/irc-logs / freenode / #whatwg / 2007-08-24 / end

Options:

# Session Start: Fri Aug 24 00:00:00 2007
# Session Ident: #whatwg
# [00:00] <Hixie> that's possible too, i haven't even looked at tabindex yet other than adding the negative thing
# [00:01] <zcorpan> yep. :) i just suddenly came to think of tabindex when reading the above
# [00:01] <Hixie> one thing i really don't know how to fix is accesskey=""
# [00:02] <zcorpan> dunno either
# [00:02] <jgraham> That isn't going to be fun.
# [00:02] <zcorpan> although opera's implementation is somewhat useful, or at least not harming the user experience
# [00:03] <Hixie> opera's implementation is unintuitive
# [00:03] <jgraham> From what I remember Mike Smith saying it sounds like different solutions are appropriate on different devices
# [00:03] <Hixie> and doesn't actually solve the problem of how to make it device independent
# [00:03] <Hixie> yeah
# [00:03] <Hixie> maybe it really is a stylistic thing
# [00:03] <Hixie> we do have key-equivalent in CSS iirc
# [00:04] <Philip`> Hixie: I would assume users wouldn't see the spam or useless content since they wouldn't be explicitly asking their tool to give them the longdesc, except in the cases where they have a good expectation that it's going to be useful, so misuse wouldn't hurt the user much
# [00:04] <Hixie> Philip`: really? how would they know to ask?
# [00:05] <Hixie> for longdesc it really seems to me that the only cases i've seen where the longdesc was actually useful and wasn't something you could have just stuffed into alt="", it was actually useful to sighted users too
# [00:05] <Hixie> and could have just been included on the page or in a link from the page
# [00:05] <Philip`> I'd assume the tool would indicate in some quick way that there is a longdesc attached to the image, similarly to how it must quickly indicate wherever there is a link
# [00:06] * Joins: doublec (n=doublec@202.180.114.137)
# [00:06] <Philip`> but I have precisely no experience of how relevant tools work in practice
# [00:06] <zcorpan> jaws says "press enter for long description" after reading the image alt, and if you press enter it will open the url in a new window
# [00:06] <zcorpan> iirc
# [00:06] <Hixie> right but how do you know it's appropriate?
# [00:07] <Philip`> (At least from what I've heard about table headers, they're not read out by default - you press some key when you've got the right cell selected, and if you see one cell has rubbish headers then you won't bother checking every single other cell - so the harm caused by misuse is similarly minimised)
# [00:09] * Quits: doublec (n=doublec@202.180.114.137) (Client Quit)
# [00:09] <Philip`> Hixie: By considering the context, like whether other images on the page have useful longdescs - e.g. you'd know it's worthwhile reading all the longdescs in the CSS spec after you've seen the first few
# [00:09] * Joins: doublec (n=doublec@202.180.114.137)
# [00:10] <Philip`> (and on sites which misuse it, you'd quickly realise you should just ignore all the rest)
# [00:10] <Hixie> that doesn't seem like the optimal user experience
# [00:10] <Hixie> it's like reminding the user continually that the page wasn't designed for them but there's this secondary set of content they can access
# [00:11] <Hixie> seems like it would make one bitter
# [00:12] * Hixie tries to build up a list of requirements for offline web apps
# [00:14] <Philip`> It seems a less significant negative point than with e.g. <input usemap> (where if your browser supports it, there are features of some sites that just don't work at all) - it might waste a bit of time to read out useless longdescs, but it's not preventing the user from using the site
# [00:15] * Joins: csarven (n=nevrasc@modemcable081.152-201-24.mc.videotron.ca)
# [00:16] <Philip`> A closer-to-optimal user experience would be good, though I don't have any ideas for that :-)
# [00:16] <Hixie> yeah the main argument for me against longdesc is that it's not useful at all, except in rare cases where frankly sighted users would benefit too, and therefore you're better off putting the content on the page itself
# [00:16] <othermaciej> Hixie: one thing that might work is to limit the number of access keys and make it a set that UAs can map in a natural way on each platform to something w/ no conflicts
# [00:16] <Hixie> it's hard to say since i've seen so few useful uses of it
# [00:16] <Hixie> othermaciej: yeah that's been suggested
# [00:16] <Hixie> othermaciej: but then the UI becomes unclear
# [00:17] <Hixie> othermaciej: i.e. discoverability drops through the floor
# [00:20] <kingryan> Hixie: having CSS-useable hooks could help with styleability
# [00:20] <Hixie> yeah
# [00:20] <kingryan> s/styleability/discoverability/
# [00:21] <Hixie> like ...:after { content: ' (' copy-key ')'; }
# [00:21] <Hixie> ... { key-equivalent: copy-key; }
# [00:21] <kingryan> yeah
# [00:21] <Hixie> or ... { key-equivalent: copy-key; } ...::after { content: ' (' key-equivalent ')'; }
# [00:21] <Hixie> ...to have resilience against the cascade
# [00:21] <kingryan> ....:access-key('n'):after { content: 'meta-key N'}
# [00:22] <kingryan> yeah
# [00:22] <Hixie> i was thinking of also having othermaciej's idea since that solves the device problem too
# [00:22] <Philip`> Subtitles on TV shows would benefit sighted users too (e.g. when they get distracted and miss a couple of words, or can't understand someone's accent), but usually they aren't displayed along with the content since they're ugly and distracting and not sufficiently useful to be shown to all users; people may have similar reasons for not wanting to put image descriptions in the normal page content, and hiding it being [D] links or londesc
# [00:23] <kingryan> Hixie: is om's idea of having a limited set?
# [00:23] <Philip`> s/being/behind/
# [00:23] <Hixie> yeah but subtitles aren't only accessible to blind users or hidden behind long properties pages, they're one-button accessible
# [00:23] <Hixie> like an <a> link would be
# [00:23] <Hixie> kingryan: yeah see above
# [00:24] <Philip`> That just seems like a browser UI issue
# [00:24] * Quits: tndH_ (i=Rob@adsl-87-102-90-11.karoo.KCOM.COM) ("ChatZilla 0.9.78.1-rdmsoft [XULRunner 1.8.0.9/2006120508]")
# [00:24] <kingryan> yeah, me wishes browsers couldn't remap stuff that the browser already handles
# [00:24] <kingryan> or allow me to re-remap it
# [00:24] <Hixie> Philip`: i don't think we should hide longdesc behind a context menu or something. i'm saying the link should be right there in the content just like for everything else.
# [00:27] <Philip`> That sounds like D-links - I think the only argument I've heard against them is they don't look very nice (but there are quite possibly other arguments I haven't heard)
# [00:27] * Quits: hendry (n=hendry@nox.vm.bytemark.co.uk) ("sleep&")
# [00:27] <othermaciej> Hixie: if the set was 0-9 for instance, it could be unmodified 0-9 on phones with a keypad, Cmd-0 - Cmd-9 on Macs, Ctrl-0 - Ctrl-9 on windows, etc
# [00:27] <othermaciej> (assuming those are actually free)
# [00:28] <Hixie> Philip`: yeah, though i wouldn't use [D].
# [00:28] <othermaciej> but I'm not sure how you address discoverability
# [00:28] <Philip`> (http://www.w3.org/TR/WCAG10-HTML-TECHS/#long-descriptions says "Invisible d-links thus provide a (temporary) solution for designers who wish to avoid visible d-links for stylistic reasons.")
