38 | <zcorpan> | Lachy: why, to make their emails more accessible? |
| 15:38 | <Lachy> | I don't know. |
| 15:39 | <Lachy> | It's just an observation. I haven't yet formed a theory to explain it. |
| 15:39 | <zcorpan> | jgraham: 4, i think |
| 15:39 | <yecril71> | I am suggesting that time intervals are not "some particular data type"; |
| 15:39 | <yecril71> | they are standardised and widely accepted and used. |
| 15:39 | <jgraham> | My hasiily constructed theory is that noone understands the tangle of accessibility related lists and so everyone just CCs them all in case one is relevant |
| 15:40 | <Philip`> | zcorpan: It's probably because Hixie selected all the recent summary mails and hit "reply all" and posted his response to sixteen addresses |
| 15:40 | <Lachy> | yecril71, vague use cases involving some unspecified "HTML grabber" to enter information into some local database with an undefined purpose doesn't make a useful use case or problem statement |
| 15:41 | <yecril71> | But it is likely that some such use case will eventually apppear. |
| 15:41 | <Philip`> | yecril71: Once it does appear, it can be addressed |
| 15:41 | <Lachy> | so if and when it does appear, then we can look at it |
| 15:41 | <Lachy> | and then we can address it. |
| 15:41 | <Lachy> | Trying to address a potential use case we haven't seen yet isn't likely to result in an adequate solution. |
| 15:42 | <yecril71> | Lachy! Since you asked about it, I thought you were interested� |
| 15:42 | <yecril71> | I am not applying or promoting anything, just discussing things. |
| 15:42 | <Lachy> | I'm interested in hearing any real use cases there may be. I'm not interested in discussing hypothetical use cases that aren't based in reality |
| 15:43 | <othermaciej> | new Safari beta, in case you folks haven't seen it |
| 15:43 | <Lachy> | othermaciej, link? |
| 15:43 | <Lachy> | is that Safari 4? |
| 15:44 | <Philip`> | othermaciej: How awesome is it, on a scale of 1 to 10? |
| 15:44 | <othermaciej> | Philip`: over 9000! |
| 15:44 | <zcorpan> | wow that's pretty awesome |
| 15:44 | <othermaciej> | Lachy: http://www.apple.com/safari/whats-new.html |
| 15:45 | <annevk> | I hope the buffer overflow is restricted to awesomeness |
| 15:45 | <gsnedders> | Wait, SFX has been renamed to "Nitro" now? |
| 15:46 | <Lachy> | othermaciej, I have an old Safari 4 developer release installed. Do i need to remove that first? |
| 15:46 | <yecril71> | � and there are forms that contain the data for didactical problems, |
| 15:46 | <othermaciej> | Lachy: it should be able to update over it just fine |
| 15:46 | <yecril71> | e.g. in mathematics or in grammar. |
| 15:47 | <yecril71> | I would use LI LABEL for bureaucratic forms, and P for educational problems. |
| 15:47 | <yecril71> | And, of course, there are numerous other possibilities, these are just the two I have come across. |
| 15:47 | <othermaciej> | gsnedders: our marketing people have a strange sense of humor apparantly |
| 15:48 | <Lachy> | thanks for not warning me I'd need to restart after it installed. That's annoying. I thought that wasn't supposed to be needed on OS X :-( |
| 15:48 | Lachy | restarts his macbook |
| 15:48 | <yecril71> | DT, DD INPUT is another possibility; however, I think LI LABEL renders better by default. |
| 15:49 | yecril71 | is skeptical about Safari |
| 15:49 | <gsnedders> | othermaciej: Nitro is so boring compared with extreme fish |
| 15:49 | <yecril71> | Safari was all broken to me when it came around April. |
| 15:50 | <othermaciej> | gsnedders: are you calling this guy boring: http://www.youtube.com/watch?v=JPqFxtWgsvQ |
| 15:50 | <yecril71> | And none of the problems I have reported have been addressed. |
| 15:50 | <yecril71> | And the ADC Web site does not even work in IE :=( |
| 15:50 | <yecril71> | So I decided to trash the whole thing. |
| 15:51 | <othermaciej> | what problems did you report? |
| 15:52 | <yecril71> | For example, ?xml-stylesheet href="#". |
| 15:52 | <karlcow> | yecril71: "extremely fishy" would amuse journalists… though if the browser is not stable nitro will find its way in the columns too. There is always a way to make fun of something |
| 15:52 | <zcorpan> | othermaciej: "This video is not available in your country." |
| 15:53 | <othermaciej> | yecril71: is that supposed to make an XML file act as its own XSLT stylesheet? or by '#' do you mean a gragment identifier? |
| 15:54 | <yecril71> | self-referencing stylesheet, yes. |
| 15:54 | <yecril71> | The workaround is you have to use the actual name. |
| 15:54 | <yecril71> | (but it breaks when you rename the resource) |
| 15:55 | <yecril71> | Or ignoring external DTDs. |
| 15:55 | Philip` | wonders what will come next after the V8 and Nitro engines |
| 15:55 | <karlcow> | I wonder what "Disable Site-specific Hacks" means |
| 15:56 | <gsnedders> | How's that a bug? There's no requirement to read external DTDs… |
| 15:56 | <karlcow> | Philip`: Nuclear |
| 15:56 | <othermaciej> | incidentally, if any of you get it installed, the little first-time intro thing is not a video, it's made with HTML5 and CSS Animation |
| 15:56 | <karlcow> | or maybe Nuked! |
| 15:56 | <yecril71> | It is a bug for me, because it is denying a useful feature. |
