/irc-logs / w3c / #html-wg / 2007-04-16 / end

Options:

# Session Start: Mon Apr 16 00:00:00 2007
# Session Ident: #html-wg
# [00:04] * Philip is now known as Philip`
# [00:08] * Quits: Sander (svl@80.60.87.115) (Quit: And back he spurred like a madman, shrieking a curse to the sky.)
# [02:08] * Disconnected
# [02:08] * Attempting to rejoin channel #html-wg
# [02:08] * Rejoined channel #html-wg
# [02:08] * Topic is 'W3C HTML WG http://www.w3.org/html/wg/ - http://krijnhoetmer.nl/irc-logs/ (logged) - http://esw.w3.org/topic/HTML/ProposedDesignPrinciples'
# [02:08] * Set by anne on Tue Mar 27 12:28:46
# [02:14] * Joins: marcos (chatzilla@131.181.99.125)
# [02:17] * Joins: edas (edaspet@88.191.34.123)
# [02:17] * Quits: edas (edaspet@88.191.34.123) (Quit: http://eric.daspet.name/ et l'édition 2007 de http://www.paris-web.fr/ )
# [02:17] * Joins: Zeros (Zeros-Elip@69.140.48.129)
# [02:34] * Quits: marcos (chatzilla@131.181.99.125) (Ping timeout)
# [02:46] * Joins: Shunsuke (kuruma@219.110.80.235)
# [03:03] * Joins: olivier (ot@128.30.52.30)
# [03:40] * Quits: Shunsuke (kuruma@219.110.80.235) (Connection reset by peer)
# [03:54] * Quits: hober (ted@69.45.6.105) (Quit: ERC Version 5.2 (IRC client for Emacs))
# [04:44] * Quits: Zeros (Zeros-Elip@69.140.48.129) (Quit: Leaving)
# [05:00] * Quits: gavin_ (gavin@74.103.208.221) (Ping timeout)
# [05:04] * Joins: gavin_ (gavin@74.103.208.221)
# [05:15] * Joins: MikeSmith (MikeSmith@mcclure.w3.org)
# [05:45] * Joins: marcos (chatzilla@131.181.211.113)
# [05:47] * Disconnected
# [05:50] * Attempting to rejoin channel #html-wg
# [05:50] * Rejoined channel #html-wg
# [05:50] * Topic is 'W3C HTML WG http://www.w3.org/html/wg/ - http://krijnhoetmer.nl/irc-logs/ (logged) - http://esw.w3.org/topic/HTML/ProposedDesignPrinciples'
# [05:50] * Set by anne on Tue Mar 27 12:28:46
# [05:54] * Joins: marcos_ (chatzilla@131.181.148.226)
# [05:55] * Quits: marcos (chatzilla@131.181.211.113) (Ping timeout)
# [05:55] * marcos_ is now known as marcos
# [06:05] * Joins: myakura (myakura@60.239.122.32)
# [06:16] * Joins: Shunsuke (kuruma@133.27.53.98)
# [06:32] * Joins: Zeros (Zeros-Elip@69.140.48.129)
# [06:44] * Quits: marcos (chatzilla@131.181.148.226) (Ping timeout)
# [06:47] * Quits: Lachy (Lachlan@124.168.27.56) (Ping timeout)
# [06:54] * Joins: anne (annevk@131.181.148.59)
# [06:54] <anne> whoa
# [06:54] <anne> I'm in Australia!
# [06:54] * Joins: htmlr (htmlr@203.206.237.84)
# [06:56] <MikeSmith> anne - welcome to the other half of the world
# [06:57] <MikeSmith> tbe better half
# [06:57] * Joins: marcos_ (chatzilla@131.181.211.113)
# [06:57] * marcos_ is now known as marcos
# [06:57] <anne> it's the warmer part anyway
# [06:57] <anne> not sure about better
# [06:57] * anne comments on the t-shirt post
# [06:58] * anne ..
# [07:01] * mjs snickers
# [07:07] * Quits: gavin_ (gavin@74.103.208.221) (Ping timeout)
# [07:08] * Quits: Shunsuke (kuruma@133.27.53.98) (Ping timeout)
# [07:12] * Joins: gavin_ (gavin@74.103.208.221)
# [07:16] * anne mumbles something about +1 messages
# [07:17] <anne> heh
# [07:17] <anne> marcos seems to have teached gmail to filter out stupid messages on public-html :)
# [07:17] <Zeros> Need a design principal that asks emails to be of substance and no +1 "I agree" messages eh?
# [07:18] <anne> something like a "HOWTO write an e-mail"
# [07:18] <marcos> let me see... we have Mike (not smith), <indent> vs. <blockquote> thread, etc
# [07:18] <anne> or "E-mail for dummies"
# [07:18] <marcos> Lachy has written one of those
# [07:19] <marcos> I think he even points to an RFC
# [07:20] <marcos> http://www.ietf.org/rfc/rfc1855
# [07:22] <Zeros> The <indent> vs. <blockquote> has been, um, interesting...
# [07:26] <sbuluf> proposals in emails could be also posted to say some wiki, and +1's could be added there. probably a better model. current number of +1's and -1's could also additionally be added to the topic of echa proposal.
# [07:27] <sbuluf> Re: <indent> vs. <blockquote> [+12, -5]
# [07:28] <anne> the use case for <indent> seems to be addressed by
# [07:28] <anne> but maybe you want some kind of block level variant that sets text apart
# [07:28] <Zeros> The number of ±1 in emails isn't that important though sbuluf, its *why* they agree or disagree that's really relevant
# [07:29] <anne> yes
# [07:29] <anne> it's about arguments, not numbers
# [07:29] <sbuluf> zeros, agreed, mostly, but it could work for clearly stated, separate proposals
# [07:29] <sbuluf> arguments could be noted in wiki
# [07:30] <anne> those can be done by formal vote
# [07:30] <anne> if necessary
# [07:30] <anne> no need to express that in e-mail
# [07:36] * karl has not yet started to read the 250 messages received this week-end. Still processing IE for the day
# [07:36] <anne> sorry about my time zone nonsense
# [07:36] <anne> didn't notice it was still at +1
# [07:36] <anne> ah, more IE?!
# [07:36] <anne> great
# [07:37] <karl> well not really that great
# [07:37] <anne> well i think so
# [07:37] <karl> everything is a question of context, anne.
# [07:38] <karl> e.ve.ry. single. thing.
# [07:38] <anne> i'm not really debating the fact that processing new IE might not be so great
# [07:39] <anne> i'm just saying that having more IE is great
# [07:46] <Zeros> I hope people don't take the  suggestion too literally
# [07:48] <anne> IE will have something like that I think
# [07:49] <Zeros> That just means that HTML6 is going to end up specifying the "[mode] special annotation for comments" or some such other non-sense
# [07:50] <anne> i think we hope to avoid that the web will rely on that
# [07:50] <anne> much like it doesn't rely too much on  now
# [07:50] <anne> (for non-IE browsers, that is)
# [07:51] <anne> if it indeed ends up in a complete fucked up way well yes... we have an issue
# [07:51] <anne> we already got quite close to vendor lock-in... this versioning proposal is a way to finish that...
# [07:52] <Zeros> close with respect to what feature?
# [07:52] <anne> all the features IE added in IE5 and IE6 that have since then been reverse engineered and implemented
# [07:53] <anne> netscape is also to blame for that I suppose
# [07:53] <Zeros> That's a failure of the specification too
# [07:53] <anne> oh yeah, HTML4 and CSS2 were horrid
# [07:53] <Zeros> Gecko has lots of special Gecko only behavior and Gecko only -moz- properties
# [07:53] <anne> but it appears Microsoft sees specifications as "Guides"
# [07:54] <anne> which may be incompatible with what other people think of then
# [07:54] <anne> them
# [07:54] <Zeros> Its just not as much a problem for Gecko since they don't have near the market permeation
# [07:54] <anne> yes, we have reverse engineered some Gecko behavior
# [07:54] * anne remembers some Range object stuff
# [07:55] <Zeros> If Gecko had 80% market share I'm sure we'd have the same problems with all the XUL features that leak into the public API and are exposed to pages that aren't actually XUL
# [07:55] <sbuluf> the history of the W3C teaches the lesson that the 800-pound gorillas
# [07:55] <sbuluf> *gotta* be listened to.
# [07:56] <sbuluf> http://lists.xml.org/archives/xml-dev/200010/msg00450.html <--from here, interesting reading, if anyone wishes, about w3c origins and mechanics
# [07:57] <sbuluf> it also got me thinking if this spec might not end up being called html 5.2, for historical consistency
# [07:57] <anne> hmm, +100,000 for me!
# [07:58] <Zeros> sbuluf, would that gorilla be MS?
# [07:58] <sbuluf> zeros =P
# [07:59] <Zeros> sbuluf, At least the w3 doesn't charge for its specs. Can you imagine if the html4 spec was $99 per pdf like ISO?
# [07:59] <sbuluf> ugh
# [08:01] <karl> Zeros, sbuluf: many IETF wgs are driven by companies too.
# [08:02] <sbuluf> i have very radical views regarding this too. but once again, i do not wish to disrupt or abuse this space, i should not. enough to say, paid specs are the polar opposite of what i wish. no cigar to w3c either, however, i'm more radical.
