/irc-logs / w3c / #html-wg / 2007-05-01 / end

Options:

# Session Start: Tue May 01 00:00:00 2007
# Session Ident: #html-wg
# [00:05] <mjs> argh, www-html has infected public-html
# [00:05] <hyatt> mjs: ?
# [00:06] <mjs> hyatt: there's been a lengthy flameware on www-html involving Philip Taylor and Tina Holmboe, and many others
# [00:06] <mjs> but someone pointed them to the grown-up table
# [00:07] <hyatt> uh-oh
# [00:07] <hyatt> i just threw my hat into that ring
# [00:07] <hyatt> should i not have? :)
# [00:07] <Philip`> Blame/thank Hixie for pointing them there :-)
# [00:08] <hyatt> i just responded to tina
# [00:08] <mjs> she's gonna /italicize/ your ass
# [00:08] <Hixie> hah
# [00:08] * Hixie thinks it is important to have the detractors as part of the wg, for two reasons
# [00:09] <Hixie> first, they can't later say "well i wasn't part of the wg when they made that mistake"
# [00:09] <hyatt> mjs: that sounds like it would hurt
# [00:09] <Hixie> and second, people tend to defend communties they're in
# [00:09] <Hixie> so if the wg makes a decision that is against their opinion, they'll often start defending it later
# [00:10] <Hixie> but more importantly, we want everyone's input
# [00:10] <Hixie> whether we agree with it or not
# [00:10] <Hixie> after all, we might be wrong
# [00:10] <Hixie> and we'd never find out if we ignored input we disagreed with
# [00:10] <mjs> does anyone /else/ find it ironic that Tina's mails are *chock full* of purely /presentational/ markup?
# [00:11] <Hixie> why is /foo/ presentational?
# [00:11] <Hixie> isn't she just _emphasising_?
# [00:13] <hyatt> what is up with this anti-browser sentiment
# [00:13] <hyatt> that's kind of weird
# [00:15] <Dashiva> Browsers are evil because they allow non-semantic markup
# [00:16] <Dashiva> At least that's how I understand them.
# [00:16] <jgraham> I found the bit about "I hope [...] you are not suggesting disagreement is somehow undesired" ironic, given the WG is consensus based... (of course I understand that's not what she means and initial disagreement can be productive)
# [00:16] <Hixie> the WG is consensus based?
# [00:16] <Hixie> that's going to be fun
# [00:17] <jgraham> Yeah
# [00:18] * Hixie doesn't plan on attempting to get consensus if he's the editor
# [00:18] <hyatt> neither do i
# [00:19] <schepers> ok, that clears it up for me, thanks
# [00:20] <jgraham> The probability of getting >400 people to agree on _anything_ is so close to 0 it makes no difference
# [00:20] <hyatt> yeah thats why it would be a waste of time
# [00:20] <hyatt> there will never be any such thing as "consensus"
# [00:21] <hyatt> especially with the hypervocal dissenters that are sure to plague almost every conversation (even when they are the extreme minority)
# [00:21] <schepers> there's a difference between getting consensus and attempting to ge consensus
# [00:21] <Hixie> yeah
# [00:21] <schepers> get, rather
# [00:21] <Hixie> in the whatwg i basically base all the decisions on argument quality rather than quantity
# [00:22] <hyatt> Hixie: yeah i agree with that although it involves keeping a very open mind and trying to be really receptive to arguments that might be opposed to your own viewpoint
# [00:22] <mjs> Hixie: she's italicizing - I don't know if it's for emphasis, to mark foreign terms, or to set apart a different mood, except from context
# [00:23] <Hixie> so like if 500 people said "we want X!" and one person said "if you do X you'll open a security hole" or "if you do X we can't implement it", then the one person would win
# [00:23] <Hixie> hyatt: for sure
# [00:23] <Hixie> hyatt: i like to think (and other people have said that this is true, which makes me happier about it) that i am reasonably good at that
# [00:23] <Hixie> the whatwg spec has several examples of things where i was strongly against what we eventually settled on
# [00:23] <jgraham> FWIW I think everyone who has had any involvement with the WHATWG disagrees with some part of the spec
# [00:23] <Hixie> yeah. including me :-)
# [00:23] <schepers> extreme strawmen like that are clear... the question is where you draw the line, obviously
# [00:24] <Hixie> schepers: yup, it's all about judgement calls.
# [00:24] <Dashiva> You could always make a list of camels
# [00:25] <schepers> I don't think it shows very good judgement to defend your position with the most extreme strawman possible
# [00:25] <schepers> and what if that one person who's worried about security is wrong?
# [00:25] <Dashiva> Then the other 500 will point that out
# [00:26] <schepers> like Hixie, I was arguing to the extreme
# [00:26] <Hixie> schepers: i'm not trying to defend my position. i was just giving an example. i feel my work in the whatwg should speak for itself, you should draw your own conclusions from those.
# [00:27] <mjs> if the general sentiment of the group frequently ends up opposed to what seem to be the strongest technical arguments as judged by experts, then we will have a problem
# [00:27] <mjs> hopefully, no strawman scenario will come to pass and this will not be a regular occurrence
# [00:28] <mjs> this this may require some degree of steering of the conversation by the editors, the chairs, and other experts who are willing to be vocal
# [00:28] * jgraham tries to think of an example of successful design-by-consensus outside the sphere of web standards
# [00:29] <mjs> I do think that fully discussing things strengthens both the design and understanding of it, so I don't mind dissent in general
# [00:29] <mjs> I am worried that some people might become perennial gadflies on topics where there is rough consensus and a well-settled idea of the best technical approach
# [00:29] <mjs> but we'll cross that bridge when we come to it
# [00:30] <mjs> I like the IETF motto of "rough consensus and running code"
# [00:30] <hyatt> well thats why i think the editors should have a certain amount of power
# [00:30] <schepers> well, don't stress about it... apparently the editors aren't concerned about consensus anyway, so you can simply ignore dissent, right?
