/irc-logs / freenode / #whatwg / 2008-01-04 / end

Options:

# Session Start: Fri Jan 04 00:00:00 2008
# Session Ident: #whatwg
# [00:00] * Quits: k0rnel (i=k0rnel@dup-dyn153.87-197-179.t-com.sk)
# [00:25] * Joins: othermaciej_ (n=mjs@dsl081-048-145.sfo1.dsl.speakeasy.net)
# [00:26] * Quits: othermaciej (n=mjs@dsl081-048-145.sfo1.dsl.speakeasy.net) (Read error: 104 (Connection reset by peer))
# [00:33] * Joins: aroben_ (n=aroben@unaffiliated/aroben)
# [00:36] * Joins: digx (n=rick@c-76-109-201-140.hsd1.fl.comcast.net)
# [00:38] * Quits: phsiao (n=shawn@nat/ibm/x-f34bb26647d32bed)
# [00:40] * Quits: othermaciej_ (n=mjs@dsl081-048-145.sfo1.dsl.speakeasy.net) (No route to host)
# [00:40] * Joins: othermaciej (n=mjs@dsl081-048-145.sfo1.dsl.speakeasy.net)
# [00:43] * Joins: Lfe (n=lfe@bergstroem.nu)
# [00:43] * Quits: Lfe_ (n=lfe@bergstroem.nu) (Read error: 104 (Connection reset by peer))
# [00:43] * Quits: aroben (n=aroben@unaffiliated/aroben) (Read error: 110 (Connection timed out))
# [00:54] <gsnedders> Hixie: http://www.digitalcitizen.info/2007/12/30/ogg-theoravorbis-as-default-for-video-scuttled-in-html5-spec-who-benefits/#comment-1487
# [01:01] * Joins: zcorpan (n=zcorpan@c-cb21e353.1451-1-64736c12.cust.bredbandsbolaget.se)
# [01:01] <zcorpan> http://tech.gtaero.net/2008/01/xhtml2-vs-html5.html -- hmm, someone should write a comment there. i don't have the energy to do so atm
# [01:04] * Joins: Lachy (n=Lachlan@cm-84.215.54.100.getinternet.no)
# [01:04] <Hixie> jruderman: spec doesn't say. feel free to send me e-mail if you think it should be defined
# [01:07] <Dashiva> gsnedders: Why is it that those people always take "won't implement something new unless there's an advantage" and turn it into "won't implement something new, ever"?
# [01:08] * jgraham thinks getElementsByClassName should be document order unless there is a good reason not to be (e.g. speed, implementation complexity)
# [01:11] <zcorpan> jgraham: agree (and i think it should be defined)
# [01:12] <zcorpan> same as getElementsByTagName
# [01:12] <jgraham> Yes, I meant it should be defined to be document order, unless there is a good reason not to make it document order in which case it should be explictly noted it is not
# [01:13] <jgraham> and the same as DOM 2 Range and DOM 3 XPath IIRC
# [01:13] <zcorpan> getElementsByTagName
# [01:13] <zcorpan> Returns a NodeList of all the Elements in document order... -- http://www.w3.org/TR/DOM-Level-3-Core/core.html#ID-A6C9094
# [01:15] <jgraham> (ah XPath has ordered and unordered as options)
# [01:22] * Philip` wonders how you can have an unordered list
# [01:22] <Philip`> Unordered sets make sense, but I'm not sure how an unordered list could work
# [01:23] * Quits: grimboy (n=grimboy@85.211.238.83) (Read error: 110 (Connection timed out))
# [01:23] * Philip` discovers that writing asynchronous JS code is sometimes nasty
# [01:29] <zcorpan> http://tinyurl.com/yv9atd -- both opera and firefox have managed to implement getElementsByClassName in document order... :)
# [01:29] * Joins: Lachy_ (n=Lachlan@cm-84.215.54.100.getinternet.no)
# [01:29] * jgraham will send email about gEBCN unless anyone else already did
# [01:30] * Joins: dglazkov (n=dglazkov@adsl-074-229-248-021.sip.bhm.bellsouth.net)
# [01:39] * Quits: Lachy (n=Lachlan@cm-84.215.54.100.getinternet.no) (Read error: 110 (Connection timed out))
# [01:40] * Quits: weinig (n=weinig@17.203.15.140)
# [01:42] * Parts: zcorpan (n=zcorpan@c-cb21e353.1451-1-64736c12.cust.bredbandsbolaget.se)
# [01:48] * Quits: dglazkov (n=dglazkov@adsl-074-229-248-021.sip.bhm.bellsouth.net)
# [01:52] * Quits: kingryan (n=kingryan@dsl092-219-050.sfo1.dsl.speakeasy.net)
# [01:52] * Joins: dglazkov (n=dglazkov@adsl-074-229-248-021.sip.bhm.bellsouth.net)
# [01:53] * Joins: weinig (n=weinig@17.203.15.140)
# [01:53] * Dashiva laughs at the puny job queue
# [02:00] * Philip` now has properly-shaded X3D text in Firefox, but realises it'd require far too much effort to bother doing in Opera (if it's even possible)
# [02:30] * Joins: othermaciej_ (n=mjs@dsl081-048-145.sfo1.dsl.speakeasy.net)
# [02:31] * Quits: othermaciej (n=mjs@dsl081-048-145.sfo1.dsl.speakeasy.net) (Read error: 104 (Connection reset by peer))
# [02:34] * Joins: doublec (n=Chris_Do@203-97-173-6.cable.telstraclear.net)
# [02:37] * Quits: Lachy_ (n=Lachlan@cm-84.215.54.100.getinternet.no) ("Leaving")
# [02:38] * Joins: Lachy (n=Lachlan@cm-84.215.54.100.getinternet.no)
# [02:38] * othermaciej_ is now known as othermaciej
# [02:49] * Joins: tomh- (i=tom@201pc233.sshunet.nl)
# [02:51] * Parts: tomh- (i=tom@201pc233.sshunet.nl) ("Leaving")
# [03:03] * Joins: webben (n=benh@91.84.19.165)
# [03:07] * Joins: kura (n=faruk@c-24-6-99-228.hsd1.ca.comcast.net)
# [03:13] * Quits: eseidel (n=eseidel@nat/google/x-c9f0be35d0a3e4fe)
# [03:19] * Joins: eseidel (n=eseidel@nat/google/x-a774e72c165cad60)
# [03:19] * kura is now known as kura[afk]
# [03:20] * Quits: eseidel (n=eseidel@nat/google/x-a774e72c165cad60) (Client Quit)
# [03:21] * Joins: eseidel (n=eseidel@nat/google/x-1e11e5785ce10976)
# [03:22] * Quits: eseidel (n=eseidel@nat/google/x-1e11e5785ce10976) (Client Quit)
# [03:27] * Disconnected
# [03:27] * Attempting to rejoin channel #whatwg
# [03:27] * Rejoined channel #whatwg
# [03:27] * Topic is 'WHATWG (HTML5) -- http://www.whatwg.org/ -- Logs: http://krijnhoetmer.nl/irc-logs/ -- Please leave your sense of logic at the door, thanks!'