| 15:56 | <othermaciej> | (well, there's a video in it, using <video>) |
| 15:56 | <annevk> | yecril71, I doubt that's a bug per spec |
| 15:56 | <Philip`> | karlcow: Fallout 3 has taught me that nuclear-powered cars blow up in a mushroom cloud after three gunshots, so that's probably not a great name |
| 15:56 | <karlcow> | hehe |
| 15:57 | <yecril71> | That is a bug per usability, not per spec. |
| 15:57 | <annevk> | there's no such bugs for technical questions |
| 15:57 | <gsnedders> | yecril71: You then have the overhead of the time taken to fetch the DTD, which is a bug per usability |
| 15:57 | <yecril71> | I prefer correctness over efficiency. |
| 15:58 | <gsnedders> | Per spec it is correct, though |
| 15:58 | <yecril71> | But I can tell it to Safari because Safari knows better. |
| 15:58 | <yecril71> | s/can/cannot/ |
| 15:58 | <othermaciej> | I don't think we have any principled reason to refuse to support that, we just haven't got around to fixing it yet |
| 15:58 | <othermaciej> | if there are sites using it in the wild we could raise the priority |
| 15:59 | <yecril71> | And I still cannot get rid if its traces in the registry after I removed it. |
| 15:59 | <othermaciej> | otherwise, we'll get to it eventually, and patches welcome |
| 15:59 | <karlcow> | othermaciej: is there an easteregg to replay the first time intro or the files somewhere on the hard drive? |
| 15:59 | <othermaciej> | http://www.apple.com/safari/welcome/ |
| 16:00 | <Lachy> | othermaciej, I like what's been done with the Speed Dial feature in Safari, and cover flow in the history search is nice |
| 16:00 | <yecril71> | patches to Safari are possible only from people who have the source code. |
| 16:01 | <othermaciej> | yecril71: http://webkit.org/ |
| 16:01 | <yecril71> | Preferences are UI, and UI is closed. |
| 16:01 | <othermaciej> | that's the engine used by Safari, it's all open source, and any code relating to XSLT will be in there |
| 16:01 | <othermaciej> | don't see what preferences have to do with it? |
| 16:02 | <yecril71> | "Read external DTDs" is a preference. |
| 16:02 | <othermaciej> | what does that have to do with ?xml-stylesheet href="#" ? |
| 16:03 | <yecril71> | We were discussing the second item. |
| 16:03 | <othermaciej> | oh |
| 16:03 | <yecril71> | href="#" can be patched in WebKit, yes. |
| 16:03 | <othermaciej> | well that's a feature we probably wouldn't add, preference or no |
| 16:05 | <yecril71> | That is, of course, up to you, but then Safari installer should not claim Safari is "the browser". |
| 16:05 | <yecril71> | Because it is too crippled and limited to be "the browser". |
| 16:06 | <karlcow> | hmmmm wonders what is /Applications/Safari.app/Contents/Resources/SnippetEditor |
| 16:06 | <annevk> | your sense of logic is not allowed in this channel yecril71, see topic ;) |
| 16:06 | <jgraham> | No Linux. Less bling than a rap star. Lame ;) |
| 16:06 | <gsnedders> | karlcow: Develop -> Show Snippet Editor |
| 16:06 | <zcorpan> | othermaciej: what's new with the "enhanced keyboard navigation"? |
| 16:07 | <Lachy> | othermaciej, http://images.apple.com/safari/welcome/media/audio.mp4 (the audio file used on the Safari welcome page) is served as text/plain :-( |
| 16:07 | <yecril71> | annevk! Was that a plonk? |
| 16:08 | <othermaciej> | zcorpan: not sure |
| 16:08 | <othermaciej> | Lachy: I will try to let the right people know |
| 16:08 | <othermaciej> | Lachy: is the video served with the right MIME type? |
| 16:08 | <karlcow> | gsnedders: had missed that. thanks. still don't know what is it |
| 16:09 | <Lachy> | othermaciej, yeah, compass.mov is served as video/quicktime |
| 16:10 | yecril71 | �s system is always loaded |
| 16:10 | <yecril71> | so I would rather that the player degraded by itself. |
| 16:12 | <yecril71> | Hixie! If I isolated the problem that slows down IE rendering your spec as it is, |
| 16:12 | <yecril71> | would you implement some workaround? |
| 16:14 | <Lachy> | othermaciej, I'm curious why the developers who made the welcome page decided to do it with separate <video> and <audio> elements, instead of just adding the audio track to the video file |
| 16:14 | <othermaciej> | Lachy: the video is shorter than the audio, and it turns out this way has better performance |
| 16:15 | <Lachy> | hmm, that's interesting. |
| 16:15 | <Lachy> | does it improve network performance by being able to download the tracks separately, or is it more performant to play them as 2 separate tracks? |
| 16:18 | <Lachy> | I think I prefer Chrome's tab design over Safari's because it's more practical. When I click on a tab in Safari while I'm still moving the mouse, it thinks I want to move the window. Whereas, Chrome makes a clearer distinction between the title bar and each tab and doesn't suffer from that problem |
| 16:19 | <karlcow> | "Show Snippet Editor: Opens a window you can use to quickly test small fragments of HTML, without requiring you to open an entire webpage." |
| 16:24 | <karlcow> | I'm pretty sure there must be com.apple.blabla to tell safari to restart with my previous sessions. Quit with two tabs, reopened with one window and no tabs. But there is a menu for it now |
| 16:25 | <Lachy> | karlcow, hasn't the "Reopen All Windows From Last Session" menu item been available in Safari for a long time? |