# [00:30] <hyatt> to put into the spec the most compelling technical arguments
# [00:31] <hyatt> and not necessarily the will of the majority
# [00:31] <hyatt> because it could turn out that in the case of the html wg that the majority is not very well-informed
# [00:31] <hyatt> this is my big concern with the html wg actually
# [00:31] <hyatt> the baseline level of competence among whatwg posters seems much higher to me so far
# [00:31] <hyatt> than the html wg
# [00:32] <Dashiva> I think that's fairly obvious to everyone, but some might disagree on just who are the uninformed ones :)
# [00:32] <mjs> I don't think there has been a lot of ignoring of dissent
# [00:32] <hyatt> Dashiva: i'm mainly concerned more about not being informed technically
# [00:32] <mjs> and I certainly hope the editors don't ignore my dissent in the future
# [00:32] <hyatt> Dashiva: people who are technically informed but have very different opinions i don't mind
# [00:32] <mjs> if well-founded
# [00:32] <hyatt> i'm more concerned about say people who can't recognize an obvious security hole for example
# [00:32] <Hixie> in the whatwg there hasn't been any ignoring at all, i reply to every e-mail
# [00:33] <Hixie> (with some exceptions but not for anything about the spec)
# [00:33] <hyatt> the recent CSS opacity thread on www-style is an example
# [00:33] <mjs> I have certainly had mailing list arguments about specs where I pointed out what seemed like major security issues, only to be dismissed with a remark that security is out of scope
# [00:33] <hyatt> most of the people talking in that thread were pretty uninformed about the technical details
# [00:33] <hyatt> and expressing opinions anyway
# [00:35] <hyatt> this is actually why i can't stand telecons or f2fs
# [00:35] <hyatt> because the people who know nothing about a topic rarely have the good sense to be quiet
# [00:35] <hyatt> so you end up wasting huge amounts of time just explaining past work and in the end the person still won't be up to speed enough to adequately participate in the technical discussions that follow
# [00:36] <hyatt> admittedly some areas are much more complex than others
# [00:36] <hyatt> most of html is pretty "simple" when compared with say margin collapsing from css
# [00:36] <hyatt> but some problems require a pretty solid understanding of how stuff works even in html (the residual style problem, adoption agency algorithm, etc.)
# [00:37] <hyatt> Hixie: btw i have a formal proposal for how to deal with pathological nesting by the way
# [00:37] <mjs> some issues relating to the HTML DOM also require deep thinking or at least careful study
# [00:37] <hyatt> Hixie: that i think should be included in the spec
# [00:37] <Dashiva> Is residual style problem part of the html5 spec?
# [00:37] <hyatt> Dashiva: it is covered by the whatwg yes
# [00:38] <Hixie> hyatt: for margin collapsing?
# [00:38] <hyatt> Hixie: no, for html parsing
# [00:38] <Hixie> oh
# [00:38] <Hixie> cool
# [00:38] <Hixie> what's your proposal?
# [00:38] <hyatt> Hixie: i have a couple of suggestions for where some depth caps should be introduced
# [00:38] <Hixie> oh you want to hard code limits in the spec?
# [00:38] <hyatt> i wanted to say that the UA can limit the # of tags it is willing to reopen for residual style
# [00:38] <Hixie> ah you just want to suggest where the limits should be
# [00:38] <Hixie> yeah
# [00:39] <hyatt> but to state that if it does limit it should do so from the inside out
# [00:39] <hyatt> favoring reopening inner ones first
# [00:39] <Hixie> send a mail with your proposal and i can add it to the list of things to do in the parser?
# [00:39] <hyatt> ok
# [00:39] <Hixie> i have a bunch of parser things to fix
# [00:39] <hyatt> my other proposal may be too implementation-specific
# [00:39] <hyatt> but it was to limit the depth of the line box tree for a given line
# [00:39] <hyatt> but that may just be too specific
# [00:39] <jgraham> Hixie: any plans to write up how the new HTML5 elements should be parsed? :)
# [00:40] <hyatt> anyway with those two depth caps, webkit can go to a pretty arbitrary nesting depth
# [00:40] <Hixie> jgraham: yeah, that's part of the things to fix :_)
# [00:40] <hyatt> and still get the correct rendering on all these malformed pages
# [00:40] <hyatt> whereas ffx just chokes and gives up
# [00:40] <Hixie> hyatt: cool
# [00:40] <Hixie> right, bbiab, getting food
# [00:41] <mjs> I just thought of a new t-shirt concept btw:
# [00:41] <mjs> < \__/ >
# [00:41] <mjs> (tag soup)
# [00:42] <Dashiva> We are not amused.
# [00:44] <schepers> speak for yourself, Dashiva
# [00:44] * Quits: Lachy (Lachlan@124.168.27.56) (Ping timeout)
# [00:49] <schepers> so, the problem of having an publicly open wg is the sheer number of people involved, with essentially no constraint on joining... this can lead to people with limited qualifications, as hyatt says... but the real problem of this, in my mind, is that this sets up a situation where the editors can justify abandoning the attempt at consensus, all while claiming that the will of the whole is being met... arguably, this is populist at best, and verging on fas
# [00:49] <mjs> John Boyer's messages are getting increasingly obscure to me
# [00:55] <Hixie> schepers: note that some of hyatt's examples were about the css group
# [00:55] <Hixie> in my experience you get experts and non-experts in equal proportions in w3c wgs whether the group is open or not, to be honest
# [00:56] <hyatt> yeah
# [00:56] <hyatt> i was thinking of the margin collapsing debate at the last css f2f
# [00:56] <hyatt> where 5 people or so deeply understand the problem
# [00:56] <hyatt> and everyone else doesn't get it at all
# [00:56] <hyatt> and yet they participated anyway
# [00:56] <hyatt> and largely derailed the discussion and wasted tons of time
# [00:57] <dbaron> yep
# [00:57] <hyatt> the point being that even with a closed group you have a mixture of experts and non-experts
# [00:57] * Joins: Zeros (Zeros-Elip@69.140.48.129)
# [00:57] <hyatt> so being open doesn't alleviate that problem
# [00:57] <hyatt> err closed
# [00:58] <hyatt> and you cannot give the voice of the non-experts equal weight when building the spec for say margin collapsing
# [00:58] <hyatt> because they don't understand the problem
# [00:58] <hyatt> and would introduce things into the spec that would make no sense
# [00:58] <hyatt> so this is one of the things i see an editor of a spec having to do
# [00:59] <hyatt> they have to make sure they are an expert on the material
# [00:59] <hyatt> since they're going to be writing about it
# [00:59] <mjs> being open also makes it easier for some people who are experts to participate, who might have a hard time under normal W3C process
# [00:59] <hyatt> and then they have to distill the consensus of the technical experts in that particular area
# [00:59] <hyatt> and put that into the spec
# [00:59] <hyatt> wading through all the noise and irrelevant mush to get to the posts from people who actually said something informed and relevant
# [01:00] <hyatt> if an editor is hamstrung by having to worry about what 300-400 people think and waiting for some sort of formal consensus process, then nothing will ever get done