# [08:03] <Zeros> karl, didn't think otherwise
# [08:03] <sbuluf> karl, i see. thanks, i did not know.
# [08:03] <sbuluf> (i did not think otherwise either, but did not know, thanks)
# [08:03] <Zeros> "But by including the (sometimes horribly crufty) features that the major vendors demanded, W3C was also able to get them to agree to implement the other (horribly crufty) features that the competition insisted on, and by getting that agreement, also able to label certain things as "not standards-compliant HTML," and have that mean something."
# [08:03] <Zeros> that's a very interesting quote
# [08:04] <Zeros> Interesting that it went from that to semantics too
# [08:05] <sbuluf> tthe article sort of says w3c has two main modes, or moments, so to speak. if the market is quiet, they innovate. if the market (vendors) moves too fast, w3c goes into treaty-making mode (like now, this very group)
# [08:05] <sbuluf> (or like html 3.2)
# [08:05] <Zeros> sbuluf, The web is actually relatively stable right now
# [08:05] <Zeros> canvas is the biggest new thing in a long time
# [08:06] <Zeros> xmlhttp existed for ages before "Web 2.0" was even imagined
# [08:06] <anne> not in other browsers
# [08:06] <anne> although in Gecko it has existed for quite some time, yes
# [08:06] <Zeros> true, though that's not really the same as the separate competing innovation that this email is talking about
# [08:07] <Zeros> we don't have a NS and MS fighting each other with duplicate features with different faces or trying to 1-up each other
# [08:07] <anne> Mozilla realized that would be bad for the web, I think
# [08:08] <anne> although, the storage APIs do break with IE
# [08:08] <anne> I believe the IE model wasn't considered good enough or so
# [08:09] <Zeros> Some agreement needs to be reached and the other big 3 need to promote that if its going to catch on
# [08:10] <Zeros> One would hope it would take the same route as canvas
# [08:10] <sbuluf> who developed canvas? and is it patented?
# [08:10] <anne> Apple developed, patented it and donated it to the WHATWG
# [08:10] <sbuluf> ahhh
# [08:10] <anne> Hixie then specced and changed it slightly and Apple adopted those changes iirc
# [08:10] <sbuluf> and it is the main, biggest feature in html5?
# [08:11] <Zeros> sbuluf, it wasn't really designed as web feature, but rather a feature for Dashboard by Apple
# [08:11] <anne> no, it's one feature
# [08:11] <Zeros> as a*
# [08:11] <sbuluf> big, however? as in "it toook quite a bit of work to be developed"?
# [08:11] <anne> it's quite big, but I think <datagrid> is larger
# [08:11] <sbuluf> mm, i see
# [08:11] <Hixie> apple didn't "donate" canvas to the whatwg
# [08:12] <anne> I thought they agreed to donate it once other vendors asked for that?
# [08:12] <anne> maybe my facts are wrong, sorry
# [08:12] <Hixie> apple implemented and shipped it, mozilla copied the api and implemented something similar, i figured it would be better for everyone involved if we had something compatible so i specified the api
# [08:12] <Zeros> Has apple made an official comment about Mozilla and Opera duplicating it?
# [08:12] <sbuluf> and all parties involved will donate rights so everyone here can implement, right?
# [08:13] <anne> sbuluf, that's part of the reason the HTML WG exists, aiui
# [08:14] <sbuluf> so...microsft joins, they give up nothing, and they get the better part of the next new things in town (some pretty big), for free
# [08:15] <sbuluf> am i too far?
# [08:15] <Zeros> well MS created the contentEditable feature
# [08:15] <Zeros> and for better or for worse that seems to have some backing for HTML5
# [08:15] <anne> lots of the WHATWG stuff is based on MS inventions
# [08:16] <anne> which makes sense
# [08:16] <sbuluf> i see, so is sort of a balanced trade
# [08:16] <anne> why reinvent
# [08:16] <anne> (also one of the design principles)
# [08:16] <Zeros> anne, flawed designs
# [08:16] <Zeros> perpetuating poorly designed APIs doesn't solve much either
# [08:16] <sbuluf> (as usual, thanks all for answers and info)
# [08:16] <Hixie> who cares if it's a balanced trade
# [08:17] <Hixie> we just want to make the web a better place
# [08:19] <anne> Zeros, well, those have not been taken (the storage API IE has is one such example iirc)
# [08:19] <anne> Zeros, same with the XHTML2 href= on every element proposal
# [08:21] * Quits: htmlr (htmlr@203.206.237.84) (Quit: htmlr)
# [08:23] * Joins: htmlr (htmlr@203.206.237.84)
# [08:25] <Zeros> anne, yes, though other things like xmlhttprequest didn't get the same kind of consideration
# [08:25] <Zeros> of course its far too late to change that now
# [08:26] <anne> XMLHttpRequest was already widely deployed and handles HTTP requests quite ok
# [08:27] <anne> it's not the most obvious API, for sure, but we're slowly improving it
# [08:27] <Zeros> its also horribly named
# [08:27] <Zeros> most "ajax" that uses it doesn't actually exchange xml at all
# [08:27] <anne> if that's your only argument than I don't really see the problem
# [08:28] <Zeros> There's also the magic numbers for onreadstatechange
# [08:28] <anne> I changed those into constants
# [08:28] <anne> for readyState, you mean
# [08:28] <Zeros> yes
# [08:29] <Zeros> for use in*
# [08:33] <Zeros> anne, http://www.w3.org/TR/XMLHttpRequest/#xmlhttprequest-members suffers from the same kind of issues though
# [08:34] <Zeros> and worse the fixes for those similar issues "may or may not be implemented by user agents"
# [08:34] <Zeros> They took the MS API and made it into a w3 spec and didn't fix anything
# [08:34] <anne> where does it say that?
# [08:34] <anne> in what context?
# [08:35] <Zeros> http://www.w3.org/TR/XMLHttpRequest/#notcovered
# [08:35] <anne> i sure as hell fixed a lot
# [08:35] <anne> Zeros, those are features that are out of scope for this version
# [08:35] <anne> Zeros, they'll be added in a later one though
# [08:35] <anne> and fully defined
# [08:36] <Zeros> I hope
# [08:36] <Zeros> didn't define the constants in there either
# [08:37] <anne> that's because you're not looking at the latest version
# [08:37] <anne> you're looking at the latest TR/ version
# [08:38] <anne> http://dev.w3.org/cvsweb/~checkout~/2006/webapi/XMLHttpRequest/Overview.html?content-type=text/html;%20charset=utf-8
# [08:40] <Zeros> ah okay
# [08:40] <Zeros> I see you're taking a page from the HTML5 spec and specifying procedural steps for processing as well
# [08:42] <anne> My work is inspired by Hixie, certainly
# [08:49] * Joins: loic (loic@90.41.0.33)
# [08:49] <anne> how's <img src="toto.jpg">La tête à toto</img> better than <img src="toto.jpg" alt="La tête à toto"> karl?
# [08:50] <anne> (at this point in time)
# [08:50] <karl> hidden metadata ;)
# [08:50] <karl> You can use markup
# [08:50] <karl> list, paragraph, you can use language information with multilingual version etc.
# [08:50] <anne> I suppose the argument about hidden metadata is not valid for this case (at least)
# [08:51] <karl> You have a direct access to edit the description
# [08:51] <anne> I agree with that, but you hardly need it
# [08:51] <anne> karl, you could have that with alt="" too
# [08:51] <Zeros> You need it in the same way you'd need it for <audio> or <video>
# [08:51] <Zeros> Why not specify alt for those instead? or the same for canvas
# [08:52] <karl> context: the discussion started in an authoring tool when you drag and drop in an image in a mailer
# [08:52] <Zeros> It was already decided that fallback inside canvas (which it was also argued by mjs is just an <img> with dynamic content) is better off with the content being the fallback
# [08:52] <karl> what is the editing scenario of the image
# [08:52] <anne> Zeros, for <canvas>, <audio> and <video> it would not be backwards compatible
# [08:53] <Hixie> the metadata in alt="" or in content is hidden either way.
# [08:53] <Hixie> the content design is better because there's no good reason to limit alternate content to text only
# [08:53] <Hixie> however at this point this is highly academic since <img> elements have no end tag in HTML
# [08:53] <Zeros> anne, those don't exist yet, what needs to be backwards compatible about them?
# [08:53] <karl> Hixie is hidden in *browsers* :) think out of the small box
# [08:54] <anne> Zeros, the fallback content
# [08:54] <Zeros> karl, IE makes the alt visible on hover
# [08:54] <karl> Zeros: don't tell me ;)
# [08:54] <Hixie> karl: the "metadata is better not hidden" concept is *entirely* about it being hidden in browsers.
# [08:54] <anne> Zeros, as the fallback content is specifically for browsers which don't support it (at least with <audio> and <video>)