| 16:27 | <karlcow> | Lachy: not sure. I barely use safari since safari 2. Just launching it usually when new versions arrive and testing for specific purpose. |
| 16:27 | karlcow | is checking % defaults read com.apple.safari |
| 16:30 | <karlcow> | made a test |
| 16:30 | <karlcow> | mwahaha |
| 16:30 | <karlcow> | "com.apple.WebKit.searchField:com.twitter.search" = ( |
| 16:30 | <karlcow> | sex |
| 16:30 | <karlcow> | ); |
| 16:33 | <karlcow> | ah good, when in private browsing mode, it doesn't save it to the com.apple.safari |
| 16:33 | <karlcow> | it just keeps the old values |
| 16:35 | <zcorpan> | othermaciej: hmm, popup windows in general don't seem to work for me (winxp); e.g. "about safari", "customize toolbar" and javascript alerts |
| 16:35 | <othermaciej> | zcorpan: strange |
| 16:36 | <zcorpan> | i like having the tabs in the title bar |
| 16:39 | <othermaciej> | zcorpan: did you update from the Safari 4 Developer Preview? |
| 16:39 | <zcorpan> | othermaciej: yes |
| 16:39 | <othermaciej> | zcorpan: I'm told what you are seeing is a known bug and should go away if you quit and relaunch Safari |
| 16:39 | <karlcow> | zcorpan: me too. but there is a little issue, with long titles when many tabs opened. At least when you over the title it displays the full title |
| 16:40 | <zcorpan> | othermaciej: it was a bit tricky to quit since trying to quit opened a popup that i couldn't see (and none of esc or enter made safari quit) :) |
| 16:41 | <karlcow> | the top sites is quite nice |
| 16:41 | <othermaciej> | zcorpan: doh! |
| 16:41 | <zcorpan> | works now though |
| 16:41 | <jgraham> | Does the coverflow thing actually work well? I tend to find that coverflow looks nicer in screenshots than it is to use in real life... |
| 16:42 | <karlcow> | jgraham: for the history? |
| 16:42 | <jgraham> | karlcow: Yeah |
| 16:42 | <jgraham> | I imagine *screenshots* for history are a good idea |
| 16:42 | <karlcow> | I would prefer an expose like thing |
| 16:43 | <karlcow> | cover flow is hard to use usually (for me), maybe other users might have a different experience |
| 16:47 | <karlcow> | ooops Top Sites does a get on previous uris which were a POST. |
| 16:49 | <karlcow> | that seems to be a bug |
| 16:55 | karlcow | really likes the Web Inspector. nice job |
| 16:55 | jgraham | apologises to zcorpan |
| 16:58 | annevk | wonders why Number(x) is not enough if parseInt(x) is not good enough because there's no way to do input validation |
| 16:59 | <karlcow> | oohhh safari messing up with mime-type and content sniffing |
| 17:00 | <karlcow> | http://www.w3.org/ contains w3c_main |
| 17:01 | <karlcow> | Web Inspector reports Resource interpreted as image but transferred with MIME type text/html. w3c_main |
| 17:01 | <jgraham> | annevk: No idea. Having two functions to do the same thing seems like a really bad idea though |
| 17:01 | <annevk> | I agree |
| 17:01 | <jgraham> | TSBO(APOO)OWTDI |
| 17:01 | <karlcow> | but w3c_main is sent with the right mime-type http://web-sniffer.net/?url=http%3A%2F%2Fwww.w3.org%2FIcons%2Fw3c_main&submit=Submit&http=1.1&gzip=yes&type=GET&uak=0 |
| 17:02 | <annevk> | I'm a bit scared by es-discuss now and then |
| 17:02 | <annevk> | it seems quite easy for things to go the wrong way |
| 17:02 | <annevk> | though hopefully that's only because it's hard to follow what's going on |
| 17:03 | <annevk> | i.e., just like some people are complaining about the WHATWG / HTML WG :) |
| 17:03 | <jgraham> | annevk: I have been on the list for mere days and I already sense a predilection for theoretical purity over solutions that will actually be deployed |
| 17:04 | <jgraham> | Although I hope that is unfair |
| 17:04 | <annevk> | well, I've been on the list for over a year and have the same kind of sense |
| 17:04 | <annevk> | dunno |
| 17:05 | <annevk> | time to take a break |
| 17:07 | karlcow | reported the mime type bug |
| 17:31 | <jgraham> | http://intertwingly.net/blog/2009/02/24/Cataloging-Venial-Sins |
| 17:46 | <jgraham> | karlcow: The example of spelling and grammar rules always suffers because spelling and grammar rules are extremely descriptivist |
| 17:47 | <jgraham> | Which could be your point but you should say so explicitly |
| 17:48 | <karlcow> | see html-wg |
| 17:48 | <jgraham> | (that maybe wasn't clear. I mean that the "rules" are mainly observations of what real people do rather than any kind of spec. And when people do try to make up a spec it usually contains nonsense like "don't split infinitives" which is based on how Latin worked...) |
| 17:54 | <karlcow> | you mean the current grammar rules are not really what really people speak, but what they should write. |
| 17:56 | Philip` | wonders what the relevance of Gödel is to HTML document conformance |
| 17:57 | karlcow | wonders what is the relevance of challenging the document conformance, or what does it achieve in the end. mesmerized |
| 18:03 | <karlcow> | hmmm the issue with the mime type for safari is happening with only of this type: Content-Type: image/png; qs=0.7 |
| 18:06 | <jgraham> | karlcow: I'm saying grammer "rules", as understood by modern linguists, are based on a description of what people actually say / write. |
| 18:08 | <karlcow> | and I'm talking about language taught at school |