# [01:01] <hyatt> i really hope that the w3c can study how the whatwg has operated
# [01:01] <hyatt> there's a lot to learn there.
# [01:01] <hyatt> more has gotten done in the short lifetime of the whatwg then in any w3c group i've ever seen.
# [01:01] <hyatt> than
# [01:01] <hyatt> primarily because the design principles were just established (and not up for debate) up front
# [01:02] <hyatt> and the editor acts as a sort of benevolent dictator
# [01:02] <hyatt> heck, it's the steve jobs model of spec development.
# [01:02] <hyatt> and it works.
# [01:03] <hyatt> as long as the editor is receptive to sound technical arguments and make the right call way more often than the wrong one.
# [01:03] <mjs> I dunno if I would want SJ to write our spec
# [01:03] <hyatt> mjs: lol
# [01:05] <jgraham> It's worth noting that no-one who has been involved with the WHATWG seems to be complaining about the lack of consensus there
# [01:05] * Joins: Lachy (Lachlan@124.168.27.56)
# [01:05] <hyatt> jgraham: i thought about that
# [01:06] <hyatt> jgraham: i think some of that has to do with the fact that the design principles/overarching philosphy of the whatwg were established up front
# [01:06] <hyatt> and not open to debate
# [01:06] <hyatt> so the people who fundamentally disagreed with the goals/principles just ended up not participating
# [01:06] <mjs> although there are sometimes people who differ wildly in basic design taste
# [01:06] <Dashiva> People seem to have no trouble arguing against the wg charter, on the other hand
# [01:07] <schepers> I think that pretty well exemplifies what I said...
# [01:09] <schepers> (about consensus, that is)
# [01:10] <jgraham> Yeah, I guess some people didn't participate because they didn't see the WHATWG as legitimate. But I haven't seen anyone come along, bring some ideas to the party and then complain about the lack of a consensus process
# [01:17] <Zeros> I know a lot of people who fundamentally disagree with the WHAT WG spec, but they also disagree with the process over there. A dictatorship is historically a very poor way to produce anything that can be perceived as middle ground.
# [01:19] * Quits: billmason (billmason@69.30.57.156) (Quit: .)
# [01:20] * Quits: mw22 (chatzilla@84.41.169.151) (Ping timeout)
# [01:25] <mjs> whatwg's dictatorship is held in check by the potential for the generals to stage a coup d'état if they feel that constitutional principles are violated
# [01:25] <mjs> (to strain the metaphor a bit)
# [01:26] <Zeros> Hixie, Was there any discussion on the whatwg list about transcripts for audio or video?
# [01:27] <schepers> to strain the metaphor even further, that's great for the dictator and the military
# [01:27] <schepers> citizens don't usually fare so well in that kind of situation
# [01:28] <mjs> except that there were no border guards posted, and yet the population kept growing
# [01:29] * mjs likes metaphors way too much, can you tell?
# [01:29] <mjs> Zeros: Apple will be proposing some things about media accessibility soonish
# [01:29] * schepers sensed that
# [01:30] <Zeros> mjs, sweet. We need some kind of transcript support for 508. I was thinking an attribute on the video and one of the <source> like a longdesc. Curious to see what Apple has in mind.
# [01:30] <Zeros> on the
# [01:31] <mjs> Zeros: our thinking is that media files often have text tracks that provide optional captions, or in some case there is alternate video that has burned-in captions
# [01:32] <Zeros> hmm, that seems like an alternative, but for someone posting third party video that may not be an option
# [01:32] * Joins: myakura (myakura@60.239.122.32)
# [01:33] <Zeros> Particularly, if we think about something like YouTube (not that they're going to transcript everything), but the user submitted videos which are converted to flv can't be guaranteed to have any kind of CCs
# [01:34] <hyatt> hmmm should i keep responding to tina
# [01:34] <Zeros> In discussions with government employee about accessibility that had a similar problem, they had to add text transcripts and link to them for all the user submitted video
# [01:35] <Zeros> A formal transcript of some such would add some more meaning, and let a regular user who got the video loaded too access it separately
# [01:35] * Quits: tH (r@87.102.32.222) (Quit: ChatZilla 0.9.78.1-rdmsoft [XULRunner 1.8.0.9/2006120508])
# [01:38] <Zeros> hyatt, webkit supports kerning?
# [01:38] <hyatt> no it doesn't
# [01:38] <hyatt> it could trivially
# [01:38] <hyatt> but it's too large a performance hit
# [01:38] <hyatt> basically we have a super-fast text rendering code path
# [01:38] <hyatt> and then the slower general code path
# [01:38] <Zeros> Oh okay. You commented that you kept it "disabled", so I wasn't sure.
# [01:39] <hyatt> the slower path is quite capable of doing kerning and ligatures
# [01:39] <hyatt> but we don't want to fall into that path in normal rendering
# [01:39] <mjs> has anyone but her asked for kerning?
# [01:40] <mjs> it certainly hasn't been on the top request list of any web developers I've talked to
# [01:40] <mjs> Zeros: yeah, not sure if the proposal considers externally attached captions
# [01:41] * Quits: gavin_ (gavin@74.103.208.221) (Ping timeout)
# [01:41] <Zeros> mjs, It was mentioned on a couple blog entries a few weeks ago. I think someone mentioned the WIRED logo as well. People I talked to about it generally said it was nifty, but no one thought it was in dire need of being added to CSS.