# [08:54] <Hixie> karl: metadata is by definition not hidden in non-browser UAs that use that metadata
# [08:55] <karl> not always true.
# [08:55] <karl> interesting.
# [08:55] <karl> beliefs
# [08:55] <karl> assumptions.
# [08:55] <karl> community
# [08:56] <Hixie> what _are_ you talking about
# [08:56] <Zeros> anne, and the alt fallback is specifically for browsers that don't do visual representation of images
# [08:57] <Zeros> there's use cases for extended alt for accessibility as well
# [08:57] <karl> I'm talking about authoring :)
# [08:57] <Zeros> I don't think adding it for authoring solves anything ;)
# [08:57] <karl> alt is often not easily accessible to the author when editing a document
# [08:58] <Zeros> karl, provided the image loads how is the content any different than the alt?
# [08:58] <anne> Zeros, <img> is already supported in the way it is, so the fallback content for browsers that don't support it reason wouldn't really apply to it
# [08:58] <Zeros> Nothing stops a smart editor from putting an input box over images to edit the alt text either
# [08:58] * karl was not talking about modifying <img>
# [08:58] <karl> :)
# [08:58] <anne> karl, in authoring tools that did your suggestion you'd just have <img src></img>
# [08:59] <anne> same problem
# [08:59] <Zeros> karl, isn't that what your <img></img> email was about?
# [08:59] <karl> people do not read mail
# [08:59] * Hixie gives up trying to work out what karl is talking about
# [08:59] <karl> Hixie: sure it is easier ;)
# [08:59] <karl> cavern
# [09:00] <karl> It promotes the idea of a model ala object
# [09:00] <karl> This would not work for backward compatibility.
# [09:00] <karl> I said it twice in the email
# [09:00] <karl> I was using |img| just as the concept example
# [09:00] <Hixie> you said something i said wasn't necessarily true
# [09:00] <Zeros> karl, what were you suggesting then?
# [09:01] <Hixie> but i have no idea (a) what i said that isn't necessary true, (b) why it isn't so, or (c) why you keep speaking in one word sentences
# [09:01] <karl> and thinking that I hesitated to use <graphics> to avoid misunderstanding
# [09:01] <karl> I should have
# [09:01] <karl> but people would have focus on the element graphics
# [09:01] <anne> how does <graphics></graphics> solve the problem?
# [09:01] <karl> bingo !
# [09:01] <karl> CQFD
# [09:01] <anne> of hidden metadata
# [09:02] <anne> I just asked the same about <img></img> btw but you didn't reply to that either
# [09:02] <anne> Hixie has been asking the same thing for a while too, but you don't seem very responsive to that...
# [09:02] <karl> *It promotes the idea of a model ala object*
# [09:02] <anne> how does that solve anything?
# [09:02] <Zeros> karl, yes, we all agree that video, audio et al. are consistent with fallback in content and that img is different... for historical reasons.
# [09:02] <karl> 1. alt="" is difficult to author.
# [09:03] <anne> won't the tools save us? :)
# [09:03] <karl> 2. How can we improve the content model that it makes it easier to author
# [09:04] <Zeros> karl, what is more difficult about alt compared to editing the content?
# [09:04] <Zeros> or was that in reference to "No markup, no multiple choices and in terms of usability difficult to edit."
# [09:04] <karl> 3. alt="" can't contain markup
# [09:05] <karl> Zeros: Have you ever author an HTML email? Have you used a wysiwyg authoring tool
# [09:05] <karl> what is simpler?
# [09:06] <anne> 1,2,3 > <object>?!
# [09:06] <karl> put the description under the image? or go to the menu to try to access the alt attribute if accessible in the editing UI
# [09:07] <anne> when you copy an image into an environment that doesn't support images it should just paste the contents of alt=
# [09:07] <Hixie> alt text isn't for descriptions in the first place, btw
# [09:07] * Joins: kazuhito (kazuhito@210.232.34.13)
# [09:07] <karl> anne: imagine that you drag and drop an image. and it generates for you the object element, open a temporary text area under the image that you can edit. ala Flickr
# [09:07] <karl> Hixie: alt text is a text equivalent of the image :)
# [09:08] <karl> if I have a photo of the beach. I can say either Beach under the sun. Beach with coconut tree, etc
# [09:08] <Hixie> those would all be stupid alt texts
# [09:08] <karl> it all depends on the descriptive equivalent you want to give to your image
# [09:08] <Hixie> alt text is for alternative text, text that is _equivalent_ to the image
# [09:08] <Hixie> not text that describes the image
# [09:08] <Zeros> karl, DW gives a UI to edit the alt attribute in the property pane, hand editing you type out alt yourself, neither seems more difficult than the other.
# [09:08] <karl> doh!
# [09:09] <Hixie> if the whole point of hte image is the image itself, then no amount of alt text will be helpful
# [09:09] <karl> arf.
# [09:09] <anne> we should have something that makes that clear
# [09:09] <Zeros> HTML email editing tools don't actually use HTML for HTML, its used a rtf serialization
# [09:09] <Zeros> as a
# [09:09] <anne> it seems better if alt= is optional for Flickr for instance
# [09:10] <anne> but I'm not sure how you'd distinguish that from the case with a missing alt=
# [09:10] <Hixie> alt: yeah, already have that noted in the list of things to deal with
# [09:11] <Hixie> er
# [09:11] <Hixie> anne:
# [09:11] <Hixie> not alt:
# [09:11] <karl> Hixie: the choice of descriptive text is entirely depending on the context. There is no absolute truth about it
# [09:11] <anne> :)
# [09:11] <Zeros> anne, what's wrong with alt for flickr?
# [09:11] <Zeros> alt="Picture of friends and me"
# [09:12] <Zeros> or better worded would be "Friends and me, at the park..."
# [09:12] * Parts: kazuhito (kazuhito@210.232.34.13)
# [09:13] <Hixie> which would be useful how?
# [09:13] <Hixie> flickr is basically useless if you don't have the images
# [09:13] <anne> that sounds more like a title= to me
# [09:13] <Hixie> why bother with alt text
# [09:13] <Hixie> what anne said
# [09:13] * Joins: kazuhito (kazuhito@210.232.34.13)
# [09:14] * karl thinks that hixie says a lot of statements :) and should try to enlarge his views
# [09:14] <karl> flickr is a community
# [09:14] <Zeros> anne, no browser displays the title in the document when the image doesn't load, or for a text mode browser.
# [09:14] <karl> where many people exchange
# [09:14] <karl> it comes in a context of narrative telling about images. sharing emotions, etc.
# [09:15] <Zeros> flickr also has comments and other content on the page
# [09:15] <karl> images are one part of it
# [09:15] * Quits: gavin_ (gavin@74.103.208.221) (Ping timeout)
# [09:17] * Quits: olivier (ot@128.30.52.30) (Quit: Leaving)
# [09:17] <Hixie> sure, there are parts of flickr that are very useful when you don't have the images. the actual images however, aren't part of that. there's no content they could convey without actually being present themselves.
# [09:17] <Hixie> doesn't matter how big a community is around an image, it's still an image
# [09:17] <karl> do you have dreams?
# [09:18] <karl> same process.
# [09:18] <karl> do you have imagination when reading poetry?
# [09:18] <karl> same process.
# [09:18] <Hixie> once again i have no idea what you're talking about
# [09:18] <Zeros> Hixie, the image ceasing to exist changes the meaning of the page
# [09:18] <karl> etc etc.
# [09:19] <karl> very very very interesting.
# [09:19] <Zeros> Hixie, without alt, when there's no image, what else is there? From an accessibility standpoint alt is also much more accessible than title since it actually gets rendered.
# [09:20] <karl> I have my train to catch and 1h30 of dreaming with the train community.
# [09:20] * Quits: karl (karlcow@128.30.52.30) (Quit: Where dwelt Ymir, or wherein did he find sustenance?)
# [09:20] <anne> how is the title relevant if you can't see the image?
# [09:20] <Zeros> anne, that was your suggestion
# [09:20] * Joins: gavin_ (gavin@74.103.208.221)
# [09:20] <anne> Zeros, i'm saying your text is not good alt= text
# [09:20] <anne> I didn't gave suggestions I think
# [09:21] <anne> Well, I said your text was better suited for title=
# [09:22] <Zeros> "Picture of friends and me" is the logical representation. The comments on the page are in reference to that.