| 18:10 | <karlcow> | and the way I write is definitely not the way I speak plus a big variety depending on the context, including respect forms, humour, etc. My language is mostly super neutral here. |
| 19:00 | <yecril71> | I still think METER is enough for marking up scalar quantities. |
| 19:11 | <rubys> | jgraham: re: "I am suggesting that using [RFC 2119] as the basis for arguing that we should have weaker document conformance constraints ignores the practical value of those constraints" +1 |
| 19:13 | <tantek> | hsivonen, I just saw a few references to the profile attribute and XMDP - I may be able to help answer questions. |
| 19:14 | <Dashiva> | "Seems like a lot of trouble for just Links/Lynx users." <-- So much for universal access :) |
| 19:14 | <yecril71> | Tantek! Do you know why OBJECT was proposed to hold dates? |
| 19:14 | <yecril71> | Seems like utter nonsense to me. |
| 19:15 | <tantek> | yecril71, because OBJECT is a semantic for embedding arbitrary data |
| 19:15 | <tantek> | many things seem like nonsense until explained |
| 19:15 | gsnedders | needs to wake up |
| 19:16 | <yecril71> | Arbitrary data, like white noise? |
| 19:16 | <tantek> | it is typically quite unscientific to conclude that something is "utter nonsense", negative proofs are extraordinarily difficult |
| 19:16 | <yecril71> | I said "seems". |
| 19:17 | <yecril71> | Object is semantics for external content, not for arbitrary data. |
| 19:17 | <gsnedders> | Fear the axiomatic proof (by contradiction)! |
| 19:19 | <yecril71> | And dates are not external content, although a calendar widget may be. |
| 19:19 | <Philip`> | gsnedders: Is proof by contradiction one of your axioms? |
| 19:19 | <tantek> | fair enough, the seems qualifier improves it yes |
| 19:19 | <Philip`> | Clearly you should be using constructivist logic instead |
| 19:19 | <gsnedders> | Philip`: But that's boring. |
| 19:20 | Dashiva | has bad memories about deducing a->a from those stupid axioms |
| 19:21 | <tantek> | expressions of doubt or lack of understanding (questions) are more conducive to scientific discourse than assertions of nonsense. i.e. better to think/say "I don't understand why {assertion}" than "{assertion} seems like (or is) nonsense". the latter statement tends to be counter-productive, and in my experience, falsely implies a negative proof. |
| 19:22 | <yecril71> | I do not understand why OBJECT would make a good element for making up dates then. |
| 19:23 | yecril71 | perceives the second version much less interesting |
| 19:23 | <tantek> | like many applications of semantic HTML, the framing is a bit different |
| 19:23 | <Dashiva> | No one said it's good, but rather that it's the default way |
| 19:23 | <tantek> | rather than asking for each element is this good/bad for marking up x |
| 19:23 | <tantek> | the question is, what is the best choice in semantic HTML for marking up x |
| 19:24 | <yecril71> | A choice that is no good cannot be the best |
| 19:24 | <yecril71> | so it should be automatically out of consideration. |
| 19:24 | <tantek> | actually it can, because best is a relative assessment, rather than absolute "good" |
| 19:24 | <yecril71> | I do not understand why it was considered. |
| 19:25 | <yecril71> | So it was just the first shot? |
| 19:25 | <tantek> | object is not just semantics for external content, hence why object has fallback, all the way down to the innertext |
| 19:26 | <tantek> | that's a very important distinction between object and img/iframe/embed/applet etc. |
| 19:26 | <tantek> | thus the semantic for providing alternate types of data |
| 19:27 | <yecril71> | But in this case you want the alternate content displayed, not the object data. |
| 19:27 | <tantek> | precisely |
| 19:27 | <yecril71> | So it is rather an antiobject. |
| 19:27 | <tantek> | and that's not difficult to do, with implementation that conform to HTML4 |
| 19:27 | <tantek> | implementationS |
| 19:28 | <tantek> | and that's why I first tried to use OBJECT as the most semantic choice in semantic HTML to represent datetimes, just over four years ago now. http://tantek.com/log/2005/01.html#d26t0100 |
| 19:29 | <tantek> | unfortunately, bugs in Safari prevented this from being a practical solution. |
| 19:30 | <yecril71> | OBJECT displays object data by default. |
| 19:30 | <yecril71> | If fallback data should be normally displayed, it cannot be OBJECT. |
| 19:33 | <tantek> | ideally, for semantic datetimes, what should be displayed is what the viewer can best understand and manipulate, which may for example be a localized version of the datetime |
| 19:33 | <tantek> | but given lack of localized-preference-aware datetime display plugins, then yes, a suitable inline fallback should be provided consistent with the statistically likely locale of the reader |
| 19:34 | <yecril71> | Localized datetime in an article given in a foreign language? |
| 19:34 | <tantek> | or for example, in a train schedule shown in a foreign language |
| 19:34 | <yecril71> | If you understand the foreign language, it serves only as a disruption. |
| 19:35 | <tantek> | any experience international traveler knows that just being able to read the dates and times of schedules etc makes them FAR more usable, even if you don't know the foreign language |