# [00:28] <othermaciej> I think the page has to provide info about shortcuts
# [00:28] <Hixie> othermaciej: i think kingryan's idea helps with that
# [00:28] <billmason> I think there's some kind of study that says numeric shortcuts aren't really free, though.
# [00:28] <othermaciej> that's how keyboard shortcuts work in native apps
# [00:28] <othermaciej> they are listed in the menu bar
# [00:29] <othermaciej> having a shortcut list with labels somewhere in the page or accessible from the chrome would be the natural analogy
# [00:29] <othermaciej> I'm not sure if styling them to mention they key helps
# [00:29] <kingryan> yeah, having it in the chrome seems reasonable
# [00:29] <othermaciej> thee UA could do a better job if it is the only thing that knows the concrete key mapping, but you need a label to go next to the shortcut
# [00:30] <othermaciej> *the UA
# [00:30] <othermaciej> it could even be a menu or submenu in the menu bar
# [00:30] <Hixie> hm yeah, for <command>s and other Command elements you could just have the UA create a menu somewhere with the key equivs
# [00:30] <othermaciej> "Page Shorcuts"
# [00:30] <kingryan> it'd be nice in interactive browsers if you could could style the access-key elements with the modifier key is pressed
# [00:30] <Hixie> they could do that now with :active
# [00:31] <othermaciej> kingryan: if the modifier is a commonly used one on the OS, that could be distracting
# [00:31] <kingryan> othermaciej: indeed it could
# [00:31] <othermaciej> something based on <command> seems like a good basic approach
# [00:31] <Hixie> oh i misread what kingryan said
# [00:31] <othermaciej> the idea is that a shorcut key actually activates a command, for which there might also be one or more UI elements
# [00:31] <kingryan> there are some places, like dialogue windows in os x, where this is done already
# [00:32] <othermaciej> it makes more sense to associate it with a command, which can then have an appropriate label
# [00:32] <othermaciej> kingryan: example?
# [00:32] <othermaciej> (I don't know of dialog windows that have keyboard shortcuts beyond the standard tab/enter/esc and such)
# [00:33] <kingryan> I remember something adding "cmd-foo" in a button when cmd is pressed down
# [00:33] <kingryan> I'm not sure where that is
# [00:33] <othermaciej> I don't think that is standard
# [00:33] <kingryan> it's not
# [00:33] <kingryan> I'm just seeing that I've seen it done before and I helped with discoverability
# [00:36] <othermaciej> usually in OS X all the keyboard shorcuts also have a menu item, so you can look in the menu system to see the shortcuts
# [00:39] <kingryan> othermaciej: true
# [00:41] <othermaciej> Windows is different since it uses the underline system sometimes, and that may apply both to menus and items in a dialog
# [00:47] * Joins: hober (n=ted@unaffiliated/hober)
# [01:22] <Hixie> ok
# [01:22] * Hixie has now read all feedback on offline storage and is coming up with ideas
# [01:23] <Hixie> Please Stand By... Processing Feedback... 12% [## ]
# [01:24] * Quits: othermaciej (i=mjs@nat/apple/x-ba32850b25eb03af) (Read error: 110 (Connection timed out))
# [01:27] * Joins: othermaciej (n=mjs@17.255.100.25)
# [01:27] * Joins: grimboy (n=grimboy@85-211-250-25.dsl.pipex.com)
# [01:31] * Parts: billmason (n=billmaso@ip156.unival.com)
# [01:41] * Joins: othermaciej_ (i=mjs@nat/apple/x-0cbad082ed829de8)
# [01:46] * Quits: othermaciej_ (i=mjs@nat/apple/x-0cbad082ed829de8)
# [01:48] * Joins: yod (n=ot@softbank221018155222.bbtec.net)
# [01:59] * Quits: othermaciej (n=mjs@17.255.100.25) (Read error: 110 (Connection timed out))
# [02:16] * Joins: karlUshi (n=karl@133.27.247.173)
# [02:19] <zcorpan> error reporting in the content-type sniffing algorithms might not be a bad idea, actually
# [02:20] <Philip`> Does the error reporting need to be more complex than just complaining if the sniffed content type != the HTTP Content-Type?
# [02:20] <zcorpan> don't think so
# [02:21] <zcorpan> the html spec might not need to say anything about it, since it isn't really in scope for html
# [02:21] <zcorpan> but we could still log such cases in the error console
# [02:24] * Quits: dbaron (n=dbaron@corp-241.mountainview.mozilla.com) ("8403864 bytes have been tenured, next gc will be global.")
# [02:24] * Quits: weinig_ (i=weinig@nat/apple/x-7e33f763c675eaa5) (Read error: 110 (Connection timed out))
# [02:24] * Joins: weinig_ (i=weinig@nat/apple/x-011e4a90c616b004)
# [02:42] * Joins: Oeighty (n=oxygen80@ip-58-28-200-134.ubs-dsl.xnet.co.nz)
# [02:49] * Quits: h3h (n=w3rd@66-162-32-234.static.twtelecom.net) ("|")
# [02:55] * Quits: bzed (n=bzed@dslb-084-059-107-157.pools.arcor-ip.net) (Client Quit)
# [02:56] * Quits: weinig_ (i=weinig@nat/apple/x-011e4a90c616b004) (Read error: 110 (Connection timed out))
# [02:57] * Joins: othermaciej (n=mjs@dsl081-048-145.sfo1.dsl.speakeasy.net)
# [02:58] * Quits: kingryan (n=kingryan@corp.technorati.com)
# [03:05] * Quits: hober (n=ted@unaffiliated/hober) ("ERC Version 5.2 (IRC client for Emacs)")
# [03:16] * Joins: tantek (n=tantek@000-126-983.area2.spcsdns.net)
# [03:31] * Joins: weinig_ (i=weinig@nat/apple/x-73474013b30819cf)
# [03:43] <Hixie> woot
# [03:43] * Hixie adds support for fragment identifiers to his issues list
# [03:53] * Quits: tantek (n=tantek@000-126-983.area2.spcsdns.net) (Read error: 104 (Connection reset by peer))
# [03:58] <Philip`> Hixie: With your idea, is there anything to stop people getting inconsistent caches if they're already in the middle of an update when the web server starts serving a newer version of the application?
# [04:00] * Quits: zcorpan (n=zcorpan@84-216-40-15.sprayadsl.telenor.se) (Read error: 110 (Connection timed out))
# [04:11] * Joins: MikeSmith (n=MikeSmit@eM60-254-218-189.pool.emnet.ne.jp)
# [04:16] * Quits: weinig_ (i=weinig@nat/apple/x-73474013b30819cf) (Read error: 104 (Connection reset by peer))
# [04:16] * Joins: weinig_ (i=weinig@nat/apple/x-077d52b93e9d8a74)
# [04:31] * Quits: aroben (n=adamrobe@unaffiliated/aroben)
# [04:50] <Hixie> Philip`: no, that's the same as the status quo, effectively
# [04:58] * Joins: dbaron (n=dbaron@corp-241.mountainview.mozilla.com)
# [05:01] * moeffju is now known as moeffju[ZzZz]
# [05:08] <Lachy> Hixie, re: "btw Lachy if you can suggest some page-only heuristics (i.e. not involving the network) for detecting bogus longdesc=""s that would have caught them in the URLs i mentioned, it would be useful"
# [05:09] <Lachy> the only other heurisitic I thought of what checking the URL to see if it ended in .jpg, .png, .gif, etc. since linking a longdesc to an image is pointless. But that would have only caught one example that I found
# [05:10] <Lachy> re [D] links and longdesc, http://joeclark.org/book/sashay/serialization/Chapter06.html#d-links
# [05:10] <Lachy> that can be summarised as: "[D]" links look awful, difficult to associate a link directly with the image, longdesc doesn't cause catastrophic bugs in browsers, and "[D]" links are as bad as "click here"