# [09:22] <Hixie> Zeros: certainly the user can be told there's an image -- e.g. [Image] -- but that's got nothing to do with alt text
# [09:22] <anne> oh yeah, the img + title can be exposed to the user
# [09:23] <anne> it should be clear though that it's a title, not text that can replace the image
# [09:23] <anne> (or that the image has replaced that text, however you phrase that)
# [09:23] <Zeros> I'm not sure img[title]::after { content: attr(title); } is a proper solution
# [09:24] <anne> ?!
# [09:26] <Zeros> that's how you'd go about exposing it to a user in a cross UA manner
# [09:27] <anne> the UA would expose it
# [09:27] <anne> not you
# [09:27] <Zeros> What UA?
# [09:27] <Zeros> No UA has that behavior yet, IE doesn't even show the title to the user at all
# [09:27] <Zeros> some make it a tooltip
# [09:27] <anne> yes
# [09:27] <anne> IE does too
# [09:29] <anne> having said all that, those things might make sense in <figure> <img> <legend> here </legend> </figure>
# [09:30] <Zeros> err, okay. you're right, it exposes alt if title is no present
# [09:32] <anne> well, that's a bug
# [09:32] <anne> alt= should only be exposed if there's actually no img visible
# [09:32] <anne> or if images are turned off
# [09:33] <kazuhito> anne: Is that a bug? Does Microsoft explain so?
# [09:34] <anne> i'm not sure if they consider it to be a bug
# [09:34] <anne> it has been reported as one though
# [09:34] <kazuhito> I just wonder that's the way Microsoft implement alt txt in their way,
# [09:35] <kazuhito> and they do not want to call it as a bug.
# [09:36] <anne> i remember them having said it was a feature
# [09:36] <kazuhito> Hmm, can be.
# [09:37] <anne> hmm, 2.44 messages an hour on public-html
# [09:37] <anne> expected total this month: 1800+
# [09:38] <anne> average a day: 58 (up from 50)
# [09:38] <Zeros> its an active list
# [09:39] * Quits: Zeros (Zeros-Elip@69.140.48.129) (Quit: Leaving)
# [09:47] <anne> if labs.opera.com is being read karl might get some more work...
# [09:47] * Joins: marcos_ (chatzilla@203.206.31.102)
# [09:56] * Quits: marcos_ (chatzilla@203.206.31.102) (Ping timeout)
# [10:10] <sbuluf> http://labs.opera.com/webstandards/ --> chaals --> The web took off because HTML was simple, easy to author, and vendor-neutral.
# [10:11] <sbuluf> sort of ironic. now is not simple, not easy to author, and...perhaps not even quite vendor-neutral either.
# [10:11] <anne> the core is still simple
# [10:11] <anne> the more complex things are now possible as opposed to impossible
# [10:12] <anne> and the quite complex things are now easier (e.g. <meta charset=utf-8>)
# [10:12] <anne> i'd therefore only agree on the last of your points
# [10:13] <anne> but even there the same argument applies
# [10:15] <sbuluf> mm, when timbl made his browser, it was a browser, a wysiwyg editor, you need not see the core, html was thought that way (till browser venderos went amok)...and i dunno if there were vendors, at all, or interop problems
# [10:15] <sbuluf> s/core/code/
# [10:15] <marcos> Sbuluf, have you read his book yet?
# [10:16] <sbuluf> marcos, which one?
# [10:16] <marcos> Weaving the Web
# [10:16] <anne> sbuluf, yes, things evolve over time
# [10:16] <sbuluf> mm, i think i did not, i read some of his design section essays, though
# [10:16] <marcos> Seriously, it's all in the book.
# [10:16] <marcos> I highly recommend you read it
# [10:17] <sbuluf> thanks. i discussed some of this with him, btw.
# [10:17] <sbuluf> "why people hand codes is the biggest mistery!" <--timbl
# [10:18] <marcos> because we like coding! its fun
# [10:18] <marcos> Working out how stuff works is also lots of fun
# [10:18] <anne> because editors don't do what we want
# [10:18] <anne> or are less trivial to work with than just doing it by hand
# [10:19] <marcos> how hard is it really to write some tags :P
# [10:19] <marcos> I'll give Timbl a copy of DreamWeaver next time I see him
# [10:22] <sbuluf> http://chatlogs.planetrdf.com/swig/2006-07-16.html#T06-23-51 <--this is the log, might be amusing for some (I'm Biblio)
# [10:23] <sbuluf> (i didn't have a clue at that time that timbl was timbl, btw
# [10:27] <marcos> Anything with rdf in the title is funny :)
# [10:30] <anne> or for that matter, most things my manager says :)
# [10:32] <marcos> yeah, he's pretty funny
# [10:32] <marcos> kinda scary looking, but funny
# [10:37] * Joins: hasather (hasather@81.235.209.174)
# [10:43] * Quits: marcos (chatzilla@131.181.211.113) (Quit: ...and I'm gone.)
# [10:46] * Quits: anne (annevk@131.181.148.59) (Ping timeout)
# [10:53] * Joins: Shunsuke (kuruma@133.27.61.181)
# [10:54] * Joins: ROBOd (robod@86.34.246.154)
# [11:01] * Quits: htmlr (htmlr@203.206.237.84) (Quit: htmlr)
# [11:01] * Joins: htmlr (htmlr@203.206.237.84)
# [11:02] * Quits: Shunsuke (kuruma@133.27.61.181) (Ping timeout)
# [11:22] * Quits: gavin_ (gavin@74.103.208.221) (Ping timeout)
# [11:27] * Joins: gavin_ (gavin@74.103.208.221)
# [11:59] * Quits: MikeSmith (MikeSmith@mcclure.w3.org) (Quit: Get thee behind me, satan.)
# [12:05] * Joins: zcorpan_ (zcorpan@84.216.43.111)
# [12:07] * Quits: kazuhito (kazuhito@210.232.34.13) (Quit: Quitting!)
# [12:16] * Quits: loic (loic@90.41.0.33) (Ping timeout)
# [12:41] * Joins: vz (nv@200.49.140.221)
# [12:41] * Quits: ROBOd (robod@86.34.246.154) (Quit: http://www.robodesign.ro )
# [12:42] * Quits: vz (nv@200.49.140.221) (Quit: vz)
# [12:42] * Quits: sbuluf (fbbfja@200.49.140.40) (Ping timeout)
# [12:43] * Joins: loic (loic@90.41.0.33)
# [13:02] * Quits: htmlr (htmlr@203.206.237.84) (Quit: htmlr)
# [13:06] * Joins: marcos_ (chatzilla@203.206.31.102)
# [13:06] * marcos_ is now known as marcos
# [13:14] * Quits: marcos (chatzilla@203.206.31.102) (Ping timeout)
# [13:21] * Joins: marcos_ (chatzilla@203.206.31.102)
# [13:21] * marcos_ is now known as marcos
# [13:29] * Quits: gavin_ (gavin@74.103.208.221) (Ping timeout)
# [13:34] * Joins: gavin_ (gavin@74.103.208.221)
# [13:44] * Quits: marcos (chatzilla@203.206.31.102) (Ping timeout)
# [13:52] * Joins: edas (edaspet@88.191.34.123)
# [13:59] * Joins: olivier (ot@128.30.52.30)
# [14:07] * Joins: gsnedders (gsnedders@86.139.123.225)
# [14:10] * Quits: polin8 (polin8@24.184.204.6) (Quit: polin8)
# [14:35] * Joins: hasather_ (hasather@81.235.209.174)
# [14:44] * Parts: hasather_ (hasather@81.235.209.174)
# [14:55] * Joins: Shunsuke (kuruma@219.110.80.235)
# [15:00] * Joins: hasather_ (hasather@81.235.209.174)
# [15:01] * Quits: olivier (ot@128.30.52.30) (Quit: Leaving)
# [15:02] * Parts: hasather_ (hasather@81.235.209.174)
# [15:05] * Joins: ROBOd (robod@86.34.246.154)
# [15:27] * Joins: Shunsuke_ (kuruma@219.110.80.235)
# [15:27] * Quits: Shunsuke (kuruma@219.110.80.235) (Ping timeout)
# [15:32] * Joins: polin8 (polin8@64.81.134.176)
# [15:37] * Quits: gavin_ (gavin@74.103.208.221) (Ping timeout)
# [15:42] * Joins: gavin_ (gavin@74.103.208.221)
# [15:51] * Quits: polin8 (polin8@64.81.134.176) (Ping timeout)
# [15:59] * Joins: h3h (bfults@66.162.32.234)
# [16:00] * Joins: polin8 (polin8@64.81.134.176)
# [16:11] * DanC tries tidy and html5lib on his bank statement... html5lib does much better.
# [16:11] <DanC> tidy gave up
# [16:12] * DanC wishes for html5lib glued to an xpath library, with python bindings
# [16:14] <hsivonen> DanC: ElementTree 1.2 seems to have XPath support. which version does html5lib support?
# [16:14] <DanC> html5lib supports elementtree?
# [16:15] * DanC does an svn update... discovers he was way behind...
# [16:15] * Quits: beowulf (carisenda@91.84.50.132) (Ping timeout)
# [16:15] <mjs> hello DanC
# [16:17] <DanC> hi
# [16:17] * Quits: polin8 (polin8@64.81.134.176) (Quit: polin8)
# [16:22] * Joins: beowulf (carisenda@91.84.50.132)
# [16:36] * Joins: polin8 (polin8@64.81.134.176)
# [16:42] <DanC> ah... it wasn't my bank statement that tidy gave up on. saving this bank statement is tricky! firefox downloads it again every time I try to save it, and the re-downloaded copy is bogus (the bank site is goofy)
# [16:42] <DanC> the only way to save it is to view source, select all, copy, paste to another editor, and save from there.