| 19:35 | <tantek> | so yes, even if in a foreign language |
| 19:35 | <tantek> | to put it another way, making *some* of the content viewable / understandable is often preferable to *none* |
| 19:36 | <tantek> | numbers, measures, dates, times fall into this category |
| 19:36 | <yecril71> | I can see it depends on the reading environment, because sometimes it can be worse. |
| 19:36 | <tantek> | in practice it is far more often better |
| 19:36 | <tantek> | as in the actual example I gave above |
| 19:36 | <yecril71> | Not in my practice. I do not travel that much. |
| 19:36 | <tantek> | of international travel and train schedules |
| 19:37 | <yecril71> | Besides, aren�t these things produced from a database rather then authored? |
| 19:37 | <tantek> | your expression of a lack of personal utility does not refute any positive assertions of utility made by others. |
| 19:38 | <yecril71> | I am not trying to refute your POV. |
| 19:38 | <tantek> | the point is, the practice increases utility for some or many |
| 19:38 | <tantek> | thus is an improvement |
| 19:38 | <yecril71> | And the database engine could produce content according to the viewer�s settings |
| 19:38 | <tantek> | it increases usability, heck accessibility for that matter |
| 19:38 | <yecril71> | without any help from the browser |
| 19:39 | <yecril71> | i.e. provide localization with no special markup |
| 19:39 | <yecril71> | necessary |
| 19:40 | <tantek> | your asserted solution has poorer scaling characteristics (requiring all publishers to code everything aware for all locales) than the alternative of simply providing the data semantically, and having the localize viewing smarts be in the user's browser |
| 19:40 | <yecril71> | Not all publishers, just travel agencies, railways and the like. |
| 19:41 | <tantek> | all publishers of date and time information |
| 19:41 | <yecril71> | They have to operate in a multilingual environment already. |
| 19:41 | <tantek> | event sites etc. |
| 19:41 | <yecril71> | What is the utility of attending an event when you do not know the language of communication? |
| 19:42 | <tantek> | maintaining 1 or 2 or a small number of variants is not the same as maintaining all variants for all locales, so the "already" argument is false. |
| 19:42 | <tantek> | you are assuming the language that the event is posted in is required to attend and take value in the event, which is a poor assumption. |
| 19:43 | <tantek> | taking a train or bus is only one example of the problem of that assumption |
| 19:43 | <tantek> | concerts are another |
| 19:43 | <yecril71> | Oh, I thought you meant social events. |
| 19:43 | <tantek> | regardless, a very poor assumption to make, and certainly insufficient to reason from |
| 19:45 | <yecril71> | It is valid for social events where people talk to each other. |
| 19:45 | <tantek> | even then, you can't know what language(s) the attendees may know and thus speak to each other |
| 19:46 | <tantek> | you might be able to statistically model some likelihood of number of attendees knowing particular languages, given the demographics |
| 19:46 | <yecril71> | The attendees agree to use a common language |
| 19:46 | <yecril71> | or the organisation hires live translators. |
| 19:46 | <tantek> | or people talk however they want to talk to each other |
| 19:46 | <yecril71> | Otherwise it does not make sense. |
| 19:47 | <tantek> | see above examples. again, asserting "it does not make sense" itself is not a helpful path for discourse. |
| 19:48 | <yecril71> | An example of an event where people talk to each other without having to understand each other I have not seen above. |
| 19:49 | <yecril71> | I can only think about the tower of Babel event when I see this. |
| 19:49 | <tantek> | people are often able to communicate some understanding to each other without having to speak the same language |
| 19:49 | <tantek> | reasoning by negative examples like that will lead you to many errant conclusions because of implied assumptions |
| 19:50 | <yecril71> | I disagree about that "often" thing. |
| 19:50 | <yecril71> | Only in extreme circumstances, most of them unplanned. |
| 19:51 | <tantek> | lack of examples = you need to accept doubt and not knowing something rather than knowing or asserting that something is not. |
| 19:51 | <tantek> | and "only" is a strong assertion to make without evidence |
| 19:52 | <tantek> | get more experience with international travel, and you will get more data for understanding cross-language cross-locale etc. issues, problems and solutions |
| 19:52 | <yecril71> | I meant, only in extreme circumstances people often� |
| 19:52 | <tantek> | right, "only" is making assumptions of negative examples |
| 19:53 | <yecril71> | You mean, travelling abroad will make me attend social events based on nonverbal communication? |
| 19:54 | <tantek> | traveling abroad -> more data for understanding cross-language cross-locale etc. issues, problems and solutions -> more appreciation of partial understanding of content in foreign languages. |
| 19:54 | <tantek> | whether that content is on the web, or in communication at social events |
| 19:54 | <yecril71> | However, understanding of dates is included in partial understanding of content. |