# [05:12] <Lachy> I think the first issue can be solved by using different link text or hiding it with CSS,
# [05:12] <Lachy> the next could possibly be solved using <figure><img><a rel=longdesc/></figure>,
# [05:13] <Lachy> and the third is countered by the argument that existing use of longdesc is more harmful
# [05:13] <Hixie> Lachy: i actually saw some .png/.jpg/.gif links that were real description of the images
# [05:13] <Lachy> oh, ok.
# [05:13] <Hixie> but yeah
# [05:14] <Hixie> i wanted to do that originally
# [05:14] <Hixie> i think i saw someone had ...descriptions...?foo.gif
# [05:14] <Hixie> or some such
# [05:14] <Hixie> heh, maciej ignored my request at the top of the e-mail :-P
# [05:14] <Lachy> oh, well you'd have to ignore the query string, but I suppose that's still no guarantee
# [05:14] <Hixie> othermaciej: which list should i reply to you on
# [05:15] <othermaciej> Hixie: what was your request?
# [05:15] <Hixie> to only reply to one list :-)
# [05:15] <othermaciej> oh, you cross-posted
# [05:15] <othermaciej> didn't even notice that
# [05:15] <othermaciej> mea culpa
# [05:15] <Hixie> hehe
# [05:15] <Hixie> no worries
# [05:15] <Hixie> i'll just post to the htmlwg one i guess
# [05:15] <othermaciej> I'm actually not sure which list contains more of the interested parties
# [05:15] <Hixie> yeah me neither
# [05:15] <Hixie> but i'm sure i'll get flamed if i use the whatwg list
# [05:16] <Hixie> actually i really want to hit the scour people with this
# [05:16] <Hixie> and i don't think they're on public-html
# [05:16] <Hixie> so screw it
# [05:16] <Hixie> whatwg it is
# [05:20] <Hixie> hey Lachy?
# [05:20] <Lachy> yo
# [05:20] <Hixie> Lachy: i didn't get mail from public-forms-tf, should i have?
# [05:21] <Lachy> the list still isn't showing up for me in the maintenance page, so I can't fix it
# [05:21] <Hixie> ah ok
# [05:21] <Hixie> danc told me you were in charge now, so i'd assumed it'd gone through
# [05:22] <Lachy> I mailed sysreq yesterday and still haven't heard back
# [05:22] <Lachy> well, sysreq initially responded and said it was all done. it's hasn't and I'm waiting to hear back. I'll check with danc later
# [05:23] <Hixie> k
# [05:26] <Lachy> http://blog.whatwg.org/omit-alt#comment-7669 he seems to ignored everything I wrote
# [05:27] <Lachy> I wonder if he actually read the spec?
# [05:29] <Hixie> unlikely
# [05:29] <Hixie> but i wouldn't worry about it
# [05:35] * Quits: dbaron (n=dbaron@corp-241.mountainview.mozilla.com) ("8403864 bytes have been tenured, next gc will be global.")
# [05:52] * Quits: doublec (n=doublec@202.180.114.137)
# [05:59] * Joins: dbaron (n=dbaron@c-71-198-189-81.hsd1.ca.comcast.net)
# [06:31] * Joins: aa (i=aa@nat/google/x-a07d158df2be9d04)
# [06:36] * Quits: dbaron (n=dbaron@c-71-198-189-81.hsd1.ca.comcast.net) ("8403864 bytes have been tenured, next gc will be global.")
# [07:10] * Joins: aroben (n=adamrobe@c-67-160-250-192.hsd1.ca.comcast.net)
# [07:10] * Quits: aroben (n=adamrobe@unaffiliated/aroben) (Remote closed the connection)
# [07:38] * Quits: csarven (n=nevrasc@modemcable081.152-201-24.mc.videotron.ca) ("http:/www.csarven.ca")
# [07:39] * Quits: weinig_ (i=weinig@nat/apple/x-077d52b93e9d8a74)
# [07:47] * Quits: KevinMarks (i=KevinMar@nat/google/x-71de07520e7be6f3) ("The computer fell asleep")
# [08:03] * Quits: MikeSmith (n=MikeSmit@eM60-254-218-189.pool.emnet.ne.jp) ("Less talk, more pimp walk.")
# [08:32] * Joins: MikeSmith (n=MikeSmit@eM60-254-219-89.pool.emnet.ne.jp)
# [08:39] <Lachy> Hixie, http://canvex.lazyilluminati.com/misc/cgi/issues.cgi/message/%3Ca9699fd20704042335r113597a2u343d8134cb28e0f4%40mail.gmail.com%3E :-)
# [08:42] <Lachy> I think it's base64 encoded, so it's unreadable
# [08:44] <Lachy> same thing has happened to quite a few messages
# [09:01] * Quits: gsnedders (n=gsnedder@host86-145-188-203.range86-145.btcentralplus.com) (Read error: 104 (Connection reset by peer))
# [09:02] * Joins: gsnedders (n=gsnedder@host86-145-188-203.range86-145.btcentralplus.com)
# [09:30] <Lachy> cool, I just noticed there's a new French translation on the blog http://blog.whatwg.org/pourquoi-le-texte-alternatif-peut-etre-omis-french - it was translated within 3.5 hours :-)
# [09:35] * Quits: karlUshi (n=karl@133.27.247.173) ("Where dwelt Ymir, or wherein did he find sustenance?")
# [09:41] * Quits: mpt (n=mpt@121-72-138-60.dsl.telstraclear.net) ("Leaving")
# [09:46] * Joins: ROBOd (n=robod@86.34.246.154)
# [09:50] <MikeSmith> I wasn't aware the it had been decided that alternate text was the first principle of accessibility.
# [09:51] <MikeSmith> So, hey, I learned something today.
# [09:55] <MikeSmith> It's like I asked Pope Ratzinger: Dude, a lot of that stuff you talk about all the time -- that you claim is gospel truth -- well, I can't actually find mention of a lot of that stuff in the Bible.
# [09:55] <MikeSmith> But at least the Catholic church calls it exactly what it is, which is Dogma.
#
# Session Start: Fri Aug 24 09:56:29 2007
# Session Ident: #whatwg
# [09:56] * Now talking in #whatwg
# [09:56] * Topic is 'WHATWG (HTML5) -- http://www.whatwg.org/ -- Logs: http://krijnhoetmer.nl/irc-logs/ -- Please leave your sense of logic at the door, thanks!'
# [09:56] * Set by Hixie on Tue Apr 03 04:10:22
# [09:57] <MikeSmith> http://en.wikipedia.org/wiki/Dogma
# [09:57] <MikeSmith> [[
# [09:57] <MikeSmith> established belief or doctrine held by a religion, ideology or any kind of organization, thought to be authoritative and not to be disputed or doubted.
# [09:57] <MikeSmith> ]]
# [09:58] <MikeSmith> "Rejection of dogma is considered heresy in certain religions, and may lead to expulsion from the religious group."
# [09:58] <Lachy> so you're saying we shouldn't be questioning the wisdom of the accessibility community, and just do what they say?
# [09:58] <krijnh> Hmm, what's http://html4all.org/ ?
# [10:02] <MikeSmith> Lachy - I can't answer that question for you. I guess it depends on whether you put your trust in Faith or Reason.
# [10:04] <Lachy> well, I've never been one to accept dogma of any kind, so at the risk of being hated by the accessibility community, I choose reason and logic
# [10:05] <Lachy> krijnh, looks like John Foliot owns that wiki
# [10:05] <krijnh> Lachy: Ah
# [10:05] <Lachy> http://www.dnsstuff.com/tools/whois.ch?ip=html4all.org
# [10:05] * Quits: yod (n=ot@softbank221018155222.bbtec.net) ("Leaving")
# [10:07] <MikeSmith> Lachy - I think it might not be all the productive to accept the usage of the term "accessibility community"
# [10:07] <Lachy> well, how else should I refer to the group of people who obviously dislike me and are typically involved with accessibility?
# [10:08] <MikeSmith> Some of those objecting to the current HTML5 spec makes it sound like there's this unified monolithic group with a single set principles that they all agree on.
# [10:08] <krijnh> The anti-Lachsessibility people
# [10:08] <MikeSmith> And that they are speaking for that community.
# [10:09] <Lachy> well, I know they don't speak for the whole community, but they're certainly the most vocal
# [10:09] <MikeSmith> like the Moral Majority
# [10:09] <krijnh> Same with some HTML5 people
# **inserts Lachy's Log** :)
# [18:10] <MikeSmith> krijnh - true
# [18:11] <MikeSmith> The lesson from it I guess it that everybody ought to be continuously checking/questioning their own assumptions.
# [18:12] <MikeSmith> It's like that moment in the Matrix, blue pill or red pill
# [18:12] -->| KevinMarks (n=KevinMar@c-76-102-254-252.hsd1.ca.comcast.net) has joined #whatwg
# [18:13] <MikeSmith> Do you want to see things as they really are, even if it's ugly and scary? Or do you want to keep seeing things as you've been comfortable seeing them and assumed they really are?
# [18:15] <Lachy> I already accepted the ugly and scary reality 3 years ago
# [18:17] <MikeSmith> Well, I think you're going to continue to have people strolling in, looking at selective parts of the HTML5 work from their own blinkered perspective, and saying, You guys are really making a big mess of things. When the reality is of course, No, the mess is already all around you, and has been for a long time.