# [01:42] <Zeros> mjs, Ok. I'll write something up and suggest it to the whatwg list then. Any idea how long Apple will take? I'd wait until after their proposal to get a better idea of where they stand if possible.
# [01:44] <mjs> Zeros: kerning isn't really a CSS issue (unless you wanted people to hand-kern their text - I'd guess what post people want is just to respect kerning pairs already in fonts)
# [01:44] <Zeros> hyatt, Is that what the ATSU setting is for in the Debug menu? (the code path)
# [01:44] <mjs> Zeros: I've got a draft in my inbox to review
# [01:44] <Zeros> okay, cool
# [01:45] <hyatt> Zeros: yeah although we have kerning disabled in the slow path too
# [01:45] <hyatt> for consistency
# [01:45] <hyatt> basically we're capable of doing kerning but don't because of perf
# [01:45] <hyatt> it's like a 25% speed hit to use ATSU over our fast code path
# [01:45] <Zeros> ouch
# [01:46] <hyatt> there might be a clever way to cache kerning pairs
# [01:46] <hyatt> for the fast code path
# [01:46] <hyatt> but honestly very few people have asked for kerning
# [01:46] <hyatt> most people don't even notice it's not there
# [01:46] <hyatt> typography weenies notice of course but they are a minority
# [01:47] * Dashiva looks up kerning in the dictionary
# [01:47] * Joins: gavin_ (gavin@74.103.208.221)
# [01:48] <karl> Dashiva: http://en.wikipedia.org/wiki/Kerning
# [01:50] * Quits: jgraham (jgraham@81.179.93.10) (Ping timeout)
# [01:51] <mjs> actually I think our most visible typography weenie glitch is lack of support for ligatures in latin text
# [01:51] <mjs> so you don't get the nice version of ffl in afflicted
# [01:52] <mjs> though you can see a difference in "Vast" if you look really close
# [01:52] <mjs> (the V and a should be kerned tighter)
# [01:52] <Zeros> Do any browsers implement ligatures?
# [01:53] <Hixie> opera does
# [01:53] <Hixie> easy way to check on Mac is to use the Zapfino font and include the word Zapfino
# [01:53] <Hixie> because that font defines a 7-letter ligature for its name
# [01:53] <karl> œ <- oe
# [01:54] <mjs> oe has a special glyph though - few fonts make it an automatic ligature
# [01:54] <hyatt> i expect firefox 3 might to kerning/ligatures
# [01:54] <hyatt> do
# [01:54] <mjs> we could do it if we found a way to do it really fast
# [01:54] <hyatt> anyway not doing these things is one of the reasons we're so fast
# [01:54] <karl> bœuf = beef
# [01:55] <karl> œuf = egg
# [01:55] <karl> :)
# [01:55] <Dashiva> whitebox + 'uf'?
# [01:56] * Joins: sbuluf (ztmqx@200.49.140.38)
# [01:56] <karl> the white is "oe"
# [01:56] <karl> http://en.wikipedia.org/wiki/%92
# [01:56] <karl> rha :)
# [01:56] <Dashiva> I was more wondering about the uf. Is it Dutch or something?
# [01:57] <karl> wikipedia for % C5 % 92
# [01:57] <karl> (putting space for avoiding colors)
# [01:57] <Zeros> Opera's highlighting of text with ligatures is awkward
# [01:58] <mjs> ligatures do lead to selection oddities too
# [01:58] <Dashiva> ah, French
# [01:59] * Joins: kazuhito (kazuhito@210.232.34.13)
# [02:00] <schepers> why do French people only have a single egg for breakfast in the morning? because one egg is un oeuf
# [02:00] * schepers signals lafftrak
# [02:03] <dbaron> karl, In Danish and Norwegian æ is considered a separate letter, not just a + e. Not sure if that's true for œ anywhere. But I don't think of those as ligatures.
# [02:06] <karl> ah david, I have to check this for French
# [02:06] <karl> Œthel (pronounced /eðəl/) is a Roman script letter (Œ, œ) used in medieval and early modern Latin, and in modern French, and also the vowel sound it represents. The letter is a ligature of o and e. In Old English the name was spelled eðel. Another name for this symbol is œgule.
# [02:06] <karl> "The letter is a ligature of o and e."
# [02:06] <dbaron> Icelandic too
# [02:07] <karl> "In French, œ (called e dans l'o, meaning "e in o" but also a pun as it sounds like œufs dans l'eau, meaning "eggs in water") is a true linguistic ligature, not just a typographic one (like the fi or fl ligatures), reflecting etymology."
# [02:08] <hyatt> i find it stunning that people want to have a standard that does not reflect reality
# [02:08] <hyatt> what is the point of that
# [02:08] <h3h> advancing obscure business agendas
# [02:10] <Zeros> hyatt, I'd be perfect, as with most things that exist outside reality.
# [02:11] <Philip`> (Did you mean "It'd be perfect"?)
# [02:11] <Zeros> err yes
# [02:12] <karl> hyatt: because reality is not one universal thing. So I guess there are common things we tend to agree with in a given community, and there are border effects. On top of that you can add general conception of the world.
# [02:12] <h3h> that sounds reasonable
# [02:13] <Zeros> iTunes isn't using Webkit for the store is it?
# [02:13] <h3h> I think what hyatt (and others) are reacting to is when people reject the majority
# [02:13] <karl> There is in France a dictionary of the French Academy of Language defining words and there is the dictionary from commercial publishers, and there is the language as it is spoken in the street.
# [02:13] <karl> In fact I should say languages
# [02:13] <karl> because we do not talk the same language in the country side, in the suburb or in a bourgeois environment
# [02:13] <karl> Now the question is "what is reality?" ;)
# [02:14] <Zeros> karl, There's a big difference depending on who you're talking to and where in the states too ;)
# [02:14] <Zeros> Do you have pop vs soda over there?
# [02:14] <karl> exactly
# [02:16] <h3h> eh
# [02:16] <h3h> black vs. white is much different from dialects of a language
# [02:16] <h3h> HTML is certainly the latter
# [02:16] <h3h> and there is certainly a core which should be obviously necessary
# [02:16] <Zeros> Is there some way we can get everyone to discuss the design principals and come to agreement on those before we go off on a sidebar about ?