| 19:55 | <yecril71> | It is basic knowledge. |
| 19:55 | <tantek> | until then, rather than asserting negatives, you should start with "I don't have the experience therefore I don't know", and ask those with experience. |
| 19:55 | <yecril71> | I do not have the experience, therefore I do not know |
| 19:55 | <yecril71> | whether people who travel abroad much |
| 19:56 | <yecril71> | are often exposed to planned social events |
| 19:56 | <yecril71> | where the attendees do not understand each other |
| 19:56 | <yecril71> | and therefore have to resort to nonverbal communication. |
| 19:56 | <yecril71> | If you have the experience needed to enlighten me in this matter, please do. |
| 19:58 | yecril71 | thinks the problems with Safari were a fortunate accident |
| 19:59 | <tantek> | see W3C i18n presentations by Richard Ishida regarding how dates and times are not basic knowledge and actually quite hard to understand well / accurately across languages and locales. |
| 20:00 | <yecril71> | Gwneud y we fyd-eang yn wirioneddol? |
| 20:01 | <tantek> | again your example above is trying to reason by negative, in this case, a strawman you constructed ("planned social events" - "do not understand" - "nonverbal communication"), which is also a logical fallacy: http://en.wikipedia.org/wiki/Straw_man |
| 20:01 | <tantek> | Hixie, do you have a Logic101 primer for the #whatwg channel, or are discussions like this (with flawed reasoning by negative examples, strawman etc) a common thing? |
| 20:01 | <yecril71> | That was actually a question, not a reasoning. |
| 20:03 | <yecril71> | Do you want me to look at the wirioneddol thing? |
| 20:04 | <tantek> | see above: presenting cases/questions with an absence of utility does not refute cases presented that demonstrate utility. |
| 20:05 | <yecril71> | Like that with concerts? |
| 20:10 | <tantek> | it appears you may have referring to a resource on the web with the previous question. URLs usually work better to reference things on the web (rather than the titles of documents). |
| 20:12 | <tantek> | this presentation is a good start: http://www.w3.org/2008/Talks/05-atmedia-ishida/slides.pdf |
| 20:12 | yecril71 | is looking at that |
| 20:13 | yecril71 | found wirioneddol and thought it may be the refernced document but it was not |
| 20:14 | <yecril71> | But it has 113 pages, and I want information about dates� |
| 20:15 | <yecril71> | and no TOC |
| 20:15 | <tantek> | this one is good too: http://www.w3.org/2007/Talks/0706-atmedia/ |
| 20:16 | <tantek> | having expressed lack of experience about language/locale issues, you should read through both of those presentation to gain more understanding |
| 20:17 | <yecril71> | There is plenty of things I should do but I do not have infinite resources :-( |
| 20:17 | yecril71 | needs a shortcut |
| 20:17 | <tantek> | here is a specific slide on dates: http://www.w3.org/2007/Talks/0706-atmedia/slides/Slide0460.html |
| 20:19 | <yecril71> | If I know that the page is in English, I also know that 03 is March. |
| 20:19 | <tantek> | for a better experience, listen to the audio from the previous link as well: http://www.w3.org/blog/International/2007/06/11/new_talk_slides_media_2007 |
| 20:19 | <yecril71> | I do not know from what culture 2.3.01 may be. |
| 20:20 | <tantek> | you don't actually know that, and here are several examples of ambiguous dates, many from English-speaking countries: http://flickr.com/photos/tags/needsisodate |
| 20:20 | <tantek> | English UK dates are written different from English US dates |
| 20:20 | <tantek> | so when is 4/5 ? in April or May? |
| 20:21 | <yecril71> | English UK is Mar 2, 2001 |
| 20:21 | <yecril71> | or March. |
| 20:21 | <tantek> | hence ambiguous, hence not basic knowledge |
| 20:22 | <yecril71> | I can easily tell 03/02/01 (US) from March 2, 2001 (UK) |
| 20:23 | <yecril71> | No ambiguity here |
| 20:23 | <tantek> | 03/02/01 is not necessarily US |
| 20:23 | <tantek> | see the Flickr link |
| 20:23 | <tantek> | some of those dates are printed in that format in the UK |
| 20:23 | <tantek> | in speaking with Richard Ishida, he agreed that using a date like "2001-03-02" is actually the most readable to the most people, worldwide. |
| 20:23 | <yecril71> | IMG_2410 is not from the Web, there is no context |
| 20:24 | <tantek> | it appears you may have referring to a resource on the web with the previous question. URLs usually work better to reference things on the web (rather than the titles of documents). |
| 20:24 | <yecril71> | YYYY-MM-DD is ANSI date, used in France. |
| 20:24 | <yecril71> | http://flickr.com/photos/tantek/3241965515/ |
| 20:24 | <tantek> | YYYY-MM-DD is ISO8601. which ANSI standard is it? |
| 20:25 | <yecril71> | ANSI SQL |
| 20:25 | <karlcow> | http://www.w3.org/TR/NOTE-datetime |
| 20:26 | <tantek> | "/" separators are used in the UK as well as the US. e.g. http://flickr.com/photos/tantek/1386536315/ |
| 20:26 | <karlcow> | http://www.w3.org/QA/Tips/iso-date |
| 20:26 | <tantek> | or ".": http://flickr.com/photos/tantek/1341747652/ |
| 20:27 | <karlcow> | http://www.w3.org/International/questions/qa-date-format |
| 20:27 | <tantek> | karlcow - thanks for the URLs. |