# [16:43] <DanC> html5lib seems a little confused: <html xmlns="http://www.w3.org/1999/xhtml"><head/><body><HTML><HEAD><TITLE>Home
# [16:43] <DanC> Banking Info</TITLE>
# [17:08] * Quits: edas (edaspet@88.191.34.123) (Ping timeout)
# [17:16] * Joins: icaaq (icaaaq@217.13.228.226)
# [17:19] * Quits: mjs (mjs@64.81.48.145) (Quit: mjs)
# [17:33] * Joins: edas (edaspet@88.191.34.123)
# [17:45] * Quits: gavin_ (gavin@74.103.208.221) (Ping timeout)
# [17:50] * Joins: gavin_ (gavin@74.103.208.221)
# [18:10] <h3h> that is terrible
# [18:12] * Joins: Voluminous (Voluminous@66.195.32.2)
# [18:12] * Joins: olli- (olli@80.203.95.229)
# [18:20] * Joins: mjs (mjs@17.255.98.24)
# [18:24] * Quits: edas (edaspet@88.191.34.123) (Quit: http://eric.daspet.name/ et l'édition 2007 de http://www.paris-web.fr/ )
# [18:41] * Quits: myakura (myakura@60.239.122.32) (Ping timeout)
# [18:42] * Joins: myakura (myakura@60.239.122.32)
# [18:49] * Parts: icaaq (icaaaq@217.13.228.226)
# [18:52] * Joins: nickshanks (nicholas@195.137.85.17)
# [18:56] * Joins: hober (ted@69.45.6.105)
# [19:00] <zcorpan_> http://lists.w3.org/Archives/Public/www-style/2007Apr/0037.html
# [19:01] <nickshanks> what does he mean by "magic" ?
# [19:01] <nickshanks> <body> isn't magic
# [19:01] <zcorpan_> in html, it is
# [19:01] <nickshanks> how so?
# [19:01] <zcorpan_> e.g. if you don't specify a background for the root element, the body element's background will cover the canvas
# [19:02] <zcorpan_> (as if it was applied to the root element)
# [19:02] <zcorpan_> not so in xhtml
# [19:02] <zcorpan_> i want it to be so in xhtml
# [19:02] <nickshanks> well that's because the body is the same size as the root
# [19:02] <zcorpan_> no
# [19:02] <zcorpan_> it's not
# [19:02] <nickshanks> sure it is
# [19:02] <nickshanks> why wouldn't it be?!
# [19:02] <zcorpan_> try to specify a border
# [19:03] <zcorpan_> it *isn't*
# [19:03] <nickshanks> the border is outside the viewport
# [19:03] <nickshanks> is that what you mean?
# [19:03] <zcorpan_> well, try in something other than ie in quirks mode... :P
# [19:04] <zcorpan_> or specify a background on the root element as well, in which case you will see where the body element is
# [19:06] <nickshanks> hmm, that's not what i thought would happen
# [19:07] <nickshanks> but it's probably correct
# [19:07] <zcorpan_> yes
# [19:07] <nickshanks> hang on, showing you
# [19:08] <nickshanks> what do you see at http://web.nickshanks.com/test
# [19:08] <nickshanks> i get red showing through on the margin, black border and blue inner
# [19:09] <nickshanks> but then i can never remember whether the body element had padding or margin
# [19:09] <nickshanks> it seems it was margin
# [19:09] <zcorpan_> i don't get any background at all there, just some text in blockquotes
# [19:10] <zcorpan_> same as http://web.nickshanks.com/test.html
# [19:10] <nickshanks> oops, wrong test.html uploadded :)
# [19:10] <nickshanks> try now
# [19:10] <nickshanks> oh crap
# [19:11] <nickshanks> what do you see at http://web.nickshanks.com/temp
# [19:11] <nickshanks> it was the link that was wrong
# [19:12] <nickshanks> as i say, i don't see anything magic happpening
# [19:12] <zcorpan_> in firefox (that supports :root -- opera doesn't yet) i get a red background and a black border around an empty block box
# [19:12] <nickshanks> just red showing through at the margin, black border and blue body
# [19:13] <zcorpan_> that page doesn't trigger the magicness
# [19:13] <zcorpan_> remove the background from the root
# [19:13] <zcorpan_> then the canvas will turn blue
# [19:13] <nickshanks> oh shit
# [19:13] <nickshanks> it's blue outside the border
# [19:13] <zcorpan_> yes
# [19:13] <nickshanks> that's a bug, it's not "magic"
# [19:14] <zcorpan_> it's not a bug
# [19:14] <nickshanks> sure it is
# [19:14] <zcorpan_> it's per css 2.1
# [19:14] <zcorpan_> why don't you trust what i say here? :)
# [19:14] <nickshanks> the default viewport colour should be used (i.e. grey on Mosaic/Netscape, White on modern browsers)
# [19:15] <nickshanks> because what you say is illogical
# [19:15] <hasather> nickshanks: zcorpan_ is right
# [19:15] <nickshanks> so the spec has an error :P
# [19:15] <nickshanks> lets fix it
# [19:15] <zcorpan_> nickshanks: the web depends on it. can't change it now
# [19:15] <nickshanks> the web is fluid, it doesn't depend on anything
# [19:16] <zcorpan_> what we *can* do is make it consistent between html and xhtml (currently this magicness doesn't apply to xhtml)
# [19:16] <zcorpan_> nickshanks: it depends on loads of stuff. browsers can't change this without losing market share
# [19:16] <krijnh> IE could change it *g*
# [19:16] <nickshanks> we need to infiltrate someone into the MSIE team
# [19:17] <nickshanks> if IE changed it, web developers would fix their mistake
# [19:17] <nickshanks> then every other browser could adjust
# [19:17] <zcorpan_> nickshanks: you said yourself you expected the body to cover the canvas. with this rule it does, or it appears to do
# [19:17] <hober> why would MS gratuitously break lots of paying customers?
# [19:18] * hober thinks this is clear case of Don't Break The Web -- update the css spec so that xhtml behaves like html in this regard
# [19:18] <zcorpan_> the web doesn't use xhtml, so not having this rule in xhtml doesn't break the web
# [19:19] <zcorpan_> but i want them to be consistent to avoid confusion among authors
# [19:19] <zcorpan_> i haven't heard a good reason for them to be different
# [19:20] <zcorpan_> the "do other xml languages do this?" is to me irrelevant. we're already doing it for html. html in xml should not be different from html in tag soup
# [19:20] <nickshanks> i think Bert is right, people writing XHTML should know the difference and i think not having the root inherit the background colour of it's content element is correct
# [19:20] <zcorpan_> ...more than necessary that is
# [19:21] <nickshanks> XHTML is just for maths geeks anyway
# [19:21] <zcorpan_> nickshanks: so you want the barrier to use xhtml be higher than it could be?
# [19:21] <zcorpan_> just because they "should know better"?
# [19:21] <nickshanks> yes
# [19:22] <zcorpan_> ok. i disagree.
# [19:22] <nickshanks> in fact, http://golem.ph.utexas.edu/~distler/blog/ should be the only site that uses XHTML
# [19:22] <nickshanks> :-)
# [19:22] * zcorpan_ fails to see the relevance
# [19:22] <nickshanks> everyone else can (and should) use HTML5
# [19:22] <zcorpan_> agreed
# [19:22] <zcorpan_> but still
# [19:22] <zcorpan_> why should xhtml be different here?
# [19:23] <nickshanks> because it's based on XML not on HTML
# [19:23] <nickshanks> it needs to have XML's solid inheritance rules
# [19:23] <nickshanks> even for things as "trivial" as CSS
# [19:23] <hober> If I happen to serialize a both-XHTML5-and-HTML5-compatible DOM into one of the two syntaxes, I can't imagine why they should render gratuitously differently in browsers
# [19:24] <zcorpan_> what inheritance rules?
# [19:24] <zcorpan_> and from an implementation's point of view, css operates on the dom, not on the serialization
# [19:24] <nickshanks> zcorpan_: ones that say :root {} should not inherit from :root > body !!
# [19:24] <zcorpan_> nickshanks: this has nothing to do with inheritance
# [19:25] <zcorpan_> this does not affect the computed value of 'background' on the root element
# [19:25] <nickshanks> well on my test page, if :root has no background-color property, it inherits the background-color of it's child body element
# [19:25] <nickshanks> then why did it go blue?
# [19:25] <zcorpan_> nickshanks: please read the spec, i'll dig up a link for you
# [19:26] * Quits: Shunsuke_ (kuruma@219.110.80.235) (Ping timeout)
# [19:26] <nickshanks> i think CSS 2.1 is a bit of a farce anyway
# [19:26] <zcorpan_> http://www.w3.org/TR/CSS21/colors.html#q2
# [19:26] <zcorpan_> 4th paragraph
# [19:26] <nickshanks> what does CSS 2.0 say about the matter?
# [19:26] <krijnh> nickshanks == mike schinkel?
# [19:27] <zcorpan_> nickshanks: css 2.0 is irrelevant
# [19:27] <nickshanks> css 2.0 is superior, because it has @font-face
# [19:27] <nickshanks> and decent typography is better then anything in CSS 2.1
# [19:28] * nickshanks mumbles something about Comic Sans*
# [19:28] <zcorpan_> css 2.1 dropped @font-face due to lack of implementations. css 2.1 has *lots* of bug fixes compared to css 2.0
# [19:29] <zcorpan_> css 2.0 is not relevant to implementors
# [19:29] <zcorpan_> you can find @font-face or some successor of it in css3, i'd presume
# [19:29] <nickshanks> yes, it's in css3/web-fonts
# [19:31] <mjs> @font-face is ridiculously over-engineered for what it does
# [19:31] * Joins: edas (edaspet@88.191.34.123)
# [19:32] <nickshanks> hmm, so CSS 2.1 §14.2 ¶4 seems like a case of "lets make what Browser X does into the official specification"
# [19:32] <nickshanks> rather than something that was logically thought through
# [19:33] <zcorpan_> nickshanks: yes
# [19:34] <mjs> zcorpan_: we should probably try to make an HTMLWG group decision on whether to advise the CSS WG on this with regards to XHTML5; but maybe it can wait
# [19:34] * Joins: kingryan (rking3@66.92.187.33)
# [19:35] <zcorpan_> mjs: thought about that too
# [19:35] <zcorpan_> should i say i disagree with the resolution or not? (would it matter?)
# [19:36] <mjs> zcorpan_: it does matter somewhat
# [19:36] <nickshanks> zcorpan_: can you come up with a list of sights blighted by this? if we could get the top 100 or so users to switch to using an XHTML-compatible background-colour declaration then perhaps we can get rid of this magicness nonsense?