# [02:16] <h3h> yet some people reject it as if it were black vs. white
# [02:16] <hyatt> i guess the philosophical difference can be summed up as follows
# [02:17] <Zeros> It seems the biggest factor contributing to making this group so slow is how quickly conversation gets sidetracked
# [02:17] <hyatt> i would like a specification that helps me render the current web (while also adding new features for html5)
# [02:17] <hyatt> but others would just like to stay strict
# [02:17] <hyatt> both points of view are valid
# [02:17] <h3h> "stay strict"?
# [02:17] <hyatt> but i'm not that interested in being involved with/working on a spec that is just a crippled subset of the real world web
# [02:18] * Quits: jmb (jmb@81.179.74.126) (Ping timeout)
# [02:18] <hyatt> so much value has been gained from the whatwg defining holes and helping moz/opera/safari achieve greater interoperability
# [02:18] <zcorpan> hyatt: in the worst case, the whatwg spec could still be a superset of the htmlwg spec :)
# [02:18] <hyatt> filling in holes
# [02:18] <hyatt> i dunno, i just look at specs and see holes and it drives me crazy
# [02:19] <hyatt> like html4
# [02:19] <hyatt> which is almost a non-spec it;'s so vague
# [02:19] <hyatt> e.g., why hasn't table layout been defined
# [02:19] <hyatt> whether it's css2 or html, somebody should have defined it
# [02:19] <hyatt> i don't care if it's hard
# [02:19] <hyatt> instead both specs just handwave and give up
# [02:19] <hyatt> that's not useful to anybody
# [02:19] <hyatt> you can't implement tables looking at either the css or html specs
# [02:20] * Joins: jmb (jmb@81.86.70.47)
# [02:20] * zcorpan adds tables to his todo list
# [02:20] <Philip`> zcorpan: We could have one version of the spec with some CSS to generate the stripped-down HTMLWG subset, with any controversial browser-conformance bits cut out to stop people complaining :-)
# [02:21] <zcorpan> hyatt: is defining quirks mode more urgent than defining table layout? :)
# [02:21] <hyatt> not sure you can really define quirks mode
# [02:21] <hyatt> since by definition browser vendors use it to add quirks they need
# [02:21] <hyatt> possibly for internal stuff etc.
# [02:21] <hyatt> there are some common cross-browser quirks i guess though
# [02:21] <hyatt> that could be defined
# [02:22] <zcorpan> yeah, i mean the quirks that would need to be implemented by new vendors to render the web
# [02:22] <zcorpan> like width:300;
# [02:24] <zcorpan> but i guess tables is more urgent given that it probably should get into css21
# [02:25] <karl> [09:13] <hyatt> i would like a specification that helps me render the current web (while also adding new features for html5)
# [02:25] <karl> [09:13] <hyatt> but others would just like to stay strict
# [02:25] <karl> you just express I think one of the problems
# [02:25] <karl> from your community point of view (browsers), you want to render the web. perfectly valid and logical.
# [02:26] <hyatt> i suppose it might be possible t ocreate a single document that is both the html5 spec from the w3c and extras from the whatwg
# [02:26] <karl> from some other people, they want to edit the web. They view it from a document point of view
# [02:26] <hyatt> and the html5 doc could be strict and leave out everything that is nasty
# [02:26] <hyatt> but the whatwg doc could define the rest
# [02:26] <hyatt> however i think that would just result in people viewing the whatwg doc as the real one
# [02:26] <Hixie> i wouldn't be editing the w3c version if that happened
# [02:26] <hyatt> and ignoring the w3c one
# [02:26] <karl> I think it might be possible to accomodate both
# [02:27] <karl> for each feature, how do I use it, how do I view it, how do I author, etc.
# [02:28] <karl> If the group is focusing on rendering, there will be high frictions I think, or people will leave and stop participating.
# [02:29] <zcorpan> we can have strict document conformance requirements at the same time as having ua conformance requirements that can handle real web content
# [02:30] <karl> zcorpan: yes class of products.
# [02:30] <h3h> I don't think the rendering stuff should be up for debate -- it's not a subjective document
# [02:30] <h3h> it's normatively descriptive based on evidence and research
# [02:30] <h3h> unlike the new features for HTML 5, which are certainly subjective and up for debate
# [02:31] <karl> s/evidence// <- too much fuziness ;) but verifiable research is a valid point
# [02:31] <h3h> again, I stress the conspicuous separation between these two goals
# [02:31] <Zeros> h3h, Isn't it still up for debate in the whatwg provided a good enough argument is made?
# [02:32] <h3h> Zeros: no... how can it be up for debate whether IE render something like X or like Y?
# [02:32] <h3h> it just does one or the other
# [02:32] <h3h> s/render/renders/
# [02:32] <h3h> the rest (when to use IE's behavior over other browsers) is up to the browser vendors
# [02:32] <karl> As I said a couple(?) of months ago. the parsing algorithm is a good thing. but it looks like, from my point of view, as an implementation guide for reading the web more than the semantics itself of the language.