# [18:22] <krijnh> Ping
# [18:22] <Hixie> at the risk of offending people, from what i can tell from speaking to people who really are part of the "accessibility comunity", most of the vocal complainers about accessibility in the htmlwg aren't actually really part of the "real" accessibility community
# [18:22] |<-- krijnh has left freenode ("...time for something different...")
# [18:22] -->| krijnh (n=krijnhoe@ktk.xs4all.nl) has joined #whatwg
# [18:23] <krijnh> Sorry, my connection dropped :/
# [18:25] <MikeSmith> Hixie - I can imagine that in some cases, they could in fact be individuals who have actually had similar combative discussions with people in the WAI WGs
# [18:25] <MikeSmith> Just hypothetically, I'm imagining.
# [18:27] <MikeSmith> Anyway, it's really hard at times to see how to engage in discussion with them productively.
# [18:27] -->| zcorpan (n=zcorpan@84-216-43-53.sprayadsl.telenor.se) has joined #whatwg
# [18:27] <MikeSmith> If they are not willing to consider questioning whether some of their views are not in fact just dogma.
# [18:28] |<-- Philip` has left freenode (Read error: 104 (Connection reset by peer))
# [18:28] <MikeSmith> like that "alternate text is the first principle of accessibility" statement in the comments to Lachy's post about Alt
# [18:28] <Lachy> Hixie, do you mean there are others more involved with accessibility who actually agree with your controversial decisions?
# [18:28] <MikeSmith> as if that is self-evident or axiomatic or something
# [18:29] -->| bzed (n=bzed@dslb-084-059-113-214.pools.arcor-ip.net) has joined #whatwg
# [18:29] * MikeSmith notes that my last statement was continuation of my previous one, not a response to what Lachy just said
# [18:30] <Hixie> Lachy: i wouldn't necessarily go that far :-P
# [18:32] * MikeSmith needs to wander off for a bit
# [18:32] <MikeSmith> back on later
# [18:32] |<-- MikeSmith has left freenode ("Less talk, more pimp walk.")
# [18:33] <Hixie> i guess tomorrow unless someone sees something wrong with it i'll write up the offline storage stuff
# [18:42] |<-- krijnh has left freenode (Read error: 110 (Connection timed out))
# [18:48] -->| MikeSmith (n=MikeSmit@eM60-254-212-215.pool.emnet.ne.jp) has joined #whatwg
# [18:49] <Dashiva> Hixie: At the risk of being quoted again, it's missing a longapplication attribute ;)
# [18:49] <Hixie> no comment! :-P
# [18:50] <Dashiva> One thing I thought about was the ability to run more than one copy of the same app, but I suppose that could be handled in the app itself
# [18:53] <MikeSmith> Hixie, Lachy - regarding the earlier discussion, I wonder why you guys might think of the idea of sort of agreeing to declare a moratorium on discussion of document conformance issues on public-html.
# [18:54] <MikeSmith> What I mean is, no discussions on public-html at this point about whether a particular element or attribute should/will be dropped, whether or not is should be conformant or non-conformant.
# [18:54] <Lachy> so we can focus more on processing requirements?
# [18:55] <MikeSmith> Exactly
# [18:55] <MikeSmith> on interoperability issues
# [18:55] <MikeSmith> On the conformance criteria for implementations, not for documents.
# [18:55] <Hixie> MikeSmith: that's up to DanC
# [18:55] <Hixie> MikeSmith: on whatwg, i welcome all feedback, on any topic
# [18:55] <Lachy> some people might object to that. There are people who think we should only define processing requirements for conforming documents
# [18:56] <MikeSmith> Lachy - I don't think that the charter would justify restricting discussion of processing requirements to only conforming documents.
# [18:56] <Lachy> indeed. I'm just saying why some people would still object
# [18:57] <MikeSmith> I know
# [18:57] <Hixie> people will object whatever we do
# [18:57] <Hixie> i don't think trying to reduce objections will be very productive to be honest
# [18:57] <MikeSmith> Hixie - why?
# [18:57] <Hixie> i think our time is better spent sifting through the feedback and summarising it into problem statements
# [18:57] <Lachy> well, something needs to be done with public-html. I'm just not sure what would help
# [18:58] <Hixie> MikeSmith: why to which?
# [18:58] <Hixie> the only thing that imho needs to be done to whatwg is the stopping of people's rants being cross-posted to whatwg
# [18:58] <Hixie> i was unhappy when i saw john's rant
# [18:59] <MikeSmith> Why you don't think it would productive to try to reduce objections, I meant. But anyway, my point was not about trying to reduce objections, it was about trying to get people's attention on the parts of the work that are a much bigger priority.
# [19:00] <MikeSmith> I would say the same thing about such rants being cross-posted to public-html
# [19:00] <Hixie> oh, because i don't think it's possible. i think you could reduce the number of vocalised objections, but bottling them up is only going to cause more pain later.
# [19:00] <Lachy> such rants shouldn't be posted anywhere
# [19:00] -->| cplot (n=cplot@c-98-193-22-194.hsd1.il.comcast.net) has joined #whatwg
# [19:01] <Hixie> Lachy: no argument from me there, but that's up to the chairs of each group. i don't find crap like that on whatwg acceptable behaviour.
# [19:01] <Hixie> as i think i made clear to him
# [19:01] <Hixie> we'll see
# [19:01] |<-- bzed has left freenode (Read error: 110 (Connection timed out))
# [19:01] <MikeSmith> I seriously wonder if we should perhaps start another list, public-html-conformance or something
# [19:01] -->| Whiskey_M (n=Richard@host-84-9-127-20.bulldogdsl.com) has joined #whatwg
# [19:02] <Lachy> we've already got 3 lists for the HTMLWG
# [19:02] <Whiskey_M> Good morning one and all :-)
# [19:02] <Lachy> I'm not sure we need a 4th
# [19:02] <Hixie> more lists doesn't help me at all, since i just end up getting the same amount of mail, except i need to work out which one to post to
# [19:03] <MikeSmith> Hixie - what would you think about freezing the document-conformance parts of the spec for now? Not making any further changes with regard to whether a particular element or attribute is conformant or not?
# [19:03] <Lachy> we've got public-html, public-html-wg-announce, public-html-comments, whatwg, whatwg-implementers, whatwg-help and now public-forms-tf. 7 lists directly related to the work on the spec is enough
# [19:04] <MikeSmith> Lachy - I guess I suggested it just as a way to reduce traffic on public-html
# [19:04] <Hixie> MikeSmith: i'm just gonna be going down the list of feedback
# [19:04] <Hixie> MikeSmith: if it involves document conformance, sobeit
# [19:04] <MikeSmith> to focus discussion on public-html, for the time being, around spec'ing implementation behavior
# [19:05] <Hixie> you won't reduce traffic to public-html
# [19:05] <Hixie> there are 400 people subscribed to it
# [19:06] <Hixie> (hmm, whatwg is up to 804)
# [19:06] <MikeSmith> Hixie - You don't think that would at least help to reduce some of the unproductive, contentious discussion?
# [19:07] <MikeSmith> I mean, how unproductive or contentious could a discussion about the tokenizer, parser, DOM behavior etc., be?
# [19:08] <Lachy> probably not, I doubt the contentious issues are still going to come up in comments until they get their own way (note, even in the design principles survey, people used it as a platform to voice their concern about various contentious issues)
# [19:08] <Hixie> MikeSmith: the topic of discussion is not why the discussion is unproductive
# [19:08] <Lachy> s/doubt/think/
# [19:08] <MikeSmith> Hixie - what is the reason, then?
# [19:10] <MikeSmith> Hixie - in going through the list of feedback, you're of course not under any obligation to do it in serial fashion. You can prioritize what to focus on, and I'm sure you do already, anyway.