# [03:27] * Set by gsnedders on Tue Dec 18 21:41:19
# [03:28] * Quits: krijnh (n=krijnhoe@ktk.xs4all.nl) (Read error: 104 (Connection reset by peer))
# [03:37] * Quits: tndH (i=Rob@adsl-87-102-34-81.karoo.KCOM.COM) ("ChatZilla 0.9.79-rdmsoft [XULRunner 1.8.0.9/2006120508]")
# [03:44] * MacDomeOut is now known as MacDome
# [04:39] * Quits: tantek (n=tantek@adsl-63-195-114-133.dsl.snfc21.pacbell.net)
# [04:45] * Quits: weinig (n=weinig@17.203.15.140)
# [04:49] <Hixie> k, commented on http://tech.gtaero.net/2008/01/xhtml2-vs-html5.html
# [04:49] * MacDome wonders if WebKIt failed to implement gEBCN in document order... I kinda doubt it
# [05:05] <othermaciej> MacDome: it's in document order
# [05:06] <othermaciej> we'd have to go out of our way to use any other order
# [05:06] <MacDome> othermaciej: I figured it would be, given that it's *right there* in the spec
# [05:06] * MacDome was failing to undersatnd the line of questioning beofre about gEBCN
# [05:06] <MacDome> no matter
# [05:07] <othermaciej> does the spec say document order?
# [05:09] <othermaciej> both getElementsByName and getElementsByClassName don't seem specify an ordering
# [05:11] <othermaciej> in contrast DOM 2 Core says this for getElementsByTagName: "Returns a NodeList of all the Elements with a given tag name in the order in which they are encountered in a preorder traversal of the Document tree."
# [05:11] <othermaciej> (preorder traversal is document order)
# [05:17] * Quits: dglazkov (n=dglazkov@adsl-074-229-248-021.sip.bhm.bellsouth.net)
# [05:19] * Joins: weinig (n=weinig@c-71-198-176-23.hsd1.ca.comcast.net)
# [05:27] * MacDome is now known as MacDomeOut
# [05:41] * Quits: aroben_ (n=aroben@unaffiliated/aroben) (Read error: 110 (Connection timed out))
# [05:50] * Quits: kura[afk] (n=faruk@c-24-6-99-228.hsd1.ca.comcast.net) (Read error: 104 (Connection reset by peer))
# [05:52] * Quits: weinig (n=weinig@c-71-198-176-23.hsd1.ca.comcast.net)
# [05:53] * Joins: kura[afk] (n=faruk@c-24-6-99-228.hsd1.ca.comcast.net)
# [06:06] * Joins: aroben (n=aroben@unaffiliated/aroben)
# [06:10] * Quits: digx (n=rick@c-76-109-201-140.hsd1.fl.comcast.net)
# [06:21] * Quits: roc (n=roc@202.0.36.64)
# [07:11] * Joins: aroben_ (n=aroben@unaffiliated/aroben)
# [07:28] * Quits: aroben (n=aroben@unaffiliated/aroben) (Read error: 110 (Connection timed out))
# [07:32] * Quits: aroben_ (n=aroben@unaffiliated/aroben) (Read error: 110 (Connection timed out))
# [07:43] * Quits: jruderman (n=jruderma@corp-241.mountainview.mozilla.com) (Read error: 110 (Connection timed out))
# [07:44] * Joins: maikmerten (n=merten@ls5laptop14.cs.uni-dortmund.de)
# [08:02] * Joins: jruderman (n=jruderma@c-67-180-15-227.hsd1.ca.comcast.net)
# [08:06] * Joins: aroben (n=aroben@unaffiliated/aroben)
# [08:09] * Quits: csarven (n=nevrasc@modemcable130.251-202-24.mc.videotron.ca) ("http://www.csarven.ca/")
# [08:18] * Joins: roc (n=roc@121-72-6-142.dsl.telstraclear.net)
# [08:31] * Quits: aroben (n=aroben@unaffiliated/aroben) (Read error: 110 (Connection timed out))
# [08:32] <kig> hard to do glows in svg, skipping
# [08:33] * Quits: Thezilch[FH] (n=fuz007@ip68-111-154-116.sd.sd.cox.net) (Read error: 104 (Connection reset by peer))
# [08:36] * MacDomeOut is now known as MacDome
# [08:36] * Joins: Thezilch (n=fuz007@ip68-111-154-116.sd.sd.cox.net)
# [08:44] * Quits: othermaciej (n=mjs@dsl081-048-145.sfo1.dsl.speakeasy.net) (Read error: 104 (Connection reset by peer))
# [08:45] * Joins: othermaciej (n=mjs@dsl081-048-145.sfo1.dsl.speakeasy.net)
# [08:54] * Quits: othermaciej (n=mjs@dsl081-048-145.sfo1.dsl.speakeasy.net) (Remote closed the connection)
# [09:05] * Joins: weinig (n=weinig@c-71-198-176-23.hsd1.ca.comcast.net)
# [09:12] * Joins: hdh (n=hdh@58.187.91.126)
# [09:14] <hdh> in <http://hdh.dyn-o-saur.com:81/~hdh/blos/twitter-160>, does anyone see the ETX mark appears after the first paragraph?