# [02:33] <h3h> karl: assuming those two things are different
# [02:33] <mjs> I wish web technologists studied linquistics more
# [02:33] <mjs> in particular the descriptivist/prescriptivist debate
# [02:33] <h3h> is French the language defined by the Academie back in 1790 or is it what's spoken today?
# [02:33] <karl> :)
# [02:33] <h3h> mjs: indeed.
# [02:34] <mjs> I can totally see some people getting bent out of shape by split infinitives
# [02:34] <karl> h3h: still in never ending production ;)
# [02:34] <h3h> karl: right.
# [02:34] <h3h> and so is the way with HTML
# [02:34] <h3h> whether or not it's in the spec
# [02:34] <h3h> so it should be in the spec
# [02:34] <Zeros> h3h, where does html5 define IE's behavior like that?
# [02:34] <h3h> (says I and others)
# [02:34] <h3h> Zeros: the parsing section is largely based on IE's behavior
# [02:35] <mjs> it's based on a mix of all the browsers actually
# [02:35] <mjs> studying various cases
# [02:35] <h3h> Zeros: as other browsers in quirks mode are largely based on IE's behavior
# [02:35] <h3h> right, but the majority seat going to IE
# [02:35] <h3h> anyway, not the point
# [02:35] <mjs> it diverges from IE mostly when the IE result is crash, hang, bad document or non-tree-structured DOM
# [02:35] <h3h> the point is that it's verifiable research based on browser behavior
# [02:35] <mjs> all of which are fairly catastrophic consequences
# [02:35] <h3h> it's not subjective feature suggestions
# [02:36] <Philip`> Do a significant (though small) number of authors look at the spec and read the document conformance requirements? Does a more significant number listen to advocates who say and and table layouts are bad and deprecated and you should use XHTML and CSS instead? If the latter, then nice strict document conformance might be more of a marketing issue than a spec-writing issue
# [02:37] <hyatt> table layouts are usable from CSS
# [02:37] <hyatt> tables are not deprecated as a form of layout
# [02:37] <h3h> here's where I think the tutorials come in, Philip`
# [02:37] <Philip`> Ah, I meant <table> layouts
# [02:37] <h3h> the majority of web authors writing HTML are going to grab snippets from random tutorial sites
# [02:37] <hyatt> if the w3c wants to define HTML5 as a pure strict language, that's ok. it's a valid point of view.
# [02:37] <h3h> so if there's an authoritative and useful tutorial (or set of tutorials), the language has a better chance of being used "correctly"
# [02:37] <hyatt> it's not really the document i'm interested in though
# [02:38] <schepers> no, the majority of web authors writing HTML *by hand* are going to grab snippets from random tutorial sites... most people will be using blog software, wikis, myspace, and other authoring tools
# [02:39] <h3h> yes
# [02:39] <h3h> sorry, assumption
# [02:39] <mjs> Philip`: I have no problem with making conformance for documents stricter than what implementations are required to accept, and indeed Web Apps 1.0 already heavily relies on this
# [02:39] <karl> hyatt: why not calling it WBIG then ? Web Browser Implementation Guide?
# [02:39] <karl> :)
# [02:39] <h3h> I don't worry as much about authoring tools like that because they're very narrow points that can be influenced much more easily than the mass mob
# [02:39] <hyatt> karl: heh
# [02:39] <karl> just noticed, that it could be We Big
# [02:39] <karl> :p
# [02:39] <Zeros> h3h, I'm not sure I understand that. You're advocating making quirks mode in browsers into the standards mode because it agrees with IE?
# [02:40] <mjs> but there's a limit to how much you want to take widely used things out of conformance
# [02:40] <h3h> Zeros: I'm not advocating anything. really, read the spec
# [02:40] <mjs> go too far, and authors won't make the effort to conform to the more reasonable requirements
# [02:40] <hyatt> you know i just thought of a really good point
# [02:41] <hyatt> which i will just phrase as antagonistically as possible. the browser vendors of the past were stupid. the browser vendors of the present aren't.
# [02:41] * Parts: hasather (hasather@81.235.209.174)
# [02:41] <hyatt> those of us that are involved now are not the people who did <blink> or
# [02:42] <hyatt> i think there is some bias being displayed against mistakes of the past made by browser vendors who were not as aware of the design philosophy of html as the current vendors are
# [02:42] <hyatt> e.g., the early netscape hackers (none of whom work on browsers any more)
# [02:42] <schepers> (or might be, but have hopefully learned from their mistake?)
# [02:42] <hyatt> yeah these aren't stupid people, i just thought i'd phrase it that way for dramatic effect :)
# [02:42] <schepers> I approve
# [02:43] <Zeros> hyatt, mistakes will be made now too. And in 8 years I'm sure similar resentments will be made about some HTML5 features.
# [02:43] <hyatt> well not the same level of mistakes though
# [02:43] <hyatt> css exists now
# [02:43] <Zeros> that I'll agree with
# [02:43] <hyatt> so there is an awareness of how to separate presentation from content
# [02:43] <hyatt> etc.
# [02:43] <hyatt> a lot of these "mistakes" stemmed from css not existing
# [02:43] <hyatt> it does now
# [02:43] <schepers> hyatt: I think some of the bitterness comes from a lack of innovation or new features from browsers (specifically IE) for the last 5-8 years
# [02:43] <hyatt> so those kinds of presentational mistakes won't be repeated
# [02:43] <mjs> ok I think I need to take any further conversation with Tina off-list
# [02:44] <hyatt> mjs: i'm not sure why she used the term "belitttled"
# [02:44] <hyatt> i haven't seen any belittling taking place
# [02:44] <mjs> she scolded the tone I used in scolding someone else for their tone
# [02:44] <schepers> shame on you, mjs
# [02:45] <schepers> I'm very hopeful that with IE on board now, we can look forward to real progress on the Web
# [02:45] <schepers> maybe I'm an optimist
# [02:46] <schepers> other browsers will no longer be held back by IE, if they act in good faith
# [02:46] <h3h> and if people on the list don't drive Microsoft off of it
# [02:46] <h3h> (or any other vendor for that matter)
# [02:46] <schepers> h3h: yes
# [02:47] <h3h> I don't understand the point of Tina's email. it's devoid of useful content
# [02:47] <Zeros> schepers, Chris seems optimistic they'll be able to implement whatever we come up with.
# [02:47] <Zeros> I think me made a comment to the effect of "ignore what Trident may or may not be capable of"
# [02:48] <Zeros> s/me/he/
# [02:48] * Quits: h3h (bfults@66.162.32.234) (Quit: |)
# [02:49] <karl> hmm time for a coffee.
# [02:49] * Quits: karl (karlcow@128.30.52.30) (Quit: Where dwelt Ymir, or wherein did he find sustenance?)
# [02:51] <Philip`> "I can recall 0 Safari bug reports requesting removal of any presentational tags or attributes" - but there's https://bugzilla.mozilla.org/show_bug.cgi?id=124699