# [19:10] <MikeSmith> You can choose to make responding to feedback about conformance a lower priority if you want.
# [19:11] <Hixie> actually my priorities tend to be strongly influenced by what browser vendors are working on
# [19:11] <Hixie> i regularly get private requests to fast track a particular feature area because they are going to fix bugs in that area shortly
# [19:11] <Hixie> it's one of the best ways to ensure that the browsers actually follow the spec
# [19:12] <MikeSmith> Hixie - right, which would seem to make me think most of the conformance criteria is not a priority at all
# [19:12] <Hixie> (and one of the vendors in question is henri, who's writing a conformance checker, so conformance is a priority sometimes)
# [19:13] <MikeSmith> So Henri's conformance checker is if equal priority to you as browser implementations are?
# [19:15] <Hixie> Henri's goodwill is of high priority to me, yes
# [19:15] <Hixie> he has been a huge help to the spec
# [19:16] <MikeSmith> I recognize that of course
# [19:17] <MikeSmith> But the issue is not Henri personally, it's whether at this point the needs of having a spec that's useful for building a conformance checker are of the same priority as having a spec that's useful for solving browser interoperability problems
# [19:18] <MikeSmith> Or the relative market demand for a interoperable browsers vs. conformance checkers
# [19:19] <Hixie> it's all important
# [19:25] -->| BenWard (i=BenWard@nat/yahoo/x-11743c5184725a4b) has joined #whatwg
# [19:28] <Hixie> nn
# [19:28] <Lachy> bye hixie
# [19:33] <MikeSmith> so, about some of the most recent messages to public-html on the Offline Web Apps thread: If, hypothetically, you see somebody posting a massive number of long messages to mailing list over a long period of time, and as far as you can tell that person's day job -- professional/educational responsibilities, whatever -- doesn't seem to related at all to the HTML work or even to Web technologies in general...
# [19:34] <MikeSmith> ...it sorta might make one wonder if they are actually doing the job that their employer/advisor/whoever is actually paying/expecting them to be doing right now
# [19:35] <MikeSmith> And if so, where they actually find the time to do that, between the sending of huge volumes of e-mail
# [19:36] -->| hendry (n=hendry@nox.vm.bytemark.co.uk) has joined #whatwg
# [19:37] <MikeSmith> But I guess it would be none of my business, anyway, to question such a person. If such were to exist, hypothetically
# **end Lachy's Log**[14:14] * Disconnected
# [14:14] * Attempting to rejoin channel #whatwg
# [14:14] * Rejoined channel #whatwg
# [14:14] * Topic is 'WHATWG (HTML5) -- http://www.whatwg.org/ -- Logs: http://krijnhoetmer.nl/irc-logs/ -- Please leave your sense of logic at the door, thanks!'
# [14:14] * Set by Hixie on Tue Apr 03 04:10:22
# [14:31] * Quits: grimboy_uk (n=grimboy@85-211-254-67.dsl.pipex.com) (Read error: 110 (Connection timed out))
# [14:31] * Joins: grimboy_uk (n=grimboy@85-211-254-165.dsl.pipex.com)
# [14:33] * Joins: grimeboy (n=grimboy@85-211-255-226.dsl.pipex.com)
# [14:38] * moeffju[1zZz] is now known as moeffju
# [14:47] * Joins: polin8 (n=brian@dsl081-134-176.nyc1.dsl.speakeasy.net)
# [14:49] * Joins: grimboy (n=grimboy@85.211.238.15)
# [14:55] * Quits: grimboy_uk (n=grimboy@85-211-254-165.dsl.pipex.com) (Read error: 110 (Connection timed out))
# [14:59] * Joins: grimboy_uk (n=grimboy@85-211-253-135.dsl.pipex.com)
# [15:04] * Quits: grimeboy (n=grimboy@85-211-255-226.dsl.pipex.com) (Read error: 110 (Connection timed out))
# [15:14] <kjetilkWork> virtuelv: I just assigned to give a talk about XML parsing vs. JSON parsing too, and why JSON is interesting
# [15:14] <kjetilkWork> of course, I have no clue about that, got any good pointers? :-)
# [15:15] <kjetilkWork> got the Opera static build running on amd64, BTW
# [15:16] * Quits: dolphinling (n=chatzill@rbpool1-96.shoreham.net) ("Off to school!")
# [15:20] * Quits: grimboy (n=grimboy@85.211.238.15) (Read error: 110 (Connection timed out))
# [15:30] <virtuelv> kjetilkWork: json.org?
# [15:32] <kjetilkWork> virtuelv: I was thinking independent speed benchmarking and that kind of stuff
# [15:32] <virtuelv> no clue
# [15:33] <virtuelv> I fear we're running wildly off-topic, though
# [15:33] <kjetilkWork> ok
# [15:43] * Quits: Hixie (i=ianh@trivini.no) (Remote closed the connection)
# [15:56] * Joins: hober (n=ted@unaffiliated/hober)
# [16:01] * Joins: Ducki_ (n=Ducki@nrdh-d9b980d4.pool.mediaWays.net)
# [16:03] * Quits: Oeighty (n=oxygen80@ip-58-28-200-134.ubs-dsl.xnet.co.nz) (Read error: 110 (Connection timed out))
# [16:07] * Joins: Hixie (i=ianh@trivini.no)
# [16:25] * Quits: Ducki (i=Ducki@nrdh-d9b98052.pool.mediaWays.net) (Read error: 113 (No route to host))
# [16:25] * Joins: billmason (n=billmaso@ip156.unival.com)
# [16:44] * Quits: bzed (n=bzed@dslb-084-059-108-228.pools.arcor-ip.net) (Client Quit)
# [17:19] * Joins: jgraham_ (n=jgraham@guest059.wtgc.org)
# [17:20] * Quits: jgraham_ (n=jgraham@guest059.wtgc.org) (Client Quit)
# [17:51] * Joins: weinig_ (i=weinig@nat/apple/x-3ffe1fa71cbb7ae0)
# [17:53] * Quits: KevinMarks (n=KevinMar@c-76-102-254-252.hsd1.ca.comcast.net) ("The computer fell asleep")
# [17:58] * Joins: a-ja (n=chatzill@70.230.180.123)
# [17:59] * Quits: a-ja (n=chatzill@70.230.180.123) (Client Quit)
# [18:15] * Quits: Ducki_ (n=Ducki@nrdh-d9b980d4.pool.mediaWays.net) (Read error: 104 (Connection reset by peer))
# [18:28] * Quits: weinig_ (i=weinig@nat/apple/x-3ffe1fa71cbb7ae0)
# [18:30] * Joins: weinig_ (i=weinig@nat/apple/x-6defccba00b70364)
# [18:31] * Quits: Whiskey_M (n=Richard@host-84-9-127-20.bulldogdsl.com)
# [18:44] * Quits: BenWard (i=BenWard@nat/yahoo/x-11743c5184725a4b) ("Fades out again…")
# [18:46] * Quits: gsnedders (n=gsnedder@host86-145-188-203.range86-145.btcentralplus.com)
# [18:50] * Quits: ROBOd (n=robod@86.34.246.154) ("http://www.robodesign.ro")
# [18:52] <zcorpan> wonder if we only want to do fancy html attribute processing in quirks mode
# [18:52] <zcorpan> color attribute processing that is
# [18:53] <zcorpan> ie7 does it in standards mode as well, but opera, safari and firefox limit it to quirks mode
# [18:57] * Joins: gsnedders (n=gsnedder@host86-145-188-203.range86-145.btcentralplus.com)
# [19:01] * Joins: maikmerten (n=maikmert@T649e.t.pppool.de)
# [19:07] * Joins: grimeboy (n=grimboy@85.211.239.84)
# [19:08] * Joins: kingryan (n=kingryan@corp.technorati.com)
# [19:26] * Quits: weinig_ (i=weinig@nat/apple/x-6defccba00b70364)
# [19:27] * Joins: weinig_ (i=weinig@nat/apple/x-5c0a84a0af185374)
# [19:29] * Joins: aroben (n=adamrobe@17.203.15.195)
# [19:29] * Quits: grimboy_uk (n=grimboy@85-211-253-135.dsl.pipex.com) (Read error: 110 (Connection timed out))
# [19:30] * moeffju is now known as moeffju[Away]
# [19:34] * Joins: KevinMarks (i=KevinMar@nat/google/x-4abaa82f0cc750fc)
# [19:47] * Joins: Codler (n=Codler@84-218-7-177.eurobelladsl.telenor.se)
# [19:52] * Quits: KevinMarks (i=KevinMar@nat/google/x-4abaa82f0cc750fc) ("The computer fell asleep")
# [20:08] * Joins: nickshanks (n=nickshan@p5498C2E3.dip0.t-ipconnect.de)
# [20:08] <nickshanks> how are things in here? sorry i've not been around for a while, have been burdened by getting a job