# [09:14] * Joins: othermaciej (n=mjs@dsl081-048-145.sfo1.dsl.speakeasy.net)
# [09:15] <hdh> konqueror4 (khtml) and qt demo browser display that way, firefox2 and opera beta put ETX before "posted at"
# [09:37] * Joins: virtuelv (n=virtuelv@pat-tdc.opera.com)
# [09:49] * Joins: digx (n=rick@c-76-109-201-140.hsd1.fl.comcast.net)
# [09:51] <hdh> http://hdh.dyn-o-saur.com:81/~hdh/misc/html5/after.html konqueror (khtml) and qtwebkit display the added content after the 1st paragraph
# [09:52] * Quits: othermaciej (n=mjs@dsl081-048-145.sfo1.dsl.speakeasy.net) (Remote closed the connection)
# [09:58] * Quits: webben (n=benh@91.84.19.165)
# [10:04] <kig> oh. firefox 2 doesn't do feGaussianBlur
# [10:11] * Quits: kura[afk] (n=faruk@c-24-6-99-228.hsd1.ca.comcast.net) (Read error: 110 (Connection timed out))
# [10:21] * Joins: ROBOd (n=robod@89.122.216.38)
# [10:28] * weinig is now known as weinig|zZz
# [10:32] * Quits: Lachy (n=Lachlan@cm-84.215.54.100.getinternet.no) ("This computer has gone to sleep")
# [10:38] * Joins: webben (n=benh@nat/yahoo/x-7a772f74caabfa67)
# [10:38] * Joins: Camaban (n=adrianle@host81-133-64-167.in-addr.btopenworld.com)
# [10:46] * Joins: Lachy (n=Lachlan@pat-tdc.opera.com)
# [10:56] * Quits: roc (n=roc@121-72-6-142.dsl.telstraclear.net)
# [11:27] <virtuelv> ugh. the double-click handler in the spec is annoying
# [11:28] <virtuelv> I can no longer triple-click to select paragraphs
# [11:44] <Lachy> virtuelv, yeah, I have the same problem with it.
# [11:44] * Quits: virtuelv (n=virtuelv@pat-tdc.opera.com) (Remote closed the connection)
# [11:45] <Lachy> Hixie, maybe the script should require the user to Shift+Double-Click or something
# [11:45] * Joins: virtuelv (n=virtuelv@pat-tdc.opera.com)
# [11:46] <hsivonen> Lachy: wouldn't that be too device-dependent?
# [11:47] <hsivonen> Hixie: Re: doctype sniffing: I think it doesn't make sense to second-guess the Mozilla quirky doctype list. having three entries of dead code is cheap relative to the data size of the list that has to be there anyway, the doctypes didn't come out of nowhere and Anne says Opera has hit one of them
# [11:49] <Hixie> lachy: k it requires ctrl+dblclick now
# [11:50] <Hixie> hsivonen: the doctypes came out of dbaron and i going through looking at what we could find at the time, iirc
# [11:50] <Hixie> but i will be looking at anne's mail in due course
# [11:52] * Joins: OmegaJunior (n=ZJr@a82-95-48-162.adsl.xs4all.nl)
# [11:56] <hsivonen> Hixie: I'm aware that dbaron collected the inital list. my point is that taking the three doctypes out of the quirky list has no real upside and has a potential downside
# [11:57] <hsivonen> I just tried googling for "David's list of doctypes" but it seems the original list isn't on the Web anymore
# [11:58] <hsivonen> archive.org doesn't have the relevant file from fas.harvard.edu, either
# [11:58] <Lachy> hsivonen, not much more device-dependent than it is already. It already requires a mouse, which not all devices have
# [11:58] <Lachy> Hixie, thanks
# [11:59] <hsivonen> Lachy: actually, it required a pointing device that can emit a double-click :-)
# [11:59] <hsivonen> but yeah, not worth tweaking
# [11:59] <Lachy> Hixie, Ctrl+DblClick is not good. Ctrl+Click triggers the context menu on Macs
# [12:00] <hsivonen> but in general, e.g. on the N800 keyboard emulation occupies the pointing device
# [12:00] <Hixie> Lachy: k, meta-dblclick
# [12:02] <Lachy> Hixie, ok, now that works on Mac. But which key is the meta key on windows?
# [12:02] <OmegaJunior> ctrl
# [12:02] <Hixie> alt, actually
# [12:02] <OmegaJunior> oh?
# [12:02] <Hixie> ctrl is, well, ctrl
# [12:02] <Lachy> Alt doesn't work in Firefox on Windows
# [12:03] <Hixie> really?
# [12:03] <Hixie> odd
# [12:03] <Hixie> maybe windows doesn't have meta
# [12:03] <Hixie> silly line of computers
# [12:04] <Hixie> k, changed it to alt
# [12:04] <Lachy> AFAIK, Windows only supports the "Alt", "Shift" and "Ctrl", and maybe "AltGraph" with some keybaords
# [12:05] <Lachy> can you make it work with either Ctrl+DblClick or Meta+DblClick
# [12:05] <OmegaJunior> Apple has option, command and shift. Option usually maps to the alt key, and command maps to the ctrl key on windows.