# [02:52] <hyatt> Philip`: we don't support <blink>
# [02:52] <hyatt> in webkit.
# [02:52] <hyatt> i don't think <blink> is a useful example, since you are conformant if you don't blink at all
# [02:52] <hyatt> so there is no need to put it in any standard
# [02:53] <hyatt> whereas cutting out would break the real web
# [02:53] <hyatt> it's used everywhere.
# [02:53] <hyatt> again, i think it is a perfectly valid point of view to want to specify a pure uncorrupted html5.
# [02:53] <hyatt> *but* i think what browser vendors are interested in is a more complete specification
# [02:54] <Philip`> (and 63458, 148365, 162300, etc)
# [02:54] <hyatt> i do not view these goals as incompatible necessarily as long as the spec labeled the "poor practice" tags as such
# [02:54] <hyatt> i guess i see value in specifying even the badly-designed stuff
# [02:54] <hyatt> the spec can even say "this is badly designed"
# [02:54] <hyatt> or "poor practice"
# [02:54] <hyatt> and should
# [02:54] <Zeros> We have to specify the badly designed stuff
# [02:54] <hyatt> but having the badly designed stuff in there allows us to define how it works
# [02:54] <hyatt> and to make sure we have a language that works with the existing web
# [02:55] <Hixie> i'm going to specify the badly designed stuff
# [02:55] <Hixie> and i'm only writing one spec
# [02:55] <hyatt> i just see such enormous value in specifying the real web
# [02:55] <Zeros> That was that was the whole point. If some third party comes along in 2010 and wants to implement a new browser they have a formal spec to implement
# [02:55] <hyatt> and not a subset of the real web
# [02:55] <Hixie> exactly
# [02:55] <mjs> Philip`: that's not a Safari bug report - I guess Mozilla attracts more standards enthusiasts
# [02:55] <hyatt> mjs: if we supported blink we'd have gotten a bug report too :)
# [02:55] <hyatt> mjs: guaranteed
# [02:55] <hyatt> oh you no what?
# [02:55] <hyatt> rofl
# [02:56] <hyatt> we have an internal radar to support blink
# [02:56] <hyatt> hahahahah
# [02:56] <hyatt> you can't win,.
# [02:56] <Hixie> hah
# [02:56] <hyatt> know what
# [02:56] * hyatt can't type today
# [02:56] <Zeros> nice
# [02:56] <hyatt> we have a bug to support it
# [02:56] <hyatt> if we added it we'd get a bug to remove it
# [02:56] <hyatt> lol
# [02:56] <mjs> hyatt: I wish we could somehow Cc Tina on that bug
# [02:56] <Zeros> ugh, Please don't support it
# [02:57] <Zeros> Might as well add background-color: seizure
# [02:57] <mjs> do we have any bug reports requesting removal of <marquee>?
# [02:57] <hyatt> mjs: i don't think so
# [02:57] <mjs> <blink> is just too much of a symbol of browser lameness
# [02:57] <hyatt> mjs: you can disable all marquees easily though from a user stylesheet
# [02:57] <mjs> compared to its value as easter egg or otherwise
# [02:57] <mjs> (for us anyway, for Mozilla I guess it is historic)
# [02:58] <mjs> (then again, they don't have a <layer> easter egg afaik)
# [02:58] <Philip`> mjs: Safari just needs more bugs - Mozilla is winning by a factor of thirty :-)
# [02:58] <mjs> Philip`: we have a lot more internally than in the public bugzilla
# [02:58] <Zeros> Philip`, unfortunately Gecko's float model is finally fixed in FF3 (since they merged reflow)
# [02:58] <Zeros> And I think they added inline-block
# [02:59] <Zeros> Webkit has a ways to go
# [03:01] <hyatt> ways to go with what, bug reports?
# [03:02] <Zeros> hyatt, Getting it compliant with CSS2, fixing a lot of outstanding layout bugs, etc.
# [03:02] <hyatt> i think webkit is ahead of gecko at this point in many areas
# [03:02] <hyatt> note i mean webkit and not safari 2.
# [03:02] <Zeros> It is, but its way behind in others.
# [03:02] <hyatt> like what?
# [03:03] <Hixie> mathml support, for one
# [03:03] <Zeros> hyatt, The table display properties have a lot of bugs
# [03:03] <hyatt> have you used the latest webkit?
# [03:03] <Zeros> Some are rather bizarre
# [03:03] <hyatt> many table display property issues have been fixedi n the last 2.5 years since safari 2.0.
# [03:03] <hyatt> safari 2 is over 2 years old
# [03:03] <hyatt> webkit is like a whole new engine when compared with safari 2.
# [03:03] <Zeros> I know, I'm using Webkit
# [03:03] <hyatt> file bugs then.
# [03:03] <Zeros> I have
# [03:03] <hyatt> would be helpful
# [03:04] <hyatt> in bugzilla?
# [03:04] <Zeros> yes
# [03:04] * hyatt hasn't seen any table display prop bugs recently
# [03:04] <Zeros> oh, let me get it :)
# [03:05] <Zeros> You also commented about the dropped footer bug being related to the table display types
# [03:05] <Philip`> http://bugs.webkit.org/show_bug.cgi?id=13538 ?
# [03:05] <Zeros> http://bugs.webkit.org/show_bug.cgi?id=12691 Very very weird
# [03:05] <Zeros> Increasing and decreasing the font duplicates all the generated content
# [03:05] <hyatt> oh i know whats up with that
# [03:06] <hyatt> darin is actually working on that duplicate issue irght now
# [03:06] <hyatt> generated content that gets wrapped in anonymous containers can get lost
# [03:06] <hyatt> since the code to try to find the generated content to update it when styles change doesn't drill into anonymous wrappers