# [20:12] * Joins: KevinMarks (i=KevinMar@nat/google/x-aa6b5f9c9928770c)
# [20:37] * Joins: ROBOd (n=robod@86.34.246.154)
# [20:39] <Hixie> John just replied to the thread I told him to not forward to whatwg.
# [20:39] * Hixie goes to cycle to work to cool off and not explode at him
# [20:41] <nickshanks> molly was talking about lots of angst on her blog too. what happened?
# [20:41] * Quits: weinig_ (i=weinig@nat/apple/x-5c0a84a0af185374)
# [20:42] <takkaria> nickshanks: lots and lots of misunderstanding and not listening
# [20:58] * Quits: Codler (n=Codler@84-218-7-177.eurobelladsl.telenor.se) ("- nbs-irc 2.21 - www.nbs-irc.net -")
# [21:01] * Quits: KevinMarks (i=KevinMar@nat/google/x-aa6b5f9c9928770c) ("The computer fell asleep")
# [21:22] * Quits: aroben (n=adamrobe@unaffiliated/aroben)
# [21:31] * Quits: nickshanks (n=nickshan@p5498C2E3.dip0.t-ipconnect.de) (Read error: 113 (No route to host))
# [21:31] <Dashiva> So, where do we get our "Member of the IRC cabal" t-shirts?
# [21:32] <Philip`> Was http://www.spreadshirt.com/shop.php?sid=108788 popular enough to justify a new design?
# [21:33] <othermaciej> I just kind of exploded at him
# [21:34] <Dashiva> I never liked that one, Philip`. It's too obviously true for peopel to ponder the implications
# [21:35] * Joins: KevinMarks (i=KevinMar@nat/google/x-54030165cc7ec3a8)
# [21:35] <Philip`> It's so obviously true that people will ponder the implications of why you felt the need to express that statement on your clothing
# [21:36] <jgraham> I hope John realises he is not doing the term "accessibility advocate" any favours
# [21:36] <hober> I get asked about it every single time I wear it.
# [21:36] * gsnedders still hasn't ordered one
# [21:36] <hober> "So what's the deal with your shirt? Of course 5 is greater than 2."
# [21:37] <gsnedders> I'd just laugh.
# [21:37] <gsnedders> as myself and othermaciej argued, the joke is lost if you actually give it any context, and it becomes too geeky
# [21:38] <Dashiva> jgraham: As indicated earlier, I really like "IRC cabal" however :)
# [21:39] <Philip`> Is it really a cabal if it publishes logs of everything that goes on?
# [21:40] <gsnedders> Philip`: we're totally deniable.
# [21:41] <gavin_> Philip`: are you looking to start log-editing conspiracies? ;)
# [21:42] * Quits: grimeboy (n=grimboy@85.211.239.84) (Read error: 110 (Connection timed out))
# [21:43] * Joins: grimeboy (n=grimboy@85-211-247-176.dsl.pipex.com)
# [21:44] <gsnedders> meh. even the small 5 > 2 shirt might be too big
# [21:44] <Philip`> It seems quite helpful to have a group of people who tend to think along similar lines and are happy to communicate with each other frequently, in terms of making progress in some direction
# [21:45] <Philip`> (It's much less good if that's a bad direction; but I'd think it's still better than making no progress at all)
# [21:54] * Joins: aroben (n=adamrobe@17.203.15.195)
# [21:55] <jgraham> Philip`: The htmlwg is making the best attempt it can at making no progress at all
# [21:56] <hober> not to promote my own joke too much, but we could always make "vast, browser-wing conspiracy" t-shirts (http://www.mail-archive.com/www-archive@w3.org/msg00554.html)
# [21:56] <hober> maybe that's too US-centric?
# [21:57] <gavin_> heh
# [22:07] * Quits: maikmerten (n=maikmert@T649e.t.pppool.de) ("Leaving")
# [22:13] * Joins: weinig_ (i=weinig@nat/apple/x-d1314cb74fc13ca3)
# [22:16] <Dashiva> The implementati
# [22:17] * Joins: G0k (n=hmason@rrdhcp212-308.redrover.cornell.edu)
# [22:17] * Quits: KevinMarks (i=KevinMar@nat/google/x-54030165cc7ec3a8) ("The computer fell asleep")
# [22:43] * Hixie sighs deeply
# [22:44] <Hixie> i didn't want to do that
# [22:44] <Hixie> i wish people would not ignore my requests and not insult the whatwg community
# [22:45] <Hixie> on an unrelated note, gsnedders: if you leave the htmlwg please do remain in the whatwg, your feedback is very useful
# [22:45] <gsnedders> Hixie: I have no intention of leaving WHATWG
# [22:45] <Hixie> cool
# [22:45] * Parts: G0k (n=hmason@rrdhcp212-308.redrover.cornell.edu)
# [22:45] * Joins: G0k (n=hmason@rrdhcp212-308.redrover.cornell.edu)
# [22:45] * Parts: G0k (n=hmason@rrdhcp212-308.redrover.cornell.edu)
# [22:45] * Quits: ROBOd (n=robod@86.34.246.154) ("http://www.robodesign.ro")
# [22:46] <gsnedders> I would not be amazed if the WG needs its charter re-examined due to implementers pulling out
# [22:46] * Joins: G0k (n=hmason@rrdhcp212-308.redrover.cornell.edu)
# [22:46] <Hixie> that would be so sad
# [22:47] <gsnedders> On an unrelated note, I've started work on an I-D regarding HTTP request/response parsing
# [22:47] <Hixie> interesting
# [22:47] <Hixie> good luck with that
# [22:48] <Hixie> the http community is not one that really believes in defined error handling
# [22:48] <Hixie> or in caring about browsers
# [22:48] <gsnedders> look at RFC2616, which already has a section on tolerant applications
# [22:48] <Hixie> (see, e.g. the cool reception that the suggestion to remove Content-Location received)
# [22:48] <Hixie> yeah
# [22:48] <Hixie> but they don't like defining what that means, in my experience :-)
# [22:49] <othermaciej> gsnedders: it would be sad if a minority of abusive people could drive out useful contributors; I hope it doesn't come to that
# [22:49] <gsnedders> I've put various things like that (cont-location) in an informative appendix, as it is really out of scope
# [22:49] <othermaciej> what's the difference between Location and Content-Location supposed to be?
# [22:49] <Hixie> Content-Location is like <base href>
# [22:49] <Hixie> except a huge number of servers send out bogus values
# [22:49] <Hixie> so implementing it per spec breaks sites
# [22:49] <Hixie> a lot of them
# [22:49] <gsnedders> IIS bug. documented by MS.
# [22:50] <othermaciej> oh
# [22:50] <Hixie> iirc it's not just IIS, but yeah
# [22:50] <gsnedders> Hixie: I looked through various things before pulling support for it myself, I never saw anything that wasn't IIS with the bug
# [22:50] <Hixie> hm ok
# [22:50] <Hixie> what did you pull support from?
# [22:51] * Hixie goes to get food and will then write up the offline web apps idea in the spec
# [22:51] <gsnedders> a _really_ shit implementation of HTTP/1.0 thrown together really quickly for a feed library, namely SimplePie
# [22:51] <Hixie> so if you have comments on that, especially "it doesn't work", mention them soon :-)
# [22:51] <Hixie> gsnedders: aah
# [22:51] <G0k> Hixie: i had a thought about the offline web apps thing. For multi-page apps, could you use <link> elements? like <link rel="application resource" href="someotherpage.html" />
# [22:51] <othermaciej> Hixie: I think there's lots of smaller issues besides the pretty big one (IMO) of not really working for multi-page apps
# [22:51] <othermaciej> Hixie: I don't think I'll have time to review it in detail today though
# [22:52] * Joins: aroben_ (n=adamrobe@unaffiliated/aroben)
# [22:52] <Hixie> i think aaron's idea (to just have an API to initiate the caching of another top-level page app) solves the multipage issue