# [12:06] <OmegaJunior> (and the Apple key maps to the Windows key ;) )
# [12:06] <Lachy> OmegaJunior, the command key is the apple key.
# [12:07] * Joins: tndH_ (i=Rob@adsl-87-102-20-204.karoo.KCOM.COM)
# [12:07] * tndH_ is now known as tndH
# [12:07] <OmegaJunior> Hmm... it's been almost a year since I stopped using Macs... I must have forgotten a lot in the mean time.
# [12:08] <Lachy> Apple's keys don't map directly one-to-one to the windows keys, unfortunately
# [12:09] <OmegaJunior> Didn't people already work out this difference for the javascript click event model?
# [12:09] <Hixie> Lachy: i could, but is there a good reason to?
# [12:10] <Hixie> anyone got any TreeWalker or NodeIterator bugs they know about?
# [12:10] <Lachy> Hixie, it depends if you want to allow both Windows and Mac users to add annotations, you either need a single key that works for both, or allow either key to be used.
# [12:11] <Hixie> Lachy: alt works for both, no?
# [12:11] <Lachy> it should
# [12:11] <Hixie> that's what it's current set to
# [12:12] <Lachy> yes, it works
# [12:13] <Lachy> Hixie, just that, IIRC, NodeIterator and TreeWalker are totally unsupported in IE.
# [12:13] <Hixie> and firefox, it seems
# [12:13] <Lachy> I thought they were supported in FF
# [12:14] <Hixie> so did i
# [12:14] * hsivonen thought Firefox had TreeWalker
# [12:15] <hsivonen> Does TreeWalked plus NodeFilter even reduce XPCOM crossings significantly compared to pure-JS DOM traversal?
# [12:15] <Hixie> it certainly has the potential to
# [12:16] <Hixie> but personally i just find it very convenient to be able to walk the tree witha filter without having to implement the whole thing myself
# [12:17] <hsivonen> I think it is code-wise easier to have the generic DOM walk loop stored somewhere so that I can copy and paste the same loop when I need it
# [12:17] <Lachy> my test shows that TreeWalker is supproted, but NodeFilter isn't in Firefox
# [12:17] <hsivonen> too bad the pure-JS walk is too slow over the whole HTML 5 spec in Gecko
# [12:18] <hsivonen> (especially if run from chrome)
# [12:18] <Lachy> same result in Opera 9.24
# [12:18] <Hixie> opera fails the node iterator tests pretty fundamentally, yet shows presence of support
# [12:18] <Hixie> wtf
# [12:19] <Lachy> Opera isn't showing any support for NodeFilter in my test
# [12:22] <Hixie> oh it's definitely doing _something_ with nodeiterator
# [12:22] <Hixie> my filter function is being called
# [12:24] <Hixie> check out test 6 in opera
# [12:24] <Hixie> it fails in expectation 2
# [12:24] <Hixie> which means expectation 1 is being called
# [12:24] <Hixie> and that's in the node filter!
# [12:33] <hdh> meta+LMB triggers window moving on x11
# [12:41] <Lachy> oops, I just realised I wrote NodeFilter instead of NodeIterator above in my last 3 messages.
# [12:55] * Joins: zcorpan (n=zcorpan@c-cb21e353.1451-1-64736c12.cust.bredbandsbolaget.se)
# [12:55] * Quits: webben (n=benh@nat/yahoo/x-7a772f74caabfa67)
# [13:01] <Hixie> i don't understand why firefox fails test 9
# [13:02] <Hixie> 35 tests to go
# [13:02] <Hixie> bed time nownn
# [13:20] * Quits: virtuelv (n=virtuelv@pat-tdc.opera.com) ("Leaving")
# [13:46] * Joins: webben (n=benh@nat/yahoo/x-3cce5d5578d6e9d8)
# [13:55] * Quits: zcorpan (n=zcorpan@c-cb21e353.1451-1-64736c12.cust.bredbandsbolaget.se) (Read error: 110 (Connection timed out))
# [14:16] * Joins: virtuelv (n=virtuelv@pat-tdc.opera.com)
# [14:21] * Joins: zcorpan (n=zcorpan@c-cb21e353.1451-1-64736c12.cust.bredbandsbolaget.se)
# [14:54] * Quits: zcorpan (n=zcorpan@c-cb21e353.1451-1-64736c12.cust.bredbandsbolaget.se) (Read error: 110 (Connection timed out))
# [15:12] * Quits: OmegaJunior (n=ZJr@a82-95-48-162.adsl.xs4all.nl) ("Trillian (http://www.ceruleanstudios.com")
# [15:24] * Quits: virtuelv (n=virtuelv@pat-tdc.opera.com) ("Leaving")
# [15:27] * Joins: zcorpan (n=zcorpan@c-cb21e353.1451-1-64736c12.cust.bredbandsbolaget.se)
# [15:44] * Joins: aroben (n=aroben@unaffiliated/aroben)
# [15:48] * Parts: zcorpan (n=zcorpan@c-cb21e353.1451-1-64736c12.cust.bredbandsbolaget.se)
# [15:57] * Quits: maikmerten (n=merten@ls5laptop14.cs.uni-dortmund.de) (Remote closed the connection)
# [16:16] * MacDome is now known as MacDomeOut
# [16:20] * Joins: zcorpan (n=zcorpan@c-cb21e353.1451-1-64736c12.cust.bredbandsbolaget.se)
# [16:20] <zcorpan> http://software.hixie.ch/utilities/js/live-dom-viewer/?