# [19:37] <nickshanks> i would imagine that 99% of sites do not set borders on their body element anyway
# [19:37] <zcorpan_> nickshanks: this has nothing to do with whether or not there are borders
# [19:37] <zcorpan_> nickshanks: most sites do set the background on body and not the root
# [19:37] <zcorpan_> and they want to fill the canvas
# [19:38] * Joins: asbjornu (asbjorn@84.48.116.134)
# [19:38] <nickshanks> we need another name for 'canvas'
# [19:38] <nickshanks> :)
# [19:39] <nickshanks> there's no harm in moving body { background-colour: puke; } to the html element, is there?
# [19:39] <zcorpan_> nickshanks: the web depends on this rule, and we cannot change the web. all browsers have implemented this, and changing it for html would break the web and thus the browser would lose market share, and thus they can't change it.
# [19:39] <zcorpan_> nickshanks: no there isn't, go ahead and do it
# [19:40] <krijnh> Although background-colour wouldn't work
# [19:40] <nickshanks> well in that case we can start by changing the recommendation from "we recommend that authors specify the background for the BODY element rather than the HTML element" to the other way around
# [19:40] <zcorpan_> nickshanks: go ahead and suggest it on www-style
# [19:41] <nickshanks> am not on that list
# [19:41] <zcorpan_> i don't disagree with removing that recommendation
# [19:41] <zcorpan_> subscribing is trivial
# [19:41] <krijnh> Anyone can post to www-style?
# [19:42] <zcorpan_> don't know
# [19:42] <zcorpan_> perhaps you have to subscribe first
# [19:42] <zcorpan_> anyone can subscribe
# [19:42] <nickshanks> zcorpan_: you're more of a "don't be disruptive, leave things as they are" type of guy, i'm more of a "lets try and get web developers to fix their sites with a bit of carrot and a bit of stick" kind of guy
# [19:42] <krijnh> zcorpan_: But you don't have to be in the WG?
# [19:42] <zcorpan_> nickshanks: i used to be your type of guy
# [19:42] <zcorpan_> nickshanks: i got over it
# [19:42] <nickshanks> you defected
# [19:43] <zcorpan_> krijnh: no
# [19:43] <nickshanks> doo hiss!
# [19:43] <nickshanks> *boo
# [19:43] <zcorpan_> nickshanks: your type of thinking is what got us into quirks mode / standards mode hell
# [19:43] <zcorpan_> afaict
# [19:43] * nickshanks hires a legion of welsh longbowmen to attack your castle*
# [19:44] <krijnh> http://lists.w3.org/Archives/Public/public-html/2007Apr/0894.html - "for every people that excel at what they do, there is an army of others who just don't care enough to dig a little deeper than what is expected from them."
# [19:44] <krijnh> So true
# [19:44] <mjs> nickshanks: your idea assumes there is some concept of "right" and "wrong" that applies to standards
# [19:45] <nickshanks> the quirks mode/standards mode is not inheriently a problem, what is a problem is that IE, FF and Safari don't say anywhere on the page "this site contains 23 errors and generated 54 warnings"
# [19:45] <mjs> nickshanks: unfortunately for you, browser developers are unwilling to punish end users to force authors to live up to your standard of purity
# [19:45] <nickshanks> mjs: you don't need to punish end users
# [19:45] <nickshanks> just make is sufficiently embarrassing for the content owner
# [19:46] <gavin> users won't blame the content owner
# [19:46] <gavin> they'll blame the browser
# [19:46] <nickshanks> you can do that in the UI and render the website as before
# [19:46] <zcorpan_> more specifically, they'll move back to their old browser or to a competitor that doesn't complain about 93% of pages being broken
# [19:46] <nickshanks> gavin: i don't think so, not if it is unobtrusive
# [19:46] <gavin> if it's unobtrusive, they'll ignore it
# [19:46] <nickshanks> right
# [19:47] <mjs> scary warnings that don't actually indicate real harm are a penalty
# [19:47] <nickshanks> but CEOs will say "hey why does the site you created for me have 23 errors"
# [19:47] <mjs> if they are so unobtrusive as to not be annoying, then they won't be noticed and do no good
# [19:47] <gavin> no, they won't
# [19:47] <gavin> exactly
# [19:48] <mjs> having a way to conformance-check your content might be a useful developer add-on to browsers
# [19:48] <nickshanks> i'm thinking here of how iCab and the Safari Tidy plug-in do this; have any of you used these?
# [19:48] <mjs> but it would probably be easier for it to just use an existing conformance checker on the web
# [19:48] <nickshanks> mjs: i believe that browsers should have developer tools built in, the web inspector is a great example of this
# [19:48] <Philip`> You need to punish the right people - Visual C++ 2003 fixed loads of standards-compliance bugs, and wouldn't compile most programs that were written for VC6
# [19:48] <mjs> both of iCab's users seem to like the way it works
# [19:49] <nickshanks> mjs: haha, don't be so cynical
# [19:49] <gavin> trying to make users care about HTML as much as you do by giving them "developer tools" isn't likely going to work ;)
# [19:49] <nickshanks> it is a useful feature both ofr end users and for content developers
# [19:49] <Philip`> and presumably it was successful since VC++2005 fixed more bugs and removed more non-standard behaviour (allowing some non-standard bits via optional compiler flags)
# [19:50] <mjs> Philip`: note, however, that Visual C++ is a developer tool - Windows still runs the old nonconformant compiled programs
# [19:50] <Philip`> and people appear to respect VC++ as being one of the best C++ compilers; which is far from the situation with IE/HTML
# [19:50] <nickshanks> end users can see "hmm, this site has 500+ errors in it's code, i'm not going to trust it much" and *casual* web developers (i.e. the ones who've never heard of the W3C) will finally discover they are emitting tag soup
# [19:51] * Quits: myakura (myakura@60.239.122.32) (Quit: Leaving...)
# [19:51] <Philip`> mjs: Indeed - I assume it works because it's annoying the people who have the power to fix the problem directly
# [19:52] * Quits: gavin_ (gavin@74.103.208.221) (Ping timeout)
# [19:56] <Philip`> (...which does not include a web site's visitors, and often not even its developers since they're relying on someone else's dynamic-site-generation code; so I don't know how you'd find the right people to automatically annoy)
# [19:57] * nickshanks was using Cyberdog the other day...
# [19:58] * Joins: gavin_ (gavin@74.103.208.221)
# [20:00] <nickshanks> Philip basically said what I have been trying to imply: the quickest way to get stuff changed is to annoy those with the power to change it. if the owner of CNN used a Mac he'd demand that the videos on cnn.com were made mac-friendly immediately. clearly he doesn't because this hasn't happened, and as such my access is denied.
# [20:01] <kingryan> nickshanks: annoying your customers is not good for business
# [20:06] <nickshanks> kingryan: are you talking about browser vendors being W3C's customers; web developers being browser vendor customers; website owners being the web developer's customers; end users being the website's customers; or ultimatly, end users being the W3Cs indirect customers. I try to think holistically from the last of those perspectives. what can we do with standards to make the end user experience better. annoying the middle men, browser vendors and web de
# [20:07] <kingryan> I'm talking about web users being the customers of browsers. You annoy them, they'll switch browsers to one that's not as annoying.
# [20:08] <nickshanks> there is no need to annoy end users whatsoever
# [20:08] * Joins: polin8_ (polin8@64.81.134.176)
# [20:08] <nickshanks> everything I mentioned can be done in a browser's "web developer mode" that users need never activate
# [20:09] * Joins: Ashe (Ashe@213.47.199.86)
# [20:09] * Quits: polin8 (polin8@64.81.134.176) (Ping timeout)
# [20:09] <nickshanks> e.g. in web developer mode you can have draconian error handling, in granny mode you can just display a little note in the status bar that this site has HTML errors and may not be worth spending your money at
# [20:10] <nickshanks> whic people will most likely ignore, but the money men hiring contract web developers will want rid of
# [20:10] * Quits: Ashe``` (Ashe@213.47.199.86) (Ping timeout)
# [20:11] * Quits: polin8_ (polin8@64.81.134.176) (Ping timeout)
# [20:12] <nickshanks> i cringe when i read www-html from 1994 and see things like "we already have so many pages, even though that solution is better we have to do such and such this way because otherwise we'll break all 500 websites out there!"