# [22:52] <G0k> like a DOM API?
# [22:52] <othermaciej> Hixie: you can't cache the other page correctly without executing all scripts in it
# [22:52] <Hixie> G0k: yeah, though we could have a declarative thing too
# [22:52] <othermaciej> that seems obviously unacceptable
# [22:52] <Hixie> othermaciej: sure, you just use the background window thing just like the update process
# [22:52] <gsnedders> the bug is fixed in IIS7, from what I can see on the web
# [22:52] <Hixie> what is obviously unacceptable?
# [22:53] <othermaciej> Hixie: how does the background window avoid executing scripts?
# [22:53] <Hixie> othermaciej: why would it need to avoid executing scripts?
# [22:53] <G0k> i feel like anything that depends on scripting being executed makes this overly complicated
# [22:53] <Hixie> othermaciej: you want it to run scripts.
# [22:53] <G0k> what if you want a web app that's static?
# [22:53] <Hixie> not much of an app, then
# [22:53] <Hixie> normal http caching already solves the "offline web page" problem
# [22:53] <othermaciej> Hixie: executing all scripts on a 50-page site just to cache all 50 pages seems like a large and probably unacceptable performance cost
# [22:54] * Quits: polin8 (n=brian@dsl081-134-176.nyc1.dsl.speakeasy.net)
# [22:54] <Hixie> othermaciej: what web app uses 50 pages?
# [22:54] <G0k> right but http caching doesn't do multi-page stuff
# [22:54] <Hixie> G0k: sure, just crawl the site. even IE4 can do that.
# [22:54] <othermaciej> Hixie: something that used history state with URI updating could easily use a lot more than 50 (where nearly all resources are actually shared)
# [22:54] <othermaciej> Hixie: I like the seductive simplicity of your idea
# [22:55] <othermaciej> Hixie: but I think many more complex cases are much better addressed with an explicit manifest of some kind
# [22:55] <othermaciej> the manifest could still be in the markup
# [22:55] <othermaciej> and you could recognize that multiple pages should use the same cache by the fact that they link the same manifest
# [22:55] <othermaciej> then you can pre-cache a bunch of stuff without the need for redundant script execution or HTML parsing
# [22:55] <Hixie> othermaciej: none of the applications i looked at (admittedly mostly google ones) actually need multipage caching, i'm somewhat reluctant to support that without some pretty concrete examples.
# [22:56] <Hixie> especially given that it makes it more complex by an order of magnitude for pretty much everyone
# [22:56] <Hixie> brb fod
# [22:56] <Hixie> fod
# [22:56] <G0k> yeah i mean i feel like we should have a real semantic way of saying "I want you to cache this other page"
# [22:56] <Hixie> foOd
# [22:56] <G0k> or maybe more simply
# [22:56] <othermaciej> there's lots of multi-page web apps out there
# [22:56] <othermaciej> I'm not sure how many would benefit strongly from an offline mode
# [22:56] <G0k> "This other page should have the same caching policy as me"
# [22:56] <othermaciej> flickr and upcoming are two very popular and obvious examples
# [22:57] <G0k> how about word processor?
# [22:57] <G0k> google write or whatever
# [22:57] <othermaciej> digg and reddit would also be examples of multi-page web apps
# [22:57] <G0k> why would you want digg offline?
# [22:58] <gsnedders> Basecamp?
# [22:58] <othermaciej> G0k: maybe I visit it when offline, and then I want to be able to read offline at any time and see more than the front page
# [22:58] <gsnedders> A lot of 37 Signals stuff could do with being usable offline
# [22:59] <G0k> why not just have the server properly give cache control instructions to the UA so that it does that?
# [23:00] <othermaciej> you mean http cache-control?
# [23:01] <G0k> yeah
# [23:02] <othermaciej> cache-control doesn't give a way to get the UA to use possibly-stale versions of the content when offline
# [23:02] <othermaciej> nor to update a group of resources atomically
# [23:02] <G0k> doesn't that what offline mode does?
# [23:03] <G0k> *isn't
# [23:03] <G0k> hm
# [23:04] <G0k> i guess cache-control doesn't have a way to say "cache me indefinitely"
# [23:04] * Quits: grimeboy (n=grimboy@85-211-247-176.dsl.pipex.com) (Read error: 110 (Connection timed out))
# [23:04] * Joins: grimeboy (n=grimboy@85.211.236.253)
# [23:05] <G0k> although you could say....make it not expire for 1000 years
# [23:05] <G0k> but then even then, the UA could decide not to cache anyway
# [23:05] <Hixie> that'd be bad, the client might never check that the file had changed :-)
# [23:06] * Quits: aroben (n=adamrobe@unaffiliated/aroben) (Read error: 110 (Connection timed out))
# [23:06] <G0k> well i feel like people should start changing the entire model of their app designs
# [23:06] <G0k> rather than sending a huge chunk of content + design
# [23:06] <G0k> just send an "empty" shell of the app
# [23:06] <G0k> then fetch content seperately
# [23:06] <G0k> thus the shell part never expires
# [23:07] * Hixie looks at flickr, upcoming, digg, reddit
# [23:07] <G0k> and doesn't have to keep getting fetched
# [23:07] <G0k> maybe use event-source to get new stories. :)
# [23:07] <Hixie> heh
# [23:07] <takkaria> I don't see that flickr is a useful thing to use offline
# [23:07] <Hixie> G0k: http://www.digg.com/spy
# [23:08] <G0k> hixie: yeah i think they using polling
# [23:08] <G0k> or the long download technique
# [23:09] <Hixie> that's what whatwg.org/issues/top uses
# [23:09] <Hixie> i hate it
# [23:09] <Hixie> bring on TCPConnection
# [23:09] <G0k> ok that's a question i have too
# [23:09] <G0k> what's the point of making TCPConnection use that weird protocol
# [23:09] <G0k> ?
# [23:10] <G0k> why not make it a real raw TCP connection?
# [23:10] <takkaria> security risks?
# [23:10] <G0k> such as?
# [23:10] <Hixie> yeah, you can't possibly have raw sockets flying about
# [23:11] <takkaria> well, what's the point in having cross-domain security restrictions if you can just write your own HTTP client in JS with none of them?
# [23:11] <G0k> well not a raw socket, just a pure tcp connection
# [23:11] <Hixie> what takkaria said
# [23:11] <Hixie> G0k: that's what i meant
# [23:11] <G0k> the TCP connections could still have the same domain security policy stuff
# [23:11] <Hixie> how?
# [23:12] <Hixie> TCP connections are IP-bound, not domain-bound
# [23:12] * Joins: grimboy_uk (n=grimboy@85-211-247-241.dsl.pipex.com)
# [23:12] <Hixie> hixie.ch and whatwg.org and damowmow.com are all at the same IP
# [23:12] <G0k> well ok, IP security at least
# [23:12] <G0k> so then what?
# [23:13] <gsnedders> see y'all tomorrow (and we'll see whether I'll still be in the WG then)
# [23:13] <othermaciej> same-domain isn't good enough since it still allows cross-protocol attacks
# [23:13] <othermaciej> the only way to offer raw TCP sockets that I've heard which is at all workable is to have a central control file served by HTTP from the server
# [23:14] <G0k> an evil person could hijack hixie.ch and then send TCP requests to some other service on that server?
# [23:14] <othermaciej> which would control access
# [23:14] <othermaciej> G0k: currently http servers running on two different ports on the same system are considered separate security domains
# [23:15] <othermaciej> takkaria: subsets of flickr could be useful offline - viewing photos I've viewed before, queuing new photos, descriptions and comments for upload on next connection, etc
# [23:15] <G0k> well this certainly wouldn't be any less secure than Java or Flash, which already allow you to connect to arbitrary services on the document's origin IP
# [23:15] <othermaciej> takkaria: it has basically the same offline usefulness as gmail
# [23:16] <takkaria> mm, OK
# [23:16] <takkaria> I use a dedicated photo upload tool so I don't get that so much
# [23:16] <G0k> i mean inventing this new protocol dramatically limits the usefulness of this thing
# [23:17] <G0k> plus it's a complete misnomer
# [23:17] <G0k> it's not a TCPConnection
# [23:17] <Hixie> yeah, the name should change
# [23:17] <Hixie> the great thing about the way it works now is that we can allow arbitrary cross-domain connections
# [23:17] <takkaria> I had visions of implementing an IRC client when I first read that name. :)
# [23:17] <G0k> well that may be a useful feature too
# [23:17] <othermaciej> well, if you have a local photo store and a native photo upload tool, then obviously you have less need for your web-based photo gallery to work offline
# [23:17] <G0k> but a real TCPConnection would be neat too
# [23:18] <othermaciej> but if you want to use your photo sharing site as the *only* place you manage your photos, then you do want it to work offline
# [23:19] <takkaria> othermaciej: yeah. fair enough. :)
# [23:20] <G0k> but yeah i mean that dom server side events thing was like...how i originally found WHAT WG
# [23:20] <G0k> working on an app which needs something just like that
# [23:21] <aa> with implicit gathering of resources for the offline application, how do apps request resources that they want to come from the server while loading?