%3C!DOCTYPE%20html%3E%0D%0A%3Cstyle%3E%0D%0Abody%20%7B%20color%3Agreen%20%7D%0D%0A%5Balign%3D%C4%B0%5D%20%7B%20color%3Ared%20%7D%0D%0A%3C%2Fstyle%3E%0D%0A%3Cbody%3E%0D%0A%3Cscript%3E%0D%0Avar%20e%20%3D%20document.createElementNS('x'%2C%20'y')%3B%0D%0Ae.setAttribute('align'%2C%20'i')%3B%0D%0Ae.textContent%20%3D%20'This%20text%20should%20be%20green.'%3B%0D%0Adocument
# [16:20] <zcorpan> A%3C%2Fscript%3E
# [16:21] <zcorpan> try that in firefox
# [16:22] <zcorpan> or [align=I] in safari
# [16:22] * Joins: phsiao (n=shawn@nat/ibm/x-91954650d3f810e0)
# [16:36] <zcorpan> it seems that in firefox, a set of attributes are unicode case-insensitive for *all* elements. in safari, *all* attributes on *all* elements are ascii case-insensitive. in opera, a set of attributes are unicode case-insensitive for all *html* elements.
# [16:39] <zcorpan> http://simon.html5.org/test/selectors/case-sensitivity/
# [16:43] <Lachy> zcorpan, that dom viewer link was too long and broke. Can you just upload it to the clipboard
# [16:45] <zcorpan> Lachy: done
# [16:49] <Lachy> zcorpan, it's not clear how to interpret the result of those case-sensitivity tests.
# [16:54] * Joins: dbaron (n=dbaron@pool-72-94-185-124.phlapa.fios.verizon.net)
# [16:55] * Quits: aroben (n=aroben@unaffiliated/aroben) (Read error: 110 (Connection timed out))
# [16:55] <zcorpan> everything listed matched the selector
# [17:00] <Lachy> ok. It would be much more useful if there was some kind of explanation and/or summary
# [17:04] * Joins: aroben (n=aroben@unaffiliated/aroben)
# [17:11] <zcorpan> opera is missing valign='', and firefox is missing ismap='', but otherwise opera and firefox have the same set of attributes that are case insensitive
# [17:13] <zcorpan> combining them gives 54 different attributes... assuming i haven't missed some attribute
# [17:14] <Philip`> Is each attribute treated the same regardless of which element it's on?
# [17:14] <zcorpan> yes
# [17:16] <kig> need a smarter thumbnail cache. 2.6GB memory use for half a million images is no fun
# [17:17] <Philip`> kig: Need fewer images :-p
# [17:17] * kig plans to scale to billions :(
# [17:18] * Philip` wonders where that many images would come from
# [17:18] <kig> internet
# [17:19] <Philip`> Oh, makes sense
# [17:19] * Joins: csarven (n=nevrasc@modemcable130.251-202-24.mc.videotron.ca)
# [17:19] <kig> ( http://dark.fhtr.org/exafile/ )
# [17:20] <kig> ... all my projects tend to be somehow strangely deranged and megalomanic
# [17:21] <Philip`> Is this somewhat related to the thingy that Microsoft Research showed a while ago?
# [17:22] <Philip`> like http://www.labs.live.com/photosynth/ though they had a different demo that was just zooming in on lots of thumbnails
# [17:22] <kig> yes and no
# [17:23] <kig> similar, same ideas, nothing else to do with it
# [17:25] <kig> here video http://fhtr.blogspot.com/2007/09/and-screencap-video-of-zoomable-file.html
# [17:26] <kig> i think the photosynth demo had something like 600 photos on screen
# [17:27] * Quits: inimino (n=inimino@c-75-71-88-233.hsd1.co.comcast.net) (Read error: 104 (Connection reset by peer))
# [17:29] <Philip`> That looks quite neat :-)
# [17:30] <Philip`> How are you meant to find a particular image when it's somewhere in a pile of fifty thousand other images?
# [17:31] <Philip`> Also, can you do this with AJAX and canvas? ;-)
# [17:31] <kig> the same way you find it when it's on the file system
# [17:31] <kig> it is ajax, actually..
# [17:31] <kig> and html
# [17:31] <Philip`> Ooh
# [17:33] * Joins: inimino (n=inimino@c-75-71-88-233.hsd1.co.comcast.net)
# [17:33] <kig> but has wacky js cross-platform issues that i have never ironed out, and hence works only in firefox 2 (ouch)
# [17:36] * Quits: inimino (n=inimino@c-75-71-88-233.hsd1.co.comcast.net) (Remote closed the connection)
# [17:37] * hdh likes the scale in Goals
# [17:40] <Philip`> "Every file produced by humanity during 30 thousand years." sounds a bit suspicious - most 30,000 year spans have resulted in zero files being produced