| 20:28 | <tantek> | from http://www.w3.org/QA/Tips/iso-date : "Albeit not perfect, ISO date format is, however, the best choice for a date representation that is universally (and accurately) understandable." |
| 20:28 | <yecril71> | http://flickr.com/photos/tantek/1341747652/ is not from the Web either |
| 20:29 | <yecril71> | My country used D.MM.RRRR, then RRRR.MM.DD, I suppose adoption of ISO is in progress |
| 20:30 | <tantek> | right, that's what people will end up using, because it makes the most sense to the most people (= most accessible) |
| 20:30 | <yecril71> | <http://flickr.com/photos/tantek/1386536315/>; is not from the Web |
| 20:30 | <tantek> | the point of those photos is to demonstrate differences in written date formats in different locales |
| 20:31 | <tantek> | ISO8601: http://en.wikipedia.org/wiki/ISO_8601 |
| 20:32 | <yecril71> | Demonstrate that different date formats are used in a single locale |
| 20:32 | <yecril71> | on pages containing invitations to events |
| 20:32 | <yecril71> | causing all sorts of misunderstandings. |
| 20:32 | <yecril71> | Consumer product validity tags are out of scope for this discussion. |
| 20:33 | <tantek> | they're not, as people travel across locales IRL, as they do when browsing the web |
| 20:34 | <yecril71> | But you cannot use OBJECTs on validity tags, so using OBJECTs in those cases would not help. |
| 20:34 | <tantek> | what are "validity tags"? |
| 20:34 | <yecril71> | Valid through |
| 20:34 | karlcow | fears the unproductive discussion :) |
| 20:35 | tantek | tends to agree with karlcow, given the frequency of logic flaws. |
| 20:38 | yecril71 | does not admit committing a logic flaw |
| 20:38 | <tantek> | by the way, for folks interested on how to "best" markup dates and times on the web, we (the #microformats community) made good progress in microformats with additions to the value-excerption pattern to workaround the cases where using abbr may present accessibility or presentation problems. for more on this topic (which may be offtopic for #whatwg - I'm not sure), feel free to /join #microformats |
| 20:39 | <yecril71> | And? |
| 20:39 | <tantek> | which is also a reasonable place to ask about XMDP (which was the subject I brought up in the first place). |
| 20:42 | <tantek> | yecril71, see above where I noted reasoning by negative (examples), strawman examples etc. every time I pointed those out, they were in response to logical flaws in your statements / arguments. acknowledging your logic flaws will help with recognizing them in the future, reduce the frequency, and result in higher quality / more productive discussions. |
| 20:53 | yecril71 | was not reasoning by negative examples |
| 20:53 | yecril71 | was merely pointing out the cases where the strategy is not really needed |
| 20:54 | yecril71 | thinks other cases exist but they are far less common |
| 21:17 | yecril71 | found his objection raised by Sarven Capadisli at <http://microformats.org/wiki/datetime-design-pattern>; |
| 21:18 | yecril71 | is glad he is not the first one |
| 21:20 | gsnedders | tries to get his head around working with RDF |
| 21:23 | <karlcow> | gsnedders: what are you trying to do? |
| 21:23 | <gsnedders> | karlcow: Using http://www.w3.org/2002/01/tr-automation/tr.rdf find the latest version of each URI and find the title, authors, and date of that document |
| 21:24 | <yecril71> | Owls are experts at getting their head around. |
| 22:12 | <karlcow> | http://280atlas.com/ |
| 22:12 | <karlcow> | gsnedders: do you need examples of scripts? |
| 22:57 | <gsnedders> | oh noes! othermaciej following me on Twitter! |
| 22:57 | gsnedders | hides |
| 22:58 | <othermaciej> | gsnedders: you followed me first dude |
| 22:58 | <gsnedders> | Yeah, sure. |
| 22:58 | <gsnedders> | I never said anything to the contrary :P |
| 22:59 | karlcow | still wonders if gsnedders needs scripts examples |
| 23:00 | <gsnedders> | karlcow: It would be helpful. |
| 23:05 | <karlcow> | gsnedders: SPARQL package in python to do query on an RDF graph http://www.ivan-herman.net/Misc/PythonStuff/SPARQL/Doc-SPARQL/ |
| 23:05 | <gsnedders> | karlcow: Yeah, I saw that. Only issue: I don't know SPARQL at all |
| 23:08 | <karlcow> | SPARQL looks like a lot SQL |
| 23:09 | <karlcow> | and if you don't want to go through the insanity of parsing RDF with XSLT (aka headaches guaranteed) you should use SPARQL |
| 23:09 | <karlcow> | http://www.ibm.com/developerworks/xml/library/j-sparql/ |
| 23:09 | <karlcow> | in this article there are very basic examples |
| 23:09 | <karlcow> | the first example is |
| 23:09 | <karlcow> | PREFIX foaf: <http://xmlns.com/foaf/0.1/>; |
| 23:09 | <karlcow> | SELECT ?url |
| 23:09 | <karlcow> | FROM <bloggers.rdf> |
| 23:09 | <karlcow> | WHERE { |
| 23:09 | <karlcow> | ?contributor foaf:name "Jon Foobar" . |
| 23:09 | <karlcow> | ?contributor foaf:weblog ?url . |
| 23:10 | <karlcow> | } |
| 23:11 | <gsnedders> | karlcow: How do I deal with not knowing what element is the child of root (e.g., it might be WD, it might be REC, etc.)? |
| 23:13 | tantek | wonders rhetorically if there might be a better IRC channel for RDF discussions. |
| 23:13 | <gsnedders> | Peh. Being on-topic would be logic, and we're meant to leave that at the door when coming in here. |
| 23:13 | <gsnedders> | s/logic/logical/ |