# [23:21] <aa> as a simple example lots of offline apps request an image from the server during load as a health check
# [23:22] * Joins: KevinMarks (i=KevinMar@nat/google/x-215eb659117afb50)
# [23:24] <Hixie> aa: you mean as opposed to the scriptable API for adding a file to the cache?
# [23:25] <G0k> has that already been specified
# [23:25] <G0k> ?
# [23:25] <Hixie> no
# [23:25] <aa> I mean the opposite: you don't want something to go to the cache
# [23:25] <aa> which is a GET
# [23:25] <G0k> you use no-cache in the HTTP headers for that?
# [23:25] <Hixie> ah
# [23:26] <Hixie> yeah, there's no way to do a GET that doesn't hit the cache in the current idea
# [23:26] <Hixie> you can do a POST
# [23:26] <aa> seems heavy handed
# [23:26] <G0k> maybe you could add a delete-from-cache API?
# [23:26] <Hixie> aa: well, what's the use case?
# [23:26] <aa> plus it means you can't use ye old new Image() trick
# [23:26] <Hixie> if you want to ping the server, getting an image seems like not the right thing to do :-)
# [23:27] * Quits: grimeboy (n=grimboy@85.211.236.253) (Read error: 110 (Connection timed out))
# [23:27] <G0k> should add an ICMPConnection. what could posssibly go wrong?
# [23:27] <aa> the use case is to determine whether your server is reachable
# [23:27] <aa> what reader and gearpad do is request x.gif?r=<rand>
# [23:27] <aa> probably that is Bad, but there it is
# [23:27] <Hixie> aa: well we can have an explicit API for that
# [23:27] <Hixie> aa: that seems like something you'd want anyway
# [23:28] <G0k> i think that's a spendid idea
# [23:28] <aa> Hixie: there are other use cases besides that
# [23:28] <G0k> *splendid
# [23:28] <Hixie> aa: like what?
# [23:28] <aa> Hixie: What if you want to talk to your server to synchronize during load.
# [23:29] <aa> not wait until onload
# [23:29] <aa> you have to use POST?
# [23:29] * Quits: G0k (n=hmason@rrdhcp212-308.redrover.cornell.edu)
# [23:31] <Hixie> aa: like to get the data?
# [23:32] <aa> yes
# [23:32] <Hixie> aa: wouldn't you just do a normal GET, and not care if you're offline or not?
# [23:33] <Hixie> hm
# [23:33] <aa> So the use case I'm imagining is using xmlhttprequest
# [23:33] <Hixie> though i see what you mean
# [23:33] <aa> maybe it could have a property to bypass the offline cache
# [23:33] <Hixie> you might want to have the data not cached
# [23:33] <Hixie> yeah
# [23:33] <aa> i think that would be enough
# [23:33] <Hixie> hmm
# [23:33] <Hixie> yeah
# [23:33] <Hixie> maybe that's the way to do it
# [23:34] <aa> another idea: maybe requests originating from javascript don't count for the auto-gathering
# [23:34] <aa> only html tags literally in the source
# [23:35] <aa> that seems weird too though :-)
# [23:35] <Hixie> yeah i don't think i like that
# [23:39] <othermaciej> it seems like when you consider all the details, implicit gathering might not be any easier to use than an explicit manifest
# [23:40] <Hixie> it's conceptually easy, you just pass everything to the app cache instead of your main cache
# [23:41] <othermaciej> I mean easier for the web app author - if the set of things they want cached needs to include items that aren't loaded and exclude items that are, then the implicit gathering plus API scattered around their app code might not be easier to use than a separate manifest file
# [23:42] <othermaciej> the manifest file can also allow low-cost checking for updates
# [23:42] <aa> FWIW, we have gotten pretty strong feedback that people don't like the manifest.
# [23:42] <othermaciej> how do they want it to work? all automatic?
# [23:43] <othermaciej> there's no way an all-automatic system can check for an update with a single conditional GET
# [23:43] <othermaciej> afaict
# [23:43] <aa> I think part of the problem is that the manifest is confusing during development because everything just works so long as you have a connection or you have the resources cached.
# [23:43] <aa> you don't find out that you forgot something until later.
# [23:43] <aa> i had proposed a mode where requests that are *not* in the manifest fail, but that of course brings up other questions.
# [23:44] <othermaciej> requests that are not in the manifest should fail when offline
# [23:44] <othermaciej> that makes it easy to test offline mode by actually going offline
# [23:46] <aa> it's not easy because most of the time during development you are not offline. also you have to be careful to clear your cache otherwise you still don't see it.
# [23:47] <othermaciej> yes, I'm saying you should not be able to get stuff from the cache when offline (at least when using a web app with a manifest)
# [23:47] <aa> what does "offline" mean?
# [23:47] <aa> file > offline?
# [23:48] <aa> maybe this is just a developer feature ... to not have the regular browser cache be consulted
# [23:48] <othermaciej> I don't understand your question
# [23:49] <othermaciej> also I need to step our for a bit, brb
# [23:49] <aa> at least in gears there is no concept of offline. so I dont know what you mean when you say the browser should change behavior when offline.
# [23:49] <othermaciej> Safari already won't load anything from the cache if your network is disconnected (maybe not the best behavior but we have not really gotten complaints)
# [23:50] <kingryan> othermaciej: I'd like to complain
# [23:50] <kingryan> :)
# [23:50] <aa> k, we can continue this later.
# [23:50] <aa> i don't want to keep you.
# [23:51] <othermaciej> one random parting thought before I step out
# [23:51] <Hixie> note that "offline" is somewhat nebulous -- you can be "online" (from the browser's point of view) but really your network is gonna be returning 302s everywhere (captive portal) or 500s (server down) or not resolving dns (wireless down) etc
# [23:51] <Hixie> i tried to handle that case too
# [23:51] <aa> Hixie: I think that is really important. Thanks.
# [23:52] <aa> (as a frequent user of unreliable wireless)
# [23:52] <othermaciej> maybe it can be an option for the manifest to say all resources from the current page load should be included, but list exceptions and additions
# [23:52] <Hixie> aa: yeah, from a personal perspective i know the feeling. :-)
# [23:52] <Hixie> othermaciej: i certainly don't mind there being a way to list resources declaratively (as well as through an API)
# [23:52] <othermaciej> Hixie: normal operating mode for the user handling that case is key, but I don't think that has to influence what happens when you pull your ethernet cable to test offline behavior
# [23:53] <Hixie> othermaciej: but i think that requiring a manifest is different from supporting a manifest as a secondary resource that causes loads
# [23:53] <othermaciej> Hixie: my brief mailing list proposal was to have an optional manifest (that would in practice be needed for multipage web apps)
# [23:54] <othermaciej> w/ the additional semantic that sharing a manifest makes you share an offline cache
# [23:54] <Hixie> yeah i'm still studying the multi-page case
# [23:54] <Hixie> i'm not sure i like the idea of having more than one "top-level page" per cache, but i'm still thinking about it
# [23:55] <othermaciej> I'm also not sure if in your proposal "subresources of the page" only includes things loaded up to onload time, or also things loaded later
# [23:55] <othermaciej> (either answer would raise issues I think)
# [23:55] <othermaciej> anyway
# [23:55] <othermaciej> bbl
# [23:56] <Hixie> ttyl
# [23:56] <Hixie> it's supposed to include anything you hit while online, the load event is only used when checking for an update to prime the new cache
# [23:58] * moeffju[Away] is now known as moeffju
# Session Close: Sat Aug 25 00:00:00 2007

The end :)