# [17:41] <kig> extrapolating with 10 billion people each producing a thousand files a year (or something like that)
# [17:42] <hdh> <meter>?
# [17:42] * Joins: cgriego (n=cgriego@216.138.69.206)
# [17:43] <Philip`> It seems hard to accurately extrapolate the population of the earth into the next century, never mind thirty thousand years into the future :-)
# [17:43] <kig> yeah, could replace it with something else
# [17:43] <Philip`> It seems hard to accurately extrapolate the idea of "humanity" that far, too :-p
# [17:43] <hdh> lol
# [17:44] <kig> http://en.wikipedia.org/wiki/Orders_of_magnitude_%28numbers%29#1018
# [17:45] <kig> "every insect on earth"
# [17:45] <kig> "crawling"
# [17:53] * Quits: zcorpan (n=zcorpan@c-cb21e353.1451-1-64736c12.cust.bredbandsbolaget.se) (Read error: 110 (Connection timed out))
# [17:54] * Quits: hober (n=ted@unaffiliated/hober) (Remote closed the connection)
# [17:54] * Joins: zcorpan (n=zcorpan@c-cb21e353.1451-1-64736c12.cust.bredbandsbolaget.se)
# [17:55] <zcorpan> ok... so how do we want case-insensitivity for values in attribute selectors to work?
# [17:56] <zcorpan> leaking case-insensitivity to attributes on non-html elements (like firefox and safari) seems clearly wrong
# [17:57] <zcorpan> but perhaps checking the element to see if the attribute should be case-insensitive is too much of a perf hit?
# [17:58] <zcorpan> e.g. should <a accept> be case-insensitive?
# [18:02] * Philip` discovers why his gl.blendFunc(gl.ONE_MINUS_DEST_ALPHA, ...) code wasn't working as expected - it's spelled ONE_MINUS_DST_ALPHA instead
# [18:03] * Philip` would like the API to warn him when passing undefined arguments
# [18:08] * gsnedders needs an XML caching format
# [18:08] <gsnedders> (for caching processed XML data)
# [18:08] * Quits: ianloic (i=yakk@glub.dreamhostps.com) ("Terminated with extreme prejudice - dircproxy 1.2.0")
# [18:11] <Philip`> gsnedders: "Processed" in what form? Just a simple tree of 'struct node { string content; map<string, string> attributes; list<node> children; }' or something?
# [18:12] * Philip` wrote something like that to cache parsed XML for a game (in C++)
# [18:12] <gsnedders> Philip`: this element/attribute has this processed content (i.e., a string) with these options
# [18:12] <gsnedders> Philip`: I need to serialise it back in XML though
# [18:13] <Philip`> (because it took far too long to parse every time the game was loaded, but writing custom binary formats is a bit of a pain, so we just cache the parsed XML in a fast binary format)
# [18:14] <Philip`> (which is easier when it's a subset of XML that doesn't have namespaces or mixed text/element children or anything fancy)
# [18:14] <Philip`> (Age of Mythology / Age of Empires III does exactly the same, incidentally)
# [18:15] <gsnedders> Ah. I need to use XML as the cached format as it's the only format PHP can natively parse into a DOM tree (and anything in userland code is far slower)
# [18:15] * Joins: inimino (n=inimino@c-75-71-88-233.hsd1.co.comcast.net)
# [18:15] <Philip`> Ah, okay
# [18:16] <gsnedders> And because I need non-lossy storage of the XML input
# [18:16] <Lachy> Philip`, where did you quote "Every file produced by humanity during 30 thousand years." from?
# [18:17] <Philip`> Lachy: From the page kig pointed at
# [18:17] <Philip`> http://dark.fhtr.org/exafile/
# [18:20] * Lachy wonders what the screen resolution would need to be to get "A quintillion files in a 100 quadrillion directory hierarchy." on screen at once
# [18:20] <hdh> subpixel hinting
# [18:21] <Philip`> You can get sixteen million images into a single pixel, such that each image still potentially has an effect on the output image
# [18:22] <Philip`> so you'd need a couple of hundred thousand pixels width/height on your monitor, else some of the input images would never have any effect and it wouldn't really be fair to say they're being displayed
# [18:23] <hdh> :)
# [18:24] <Lachy> Philip`, I don't think the number of colours that can be represented by a pixel equals the number of files that could be represtented with that single pixel in the way you describe
# [18:24] <Philip`> Uh, that's true
# [18:25] <Philip`> I should have said 24 instead of sixteen million
# [18:25] <Lachy> why 24?
# [18:27] <Philip`> Because one pixel is 24 bits, so I was thinking that if you had more than 24 images then at least one of them couldn't possibly affect the output since there aren't enough bits, except actually I'm being stupid
# [18:27] <Philip`> because it's easy to have fractional bits of information
# [18:27] <gsnedders> silly Philip`
# [18:27] <gsnedders> I mean, I'm _never_ silly.
# [18:27] <Philip`> e.g. compute the MD5 of all the images then take the first 24 bits
# [18:27] * Joins: mpt (n=mpt@canonical/launchpad/mpt)
# [18:27] <Philip`> so every image will have an effect on the output
# [18:28] <Philip`> though that'd also be a silly way to represent multiple images
# [18:28] <Lachy> you'd have to find the average colour of every file, and then find the average of all those averages. Given an even distribution of colours, eventually the pixel would end up being a shade of grey.
# [18:28] <Lachy> or white
# [18:29] <Lachy> I don
# [18:29] <Philip`> If you're just taking the mean, then the limit would be 256
# [18:30] <Philip`> because if you have 256 images, an additional image would contribute < 1/256 to each component, so it would have no effect
# [18:30] <Lachy> no, say you have 256 predomindately red files. Then you add a bunch of predominately blue files to the collection. The average colour would still be shifted towards blue.