# [03:08] <Zeros> hyatt, I wish I could show you the admin interface for the IceWarp mail server too.
# [03:08] <hyatt> Zeros = elliot sprehn?
# [03:08] <Zeros> yes
# [03:08] <Zeros> The interface is incredibly broken in Webkit
# [03:08] <hyatt> ah cool hi
# [03:08] <Zeros> hey :)
# [03:08] <hyatt> that bug should be fixed very soon
# [03:08] * Quits: kingryan (rking3@66.92.187.33) (Quit: kingryan)
# [03:08] <hyatt> the generated content one
# [03:08] <Zeros> sweet, thanks
# [03:08] <hyatt> someone is working on it
# [03:09] <hyatt> generated content + table display types = wow obscure :)
# [03:09] <hyatt> since neither work in winie, you don't tend to see those :)
# [03:09] <hyatt> but there are other areas where we're ahead of gecko
# [03:09] <hyatt> e.g., inline-block, floats
# [03:09] <hyatt> ffx3 has changed a lot to be more like us in that regard
# [03:09] <hyatt> it's been exciting watching that
# [03:09] <hyatt> since hopefully ffx's market share will help the sites change
# [03:10] <hyatt> to be more like the standard when ffx3 comes out
# [03:10] <mjs> does anyone here think I was rude to Tina on the list?
# [03:10] <hyatt> but yeah i think gecko probably does better on a lot of things by virtue of being older and receiving wider testing
# [03:10] <Hixie> mjs: no
# [03:10] <hyatt> we'll get there though.
# [03:11] <Hixie> mjs: do you have a specific e-mail in mind?
# [03:11] <hyatt> anyway, it's not a competition (except when it is) ;)
# [03:11] <mjs> I wrote her privately and told her she sounded overheated and frankly somewhat rude, and may want to try to present her arguments more calmly
# [03:11] <mjs> and she said I was the one who had been incredibly rude
# [03:11] <mjs> just trying to get a second opinion in case I am delusional here
# [03:12] <Zeros> hyatt, Wasn't trying to insult Webkit or your work, gecko has definitely had more time to become mature, and Webkit has come a long long way since Safari 2
# [03:13] <Hixie> mjs: i'd ask for a pointer so that we can make an objective judgement
# [03:14] <Zeros> hyatt, One of the things I think that needs addressing is how plugins sleep in Webkit. Keeping a window in the background too long, minimizing it or clicking on a different tab has a tendency to timeout network connections. I think I remember you commenting that was a feature?
# [03:14] <hyatt> we no longer throttle background windows down
# [03:14] <hyatt> like we do in safari 2
# [03:15] <Philip`> http://krijnhoetmer.nl/irc-logs/html-wg/20070416#l-375 - "<anne> expected total this month: 1800+" - sadly expectations have not been met, with a mere 1773 messages
# [03:16] <hyatt> killed my whole afternoon on this channel/mailing list lol
# [03:16] <Zeros> poor hyatt
# [03:16] <hyatt> did you reduce that canvax positioning thing yet?
# [03:16] <hyatt> canvex
# [03:17] <mjs> Hixie: I think she was just referring to my recent emails on public-html, and I to hers
# [03:18] <Zeros> hyatt, no, had other things going on. Its some kind of mix between the svg, and the absolute positioning from where I got I think. If the svg doesn't load it displays properly
# [03:18] <Hixie> i don't think any of those mails were especially constructive, but they weren't really rude
# [03:19] * hyatt got in a few digs over the crappy css3 font stuff
# [03:19] <hyatt> so hyatt is happy
# [03:19] <Zeros> hyatt, Let me take a crack at that right now
# [03:19] <Zeros> Philip`, did you could www-html too?
# [03:19] <hyatt> Zeros: it's probably something easy
# [03:19] <Zeros> count*
# [03:20] <Philip`> Is that if any of the SVGs don't load, or only specific ones? (The bottom one is the only one whose extends outside the view box, but I don't know if that matters)
# [03:20] <Philip`> *whose contents extend
# [03:20] <Philip`> *whose content extends
# [03:21] <hyatt> ok dinner time
# [03:21] * Quits: hyatt (hyatt@24.6.91.161) (Quit: hyatt)
# [03:21] <mjs> I found her referring to my comment as "condescending and childish" to be somewhat rude, but perhaps I was oversensitive
# [03:22] <Hixie> no more rude than your equivalent comment to her
# [03:22] <mjs> the comment she replied to wasn't even to her
# [03:23] <Philip`> Zeros: I didn't - I guess that would bump it up to 2018, but it's cheating since it's counting some XHTML and XBL discussions too
# [03:25] <Philip`> (public-xhtml2: 37 messages. Are they still doing a lot of work in private, or are they not doing any work that requires communication at all?)
# [03:26] <Hixie> that's about the volume they were getting on their private list
# [03:26] <Hixie> last e-mail to their private list was on Wednesday, 25 April
# [03:27] <Hixie> the 2006 and 2005 quarters had the following volume of e-mail for their private list (Q4 2006 to Q1 2005): 86 211 107 189 73 97 260 171
# [03:28] <Hixie> so i guess this is lower than usual, but they just went through a rechartering, so it's not that surprising
# [03:29] <Philip`> Ah, okay - sounds like it's just a naturally quieter life than in public-html
# [03:30] <mjs> it's quieter outside the spotlight
# [03:34] * Parts: zcorpan (zcorpan@217.211.77.236)
# [03:36] * Joins: zcorpan (zcorpan@217.211.77.236)
# [03:40] * Parts: zcorpan (zcorpan@217.211.77.236)
# [03:47] * Quits: mjs (mjs@17.255.99.124) (Quit: mjs)
# [03:49] * Quits: gavin_ (gavin@74.103.208.221) (Ping timeout)
# [03:50] * Joins: Shunsuke (Shunsuke@133.27.59.8)
# [03:51] * Quits: marcos (chatzilla@131.181.148.226) (Ping timeout)
# [03:51] * Quits: myakura (myakura@60.239.122.32) (Ping timeout)
# [03:55] * Joins: gavin_ (gavin@74.103.208.221)
# [04:05] * Quits: Shunsuke (Shunsuke@133.27.59.8) (Ping timeout)
# [04:07] * Joins: marcos (chatzilla@131.181.148.226)
# [04:22] * Joins: owner (chatzilla@70.181.71.135)
# [04:30] * Joins: Shunsuke (Shunsuke@133.27.53.221)
# [04:33] * Joins: DanC_lap (connolly@128.30.52.30)
# [04:34] * Joins: hyatt (hyatt@24.6.91.161)
# [04:38] * Quits: Shunsuke (Shunsuke@133.27.53.221) (Ping timeout)
# [06:40] * Disconnected
# Session Close: Wed May 02 00:00:00 2007

The end :)