# [20:12] <nickshanks> and think to myself "if only they had done it the other way and fixed those 500 sites...
# [20:12] * xover notes IE does essentially this for JS errors...
# [20:13] <mjs> nickshanks: you're probably not going to like the output of this group, then
# [20:14] <nickshanks> well there are only a few tens of billions of pages now; in 50 years' time we may have 10^30 web pages and the 10^10 of them coded in tag soup HTML 4 and "last updated in 2007" will seem utterly insignificant
# [20:16] * nickshanks shrugs
# [20:16] <nickshanks> i guess i just think forwards-compatibility is more important than backwards-compatibility, especially when most sites and browsers are being actively maintained and can be easily changed.
# [20:17] <hober> I don't think I understand what you mean by "can be easily changed"""
# [20:17] <nickshanks> well website owners can upload new HTML files to their sites; browser vendors can change their code and put out updates
# [20:18] <hober> where do you get "easily" from that?
# [20:18] <nickshanks> does anyone do surveys of last modified dates on web pages?
# [20:19] * Joins: polin8 (polin8@64.81.134.176)
# [20:19] <nickshanks> it takes me about 0.2 seconds to update a web page, basically the time it takes to press command and S together
# [20:19] <nickshanks> that's fairly easy by most regards
# [20:20] <nickshanks> i appreciate most people may have to use FTP, ssh or rsync to update their websites, but i don't see how it can be anything other than easy
# [20:20] <krijnh> Better spend that 0.2 seconds figuring out how reality works :)
# [20:22] * Quits: polin8 (polin8@64.81.134.176) (Quit: polin8)
# [20:22] <xover> Rather a defeatist attitude for leading the web to its full potential...
# [20:23] <nickshanks> and what is reality if it's not a case like my own?
# [20:23] <nickshanks> if updating websites was difficult, do you think we'd have so many?
# [20:24] * Joins: polin8 (polin8@64.81.134.176)
# [20:24] <nickshanks> if updating browsers was difficult, do you think vendors would keep at it and not just give up (okay, so MS gave up at one point, but that's not my point :)
# [20:25] <hober> I don't think you're appreciating the level of complexity present in pretty much all web browsers and in many web applications...
# [20:26] <mjs> updating browsers is difficult
# [20:26] <nickshanks> mjs: have you seen nightly.webkit.org and updated browser every day, created automatically! wow, that's really hard
# [20:26] <nickshanks> *an
# [20:27] <krijnh> lol
# [20:27] * hober boggles
# [20:27] <krijnh> yeah mjs, have you?
# [20:27] <mjs> nickshanks: experimental nightly build != browser update
# [20:27] <mjs> nickshanks: if I tried to ship a random nightly I would probably be fired from my job
# [20:28] <hober> existence of automated build procedure for complex project != "the project is not complex"...
# [20:28] <mjs> anyone here who has actually shipped a browser raise your hand
# [20:29] <xover> Appeal to authority?
# [20:29] <mjs> appeal to experience
# [20:29] <mjs> I have first-hand experience on how hard it is to update a browser
# [20:30] <mjs> you don't have to trust me, but my opinion is likely to be more informed than that of someone who hasn't
# [20:30] <gavin> what do you mean by "update a browser"?
# [20:30] <mjs> I dunno, nickshanks's term, not mine
# [20:30] <nickshanks> gavin: me or maciej?
# [20:30] <gavin> either of you
# [20:31] <gavin> presumably you're talking about the same thing :)
# [20:31] <mjs> I assumed he meant "ship a new version of an existing browser"
# [20:31] <nickshanks> i was referring to getting a new version into as many current user;s hands as possible, and getting the old version out of use
# [20:31] <mjs> but if you think his term is ambiguous, better for him to explain
# [20:32] <nickshanks> e.g. the old version has bug X; the new version doesn't, lets make it as easy as possible for users to a) realise they are out of date, and b) download the new version
# [20:32] * Quits: zcorpan_ (zcorpan@84.216.43.111) (Ping timeout)
# [20:33] <mjs> those parts are easy
# [20:33] <nickshanks> Safari has it easy in this regard because Apple Software Update nags users every week (by default) to update Saafari
# [20:33] <mjs> the hard part is making sure the new version doesn't introduce new bugs Y and Z
# [20:33] * hober thinks mjs et al. have done a great job of that
# [20:33] <nickshanks> mjs: well i am assuming we're already past the QA stage of the release
# [20:34] <gavin> sounds like mjs is talking about code->bits, and nickshanks is talking about bits->users :)
# [20:34] <mjs> nickshanks: ah, ok, so you are saying the easy part of delivering a new browser version is easy
# [20:35] <mjs> well, no argument there
# [20:35] <mjs> but don't forget about the hard part of actually doing the development and testing
# [20:35] <nickshanks> I'm saying CSS 4.0 introduces new-fangled-rule, and IE12 and FF 4 implement it; it's easy for web developers ot add new-fangled-rule to their CSS files and upload it, and it (should be) just as easy for end users to update their browsers
# [20:36] <nickshanks> (mjs: i concur, but someone asked how easy it was for people to update their websites)
# [20:40] <nickshanks> mjs: also, i think all current vendors have a problem, and that is differentiating themselves from the competition. if all they were concerned about was fixing extant bugs in their code, and not trying to win new users by adding cool features (not denying that cool new features aren't appreciated) then i think it would be easier to put out updates. it would also help if updates *to consumers* were more frequent, say every two weeks, with the trunk code
# [20:40] <nickshanks> currently this only happens with security fixes
# [20:40] <mjs> nickshanks: actually, fixing web content bugs is *much* more risky and more difficult than adding new UI features
# [20:40] * Quits: polin8 (polin8@64.81.134.176) (Quit: polin8)
# [20:41] <gavin> indeed
# [20:41] <mjs> nickshanks: so, I kinda think you don't know what you are talking about
# [20:42] <nickshanks> mjs: yeah, i've never shipped a browser (but would like to!)
# [20:42] * Joins: polin8 (polin8@64.81.134.176)
# [20:42] <nickshanks> what do you mean by "risk" ?
# [20:42] <mjs> risk of causing regressions
# [20:42] * Joins: Ashe`` (Ashe@213.47.199.86)
# [20:43] <nickshanks> risk that joebloggsesflashywebsite.com will not look like it did before?
# [20:43] <mjs> web content bug fixes have a very high risk of breaking other web sites
# [20:43] <nickshanks> and that's okay
# [20:43] <kingryan> nickshanks: no it's not
# [20:43] <nickshanks> breaking a few websites here and there (that were using broken code to start with) should not be considered a problem
# [20:43] <mjs> risk that cnn.com will have its layout mangled into unreadability; risk that yahoo.com will crash; risk of introducing security holes or hangs
# [20:44] <kingryan> when a browser breaks my grandma's online banking site, then she can't buy me a birthday present, that's *not good* :D
# [20:44] * Quits: Ashe (Ashe@213.47.199.86) (Ping timeout)
# [20:44] <nickshanks> i'm not talking about mangled beyond readability
# [20:44] <nickshanks> i'm talking from the viewpoint of my own experience, which is fixing CSS bugs in Safari
# [20:44] <nickshanks> things like making a ~ b work
# [20:45] <nickshanks> that fix should have gone out 1.5 years ago when i submitted the patch and it was accepted
# [20:45] <nickshanks> presently it's only available via nightly builds
# [20:45] <mjs> adding new features is somewhat less risky, but still risks crashes, hangs, security holes, bad performance in pathological cases, etc
# [20:48] <nickshanks> which we test for, you can't guarantee that any line of code in any piece of software is 100% safe, but if it works in all conceivable usage cases, is that not "good enough" ?
# [20:50] * xover notes Netscape Navigator 4.x is finally dead and content workarounds for its bugs can be safely dispensed with…
# [20:52] <DanC> ok, I caught up on the "formal definition" thread and chimed in. here's hoping that doesn't stir up more discussion of it, or if it does, that the discussion is productive.
# [20:55] <DanC> on the "poor authorship" thread, I have wondered about "best practices" enough to check the charter once or twice; it's not in there explicitly.
# [20:58] * Joins: briansuda (briansuda@130.208.155.180)
# [20:58] <xover> Hmm. DanC: You do not view formal definition in a machine parseable format (EBNF/DTD/XSD/Foo/Etc.) as a fundamental requirement for this WGs output standard?
# [20:58] * DanC is happy to leave that thread alone.
# [20:58] <DanC> indeed, I do not, xover.
# [20:59] * xover is somewhat suprised…
# [20:59] * Quits: polin8 (polin8@64.81.134.176) (Quit: polin8)
# [20:59] <DanC> I'm happy for us to produce one, but I can imagine us succeeding without one.
# [21:00] * Joins: dbaron (dbaron@63.245.220.242)
# [21:01] <DanC> dbaron, hi... mjs was saying a day or so ago that he thinks it's time to put the question on the design principles. after I asked a few questions, he went to check on a few more things...