# [18:31] <Philip`> It wouldn't, since the output blue would be (256*0.0 + 1*1.0)/257 = 0 (rounded to the display's colour precision)
# [18:31] <Philip`> ...if you're only adding one blue image
# [18:31] <Lachy> I said a *bunch*
# [18:32] * Quits: mpt (n=mpt@canonical/launchpad/mpt) (Remote closed the connection)
# [18:32] <Philip`> If you add a bunch, you could remove one of them and have no effect on the output, so that image isn't really in the output
# [18:33] <Philip`> I'm not at all convinced by myself, though
# [18:33] <Lachy> add 256 blue files to the collection of 256 red files. The average will then be a shade of magenta
# [18:33] <Philip`> Remove one of those red files, and the output will be precisely the same as before
# [18:34] <Philip`> so all you're really displaying is 256 blue and 255 red files
# [18:34] <Lachy> there is no limit to the number of files. However, as the number of files increases, the effect of a single file reduces
# [18:35] <Philip`> There is a limit because eventually the effect of a single file reduces to zero :-)
# [18:36] <Lachy> no, the effect of a single file just asymptotically approaches zero
# [18:36] <Philip`> The output is quantised to 24 bits, so the effect does reach zero
# [18:36] <Philip`> (Er, I guess it's more correct to say it's quantised to 8 bits)
# [18:36] <Lachy> that's a rounding error that occurs after the calculation
# [18:36] * gsnedders might actually write a spec in terms of ABNF! :o
# [18:37] <Lachy> it doesn't affect the calculation itself
# [18:37] <Philip`> It's not a rounding error - it's a fundamental part of the question of how many images you can meaningfully represent in a pixel on a monitor :-)
# [18:38] * Quits: dbaron (n=dbaron@pool-72-94-185-124.phlapa.fios.verizon.net) (Read error: 110 (Connection timed out))
# [18:38] <Philip`> (...assuming it's not a silly LCD monitor that does 18-bit-plus-dithering or something)
# [18:39] <Lachy> IIRC, the human eye can only distinguish about 10 million colours anyway
# [18:39] <Philip`> Bah, humans weren't part of the question
# [18:40] <Philip`> In 30,000 years we will all be robots and we'll plug ourselves directly into the VGA socket
# [18:40] <Lachy> You think we'll still have VGA in 30,000 years? Surely we'll have upgraded to at least Dual-DVI by then :-)
# [18:41] * Joins: dbaron (n=dbaron@pool-71-246-1-92.phlapa.fios.verizon.net)
# [18:41] <Philip`> Don't underestimate the problems of backward compatibility in technology!
# [18:41] <Lachy> (hopefully HDMI will have been thrown out by then)
# [18:42] <Lachy> DVI is backwards compatible with VGA, so not a problem
# [18:42] <Lachy> just needs a simple DVI-VGA adapter
# [18:42] <gsnedders> hmmm… ABNF doesn't specify which match to use if multiple alternatives match.
# [18:42] <Philip`> Only DVI-I
# [18:43] <Lachy> what about DVI-A?
# [18:43] <Philip`> gsnedders: Probably the idea is that you don't have multiple alternatives matching, because that would be an ambiguous / non-deterministic grammar
# [18:43] <gsnedders> Lachy: Dual-DVI? Not Dual-link DVI? Why would we want two sockets? :P
# [18:43] <gsnedders> Philip`: yeah, I guess I'll just use something like, "If xxx doesn't match, yyy is a string."
# [18:44] <Philip`> Is DVI-A something other than VGA but in an incompatible shape?
# [18:44] <Lachy> yeah, I meant Dual Link DVI
# [18:44] <gsnedders> Philip`: DVI-A is VGA only
# [18:44] <gsnedders> Philip`: DVI-I is VGA and DVI-D
# [18:45] <Lachy> but 2 sockets are necessary for 2 monitors. Or, when we are robots, one for each eye
# [18:45] <gsnedders> (where DVI-D is the native digital one)
# [18:45] <Philip`> Anyway, when the robots invent time travel, they'll make us pay for inventing all these legacy technologies that they'll be shackled to for the rest of eternity
# [18:48] <Lachy> in that case, we should just deposit $1 in a long term savings account, and by the time the robots come to collect their compensation in 30,000 years, then interest earned should cover it.
# [18:48] <Dashiva> Dunno, there's inflation to consider too
# [18:49] <Philip`> There's the matter of writing down the PIN number too
# [18:50] * gsnedders once managed to get through five PIN numbers in a year.
# [18:52] <Lachy> nah, just write it in your Will that the money in the account is left to the robots. Your heirs generally don't need your PIN to access their inheritence, and nor should the robots
# [18:52] * weinig|zZz is now known as weinig
# [18:53] * Lachy has to go home. Back soon.
# [18:53] * Quits: Lachy (n=Lachlan@pat-tdc.opera.com) ("This computer has gone to sleep")
# [18:54] * Quits: zcorpan (n=zcorpan@c-cb21e353.1451-1-64736c12.cust.bredbandsbolaget.se) (Read error: 110 (Connection timed out))