| 23:17 | <tantek> | being logical can often improve the economic use of your resources such as time, assuming that is part of your self-interest. |
| 23:19 | <danbri> | I guess SQL is in scope nowadays? |
| 23:20 | <karlcow> | gsnedders: I just created one for you |
| 23:20 | <karlcow> | so you can understand |
| 23:20 | <karlcow> | Use the SPARQL endpoint http://sparql.org/sparql.html with the form |
| 23:20 | <karlcow> | and type |
| 23:20 | <karlcow> | PREFIX dc: <http://purl.org/dc/elements/1.1/>; |
| 23:20 | <karlcow> | PREFIX rec: <http://www.w3.org/2001/02pd/rec54#>; |
| 23:20 | <karlcow> | SELECT ?title |
| 23:20 | <karlcow> | FROM <http://www.w3.org/2002/01/tr-automation/tr.rdf>; |
| 23:20 | <karlcow> | WHERE { |
| 23:20 | <karlcow> | ?spec a rec:REC . |
| 23:20 | <karlcow> | ?spec dc:title ?title . |
| 23:21 | <karlcow> | } |
| 23:21 | <danbri> | i've been nudging the http://www.w3.org/2005/Incubator/rdb2rdf/ folk (working on sql / sparql mappings) to look at html5 sql apis |
| 23:21 | <karlcow> | here I'm asking to display the title of all documents which are a Rec |
| 23:21 | <danbri> | eg. in last week's semweb cg call |
| 23:22 | <karlcow> | tantek: you can ignore the lines too ;) I very rarely read the logs of this channel for example, and that's fine. |
| 23:23 | <tantek> | karlcow, focus / and on-topic scales better in terms of communication for more people, rather than telling everyone to ignore lines and speak whatever in whatever context. |
| 23:24 | <danbri> | tantek, are you new here? #whatwg isn't famed for ontopicity... |
| 23:25 | <tantek> | danbri, I lurk often, and most of the time see quite on topic discussions between Hixie, hsivonen, et al |
| 23:27 | <karlcow> | tantek giving a meta class related to good use of channel (not respecting the channel topic then) aka respecting the topic (here html5). Conclusion: eat your own dog food. :))) |
| 23:28 | <karlcow> | gsnedders: I realize that I have forgotten to give you the link to the description of the vocabulary so you can play with other items of the graphs. http://www.w3.org/2001/02pd/rec54# |
| 23:28 | <tantek> | karlcow, html5 being ontopic I presumed that semantic use of HTML4 was also ontopic - pardon me if that was an error. |
| 23:29 | <tantek> | for microformats specific discussions I do redirect folks to #microformats |
| 23:29 | karlcow | is just talking about the meta discussion we are just having now about IRC usage ;) |
| 23:29 | <Hixie> | there's no topic in this channel |
| 23:29 | <Hixie> | people can talk abotu whatever they want to |
| 23:29 | <Hixie> | it's just a place for people who work on html5 to wind down |
| 23:30 | karlcow | doesn't really care. gsnedders is a nice chap. He asked for help, I'm giving a bit of help. And it might benefit others. end of the metadiscussion on my side. |
| 23:30 | <gsnedders> | karlcow: I'll have a look at all that tomorrow |
| 23:30 | <Hixie> | helping people is always welcome on this channel :-) |
| 23:30 | <gsnedders> | karlcow: I have to go sleep now |
| 23:30 | <karlcow> | gsnedders: have a good night |
| 23:31 | <gsnedders> | Besides, I could always go back to discussing it with MikeSmith in #html-wg where it almost certainly is off-topic :P |
| 23:31 | <karlcow> | hehe |
| 23:32 | <tantek> | Hixie, apologies, for some reason I gotten the impression that the purpose/focus of this channel was HTML5 ;) |
| 23:32 | <Hixie> | nah the purpose is just so i have a rapturous audience to listen to my rants |
| 23:32 | Hixie | ducks |
| 23:32 | <gsnedders> | Hixie: Re: your email, that means I can please both you and Anne at once. Wow. :) |
| 23:33 | <Hixie> | gsnedders: such a rare opportunity! |
| 23:37 | Philip` | wonders how off-topic he could attempt to drag the channel before it just got very irritating to everyone |
| 23:37 | <Philip`> | So, is anyone else waiting for the TF2 scout update? |
| 23:37 | <gsnedders> | Philip`: Oh come on. Try harder. |
| 23:38 | <Hixie> | Philip`: no but I _am_ waiting for Half-Life 2 Episode 3 |
| 23:39 | <Philip`> | The double-Payload map looks quite interesting |
| 23:40 | <Philip`> | Hixie: That's not quite the same type of waiting, since Ep3 is a lot of months away and the TF2 update is hours away :-) |
| 23:40 | <Hixie> | :-) |
| 23:41 | <gsnedders> | Anyone else got GTA IV: TLAD yet? |
| 23:42 | <Philip`> | (They seem to have totally failed at the idea that episodic content is meant to be faster to produce than a full game) |
| 23:42 | gsnedders | on the whole thought it was a bit short only taking seven hours to get through |
| 23:42 | <gsnedders> | Though, to be fair, that isn't actually that much shorter than GTA IV itself (11 hours, IIRC) |
| 23:43 | karlcow | is kicking gsnedders to bed |
| 23:43 | <gsnedders> | Oh, yeah. I did mean to do that |
| 23:43 | gsnedders | thanks karlcow for the reminder, and goes |
| 23:43 | <Philip`> | gsnedders: You should play TF2! I must have spent well over three hundred hours on it, which is pretty good value for money |
| 23:43 | <Hixie> | Philip`: indeed |
| 23:43 | <Hixie> | bbiab |
| 23:44 | <gsnedders> | karlcow: Molly sends a hug and kiss, FWIW |
| 23:47 | <karlcow> | yeeeha |
| 23:48 | karlcow | wonders if gsnedders goes to bed with Molly. hmmmm no no don't answer. TMI |