# [21:01] * Joins: polin8 (polin8@64.81.134.176)
# [21:01] <dbaron> hi
# [21:01] <DanC> I haven't finished reviewing them. but I wonder if I should consider myself way behind.
# [21:01] <mjs> DanC: mainly I wanted to read and respond to new feedback - I thin I am caught up now
# [21:01] <dbaron> I don't think I can pretend to be following the list anymore -- too much traffic.
# [21:02] <mjs> DanC: I think the non-disputed ones are ready to be put to a decision
# [21:02] <DanC> I think a lot got lost between dbaron's 2 pages or so on "don't bread the web" and the 1 paragraph that made it to the wiki.
# [21:02] <mjs> DanC: I can write up an actual proposition and you and Chris W. can decide how best to actually decide the matter
# [21:03] <DanC> well, I suppose a proposition is always in order... though I'm not sure when is the best time for it just yet.
# [21:03] <DanC> I did some updating of my issues list last Friday.
# [21:03] <DanC> I really should get over my reluctance to schedule a teleconference.
# [21:04] <DanC> a teleconference can really help with the "I don't think I can pretend to be following the list anymore -- too much traffic." syndrome. a teleconference agenda is a natural feedback mechanism that puts a damper when there's too much traffic.
# [21:04] <nickshanks> is breaded web better than webbed bread?
# [21:04] * DanC was waiting for the break/bread joke
# [21:05] <nickshanks> sorry, i was in another channel ;)
# [21:05] <DanC> Chris hasn't responded, directly, to your last proposal, mjs. I don't know how big of a pipeline you want to maintain ;-)
# [21:06] <DanC> I should talk with Chris about it.
# [21:06] <DanC> "Why even care about incompetent web developers?" oh my.
# [21:07] <mjs> DanC: well, the last proposal needed discussion - although that now seems to have settled down
# [21:08] <DanC> yeah... so now I need to find out if Chris is happy to put the question (or have me do it) or Microsoft wants more time to discuss it.
# [21:09] <DanC> "being a web developer requires some skill that a lot of people just don't have". but there's an awful lot of web content not produced by so-called "web developers".
# [21:09] <mjs> DanC: it sounded like he wanted to give some feedback from Microsoft's POV
# [21:10] <xover> But should it be created directly by such users, or by tools with suitable levels of abstraction?
# [21:10] <DanC> his "versioning and html[5]" message might be a response to the HTML5 proposal. not entirely clear.
# [21:11] <DanC> xover, the question of how it *should* be created is somewhat academic. The fact is, it *is* created by a decreasingly sophisticated authorship.
# [21:11] <mjs> DanC: I think his message is solely about whether to have version syntax, not a response to the HTML5 proposal - he was planning to write it since before the HTML5 proposal was made
# [21:12] <xover> And increasingly by way of tools such as Microsoft Word, Bloxom, blogger.com, etc.
# [21:12] <DanC> bloxom and blogger let authors type markup, no?
# [21:12] <DanC> oh crud. I missed a 2pm appointment.
# [21:13] <xover> Is your "unsophisticated authorship" really interested in editing raw HTML?
# [21:15] <DanC> no; they're interested in photos of the party last weekend. but raw HTML is the shortest path to the target, currently. (assuming myspace lets you type OMG and such; I don't even know.)
# [21:15] <kingryan> xover: they're also interested in copy/pasting raw html (ref: myspace themes)
# [21:16] <xover> kingryan: the theme, yes; but the HTML itself?
# [21:17] * Quits: dbaron (dbaron@63.245.220.242) (Ping timeout)
# [21:17] <kingryan> xover: the theme is written in html/css/js
# [21:17] * Joins: bewest (ben@209.237.236.227)
# [21:18] <xover> I would assume myspace lets you click on “Upload Image” to achieve that goal.
# [21:19] <DanC> but what about captions and comments? OMG was intended as part of a comment
# [21:19] <xover> Have a look at Apple's .Mac Homepage and Picture Albums service. Not a bracket in sight...
# [21:20] <DanC> and the market share is...?
# [21:20] <xover> I'm sure mjs could tell you if you're really interested...
# [21:21] <kingryan> xover: I'm not sure of the connection between "upload image" and the html/css/js needed to layout and style the page
# [21:21] <DanC> meanwhile, flickr allows html in comments and captions. flickr gets more exposure in my world. I dunno the real market numbers.
# [21:21] * DanC suspects "upload image" refers to "pictures of the party last weekend"
# [21:22] <xover> More to the point, the current activity is presumably meant to produce something of some longevity which will hopefully lead the web in the direction of its full potential.
# [21:23] <DanC> the web is designed to reflect the breadth of human experience. 80% of everything is drek (or 90% or 97%, depending on who you ask); I don't expect the web to be different.
# [21:24] <DanC> if you try to constrain away the 80%, you risk losing the really best 1%
# [21:24] <xover> For instance, better support of web applications (think AJAX, just for frame of reference) should lower the bar for providing the decreasingly sophisticated authorship with increasingly sophisticated tools.
# [21:26] <DanC> sure. I'm losing the relevance to the HTML WG's work.
# [21:26] <xover> Hence, the spec need not be authored to "pander" to a decreasingly sophisticated authorship, provided it "panders" to the increasingly sophisticated toolsmiths.
# [21:27] <DanC> ok, that makes sense. It involved not *ignoring* the "incompetent web developers", but making extra effort to balance them.
# [21:28] <DanC> involves
# [21:28] <DanC> now I must dash...
# [21:28] <xover> ttyl (and thanks for your time)
# [21:55] <mjs> DanC: Alexa and Google PageRank can tell you how visited and how linked things are, though can't really tell you to what extent users upload content
# [21:56] <mjs> DanC: more and more platforms for end-user generated content are using rich text editing that hides the angle brackets, though many have escapes for experts
# [22:00] * Quits: gavin_ (gavin@74.103.208.221) (Ping timeout)
# [22:05] * Joins: gavin_ (gavin@74.103.208.221)
# [22:14] * Joins: dbaron (dbaron@63.245.220.242)
# [22:26] * Quits: gsnedders (gsnedders@86.139.123.225) (Ping timeout)
# [22:34] * Joins: gsnedders (gsnedders@86.139.123.225)
# [22:38] * DanC returns from errands
# [22:38] <DanC> hmm... we passed 300 at some point. 342 now. http://www.w3.org/2000/09/dbwg/details?group=40318&public=1
# [22:39] <DanC> Miller Abel is new to me... from Microsoft.
# [22:40] <DanC> hmm... I don't recall hearing much from AOL.
# [22:40] <DanC> "Galen O\'Hanlon". sigh. wonder where that bug is.
# [22:41] <kingryan> php/mysql?
# [22:41] <DanC> pretty likely
# [22:43] <kingryan> magic_quotes_gpc is the likely cuprit then
# [22:43] <kingryan> culprit*
# [22:47] * Quits: ROBOd (robod@86.34.246.154) (Quit: http://www.robodesign.ro )
# [22:50] <bewest> DanC: is it feasible to put the stats summarizing the page at the top of the page instead the bottom?
# [22:51] * Quits: loic (loic@90.41.0.33) (Ping timeout)
# [22:51] <DanC> I'm sure it's a matter of a few lines of php... but it's a scrip that's shared by all WGs. The shortest path is probably for you to mail that suggestion to site-comments@w3.org, cc me and karl.
# [22:52] <DanC> site-comments has a public archive
# [22:52] <bewest> ok
# [22:57] <DanC> when I read Chris's "I'm just waiting for the press to pick up one of these threads" messages, I was tempted to reply "smile, you're on slashdot!". But didn't want to encourage that sort of thing.
# [22:58] <DanC> http://developers.slashdot.org/comments.pl?sid=230515&cid=18704035
# [23:01] * Quits: dbaron (dbaron@63.245.220.242) (Quit: 8403864 bytes have been tenured, next gc will be global.)
# [23:14] * Quits: briansuda (briansuda@130.208.155.180) (Quit: briansuda)
# [23:17] * Joins: dbaron (dbaron@63.245.220.242)
# [23:23] <DanC> hoot! [[
# [23:23] <DanC> Apparently the standard was practicing catch-and-release, because it
# [23:23] <DanC> wasn't captured firmly enough to actually be interoperable.
# [23:23] <DanC> ]]
# [23:23] <mjs> DanC: I'm sure he has noticed by now, at least if Microsoft's marketing folks pay as much attention to web commentary as ours do
# [23:24] * DanC is really embarrased about <object>.
# [23:28] <Dashiva> I haven't seen Chris respond to the "If you release often, people won't have time to lock into bugs" line of talk yet
# [23:31] <Voluminous> Chris is out of office for today and tomorrow I thought.
# [23:58] * Quits: mjs (mjs@17.255.98.24) (Quit: mjs)
# Session Close: Tue Apr 17 00:00:01 2007

The end :)