# [18:55] <gsnedders> Hixie (or anyone else who knows): is the spec-gen available at all?
# [19:01] <Philip`> gsnedders: Do you mean the source code, or just remote access to the service?
# [19:01] <gsnedders> Philip`: either
# [19:01] * Parts: Camaban (n=adrianle@host81-133-64-167.in-addr.btopenworld.com)
# [19:02] * gsnedders vaguely remembers something being said about it being by Bert Bos and being remotely available to W3C members, but isn't very sure
# [19:04] * Philip` remembers the same
# [19:04] <Philip`> but all I can find is http://www.w3.org/Tools/HTML-XML-utils/ which doesn't look like the same thing
# [19:05] * Parts: hdh (n=hdh@58.187.91.126)
# [19:09] * Joins: eseidel (n=eseidel@nat/google/x-9a203b75b79f7a6c)
# [19:10] * Joins: Lachy (n=Lachlan@cm-84.215.54.100.getinternet.no)
# [19:12] * Joins: maikmerten (n=maikmert@Lae7d.l.pppool.de)
# [19:13] * Quits: Lachy (n=Lachlan@cm-84.215.54.100.getinternet.no) (Client Quit)
# [19:13] * Joins: Lachy (n=Lachlan@cm-84.215.54.100.getinternet.no)
# [19:14] <Lachy> gsnedders, the spec generator is for W3C members only. I could look up the URL for you, but wouldn't do you much good
# [19:15] <Lachy> it was originally made for the CSSWG
# [19:15] <gsnedders> Lachy: yeah, that's what I thought.
# [19:16] * Quits: webben (n=benh@nat/yahoo/x-3cce5d5578d6e9d8) (Connection timed out)
# [19:16] <Lachy> it would be nice if someone could write a similar system using html5lib
# [19:16] <Lachy> and an HTML5 serialiser
# [19:17] * gsnedders wonders if he can, until then, occasionally nag someone into putting stuff through it for him
# [19:17] <Lachy> what do you need done?
# [19:18] <gsnedders> now? nothing.
# [19:18] <gsnedders> just nice to decide what format to write a spec in before writing it :)
# [19:18] <Lachy> ok, sure. I can do nothing for you now :-)
# [19:18] <gsnedders> Lachy: hard working guy, eh? :)
# [19:19] <Lachy> I could do it for you occasionally
# [19:20] <gsnedders> I expect I could get others to do so occasionally too: most likely whoever happens to be around here when I need it done :)
# [19:23] * gsnedders tries to reverse-engineer what it does
# [19:23] <gsnedders> .no-num, .no-toc, and  are the obvious parts
# [19:24] <gsnedders> IIRC it does cross-references too
# [19:46] * Joins: kingryan (n=kingryan@dsl092-219-050.sfo1.dsl.speakeasy.net)
# [19:59] * aroben is now known as aroben|lunch
# [20:06] <Dashiva> > The 'height' property doesn't apply. The height of the content area should be based on the font, but this specification does not specify how.
# [20:07] <Dashiva> Anyone know off-hand if this is interop and if so, what is used?
# [20:18] * Joins: virtuelv (n=virtuelv@65.80-202-82.nextgentel.com)
# [20:19] * Joins: hober (n=ted@unaffiliated/hober)
# [20:32] * Quits: weinig (n=weinig@c-71-198-176-23.hsd1.ca.comcast.net)
# [20:58] * aroben|lunch is now known as aroben
# [21:00] * Joins: roc (n=roc@121-72-6-142.dsl.telstraclear.net)
# [21:02] * Quits: roc (n=roc@121-72-6-142.dsl.telstraclear.net) (Client Quit)
# [21:06] * Joins: roc (n=roc@121-72-6-142.dsl.telstraclear.net)
# [21:11] * Philip` doesn't like having to depth-sort transparent objects
# [21:29] * Quits: virtuelv (n=virtuelv@65.80-202-82.nextgentel.com) ("Leaving")
# [21:44] * Quits: digx (n=rick@c-76-109-201-140.hsd1.fl.comcast.net)
# [21:59] * Quits: ROBOd (n=robod@89.122.216.38) ("http://www.robodesign.ro")
# [22:21] * Joins: digx (n=rick@c-76-109-201-140.hsd1.fl.comcast.net)
# [22:32] * Joins: zcorpan (n=zcorpan@c-cb21e353.1451-1-64736c12.cust.bredbandsbolaget.se)
# [22:33] * Joins: othermaciej (n=mjs@dsl081-048-145.sfo1.dsl.speakeasy.net)
# [22:37] * Parts: zcorpan (n=zcorpan@c-cb21e353.1451-1-64736c12.cust.bredbandsbolaget.se)
# [22:41] * Joins: virtuelv (n=virtuelv@65.80-202-82.nextgentel.com)
# [22:47] * Quits: maikmerten (n=maikmert@Lae7d.l.pppool.de) (Remote closed the connection)
# [23:07] * Quits: virtuelv (n=virtuelv@65.80-202-82.nextgentel.com) ("Leaving")
# [23:09] * Quits: dbaron (n=dbaron@pool-71-246-1-92.phlapa.fios.verizon.net) ("8403864 bytes have been tenured, next gc will be global.")
# [23:30] * Joins: weinig (n=weinig@17.203.15.140)
# [23:32] * eseidel is now known as eseidelOut
# [23:35] * Joins: aroben_ (n=aroben@unaffiliated/aroben)
# [23:41] * Quits: weinig (n=weinig@17.203.15.140) (Read error: 104 (Connection reset by peer))
# [23:42] * Joins: weinig (n=weinig@17.203.15.140)
# [23:43] * Quits: phsiao (n=shawn@nat/ibm/x-91954650d3f810e0)
# [23:49] * Joins: weinig_ (n=weinig@17.203.15.140)
# [23:49] * Quits: weinig (n=weinig@17.203.15.140) (Read error: 104 (Connection reset by peer))
# [23:51] * Quits: aroben (n=aroben@unaffiliated/aroben) (Connection timed out)
# Session Close: Sat Jan 05 00:00:00 2008

The end :)