/irc-logs / w3c / #html-wg / 2007-10-13 / end

Options:

  1. # Session Start: Sat Oct 13 00:00:00 2007
  2. # Session Ident: #html-wg
  3. # [00:19] * jgraham was rather hoping to meet zcorpan_
  4. # [00:20] <shepazu> zcorpan_, why not hide in Lars Erik's luggage?
  5. # [00:20] <anne> dude
  6. # [00:20] <anne> W3C versus Thailand
  7. # [00:20] <anne> that's no match
  8. # [00:20] <jgraham> Fair point
  9. # [00:20] <shepazu> zcorpan_ is going to Thailand?
  10. # [00:20] <zcorpan_> yup
  11. # [00:20] <shepazu> dang.
  12. # [00:21] <jgraham> Maybe we should all go to Thailand then?
  13. # [00:21] <zcorpan_> heh
  14. # [00:22] * anne would prefer that
  15. # [00:22] * anne also wants this meeting on a boat chaals was going to arrange for the Web API WG
  16. # [00:25] * shepazu has looked into some of the logistics... preparing offline resources, etc.
  17. # [00:29] <anne> few more years and maybe this offline web app stuff will be implemented
  18. # [00:32] <anne> shepazu, any idea why #tp is password protected?
  19. # [00:33] <shepazu> paranoia?
  20. # [00:33] * Parts: hober (ted@68.107.112.172) (ERC Version 5.3 (devel) (IRC client for Emacs))
  21. # [00:33] <anne> I suppose http://www.w3.org/QA/2007/10/w3c-meets-web-community.html should be updated then
  22. # [00:34] <shepazu> maybe it's just not open yet
  23. # [00:34] <shepazu> write the sysadmins
  24. # [00:34] * Quits: Sander (svl@86.87.68.167) (Quit: And back he spurred like a madman, shrieking a curse to the sky.)
  25. # [00:35] <Hixie> there's a session on html5 and xhtml2?
  26. # [00:35] <Hixie> that's news to me
  27. # [00:35] <anne> it also became public who's on it
  28. # [00:35] <Hixie> oh?
  29. # [00:36] * anne looks up the link in his twitter account
  30. # [00:36] <anne> http://www.w3.org/2007/11/07-TechPlenAgenda.html
  31. # [00:36] <Hixie> that's funny
  32. # [00:36] * anne wasn't aware about the [slides] bit
  33. # [00:36] <Hixie> i'm on two panels but the one panel about my day job, i'm not on
  34. # [00:37] <anne> "URI-Based Extensibility" fun!
  35. # [00:37] * anne looks forward to that
  36. # [00:38] <Hixie> haha, i'm also not on the panel about video, despite being the guy who specced <video> :-P
  37. # [00:39] <anne> what other panel are you on?
  38. # [00:40] <anne> I think the video panel is about video formats mostly, as I understand it
  39. # [00:40] <anne> openness maybe?
  40. # [00:40] <Hixie> openness yes
  41. # [00:40] <shepazu> and it was specced (and implemented) in SVG 1.2 and in SMIL before that
  42. # [00:40] <anne> makes sense
  43. # [00:40] <Hixie> i love the first sentence of session 2's description
  44. # [00:42] <zcorpan_> lol
  45. # [00:42] <mjs> "sentence"
  46. # [00:46] <Hixie> wow, the whatwg work has really affected the w3c. sessions 2, 3, 6, 4, and 8 are all directly or closely related to the html5 work.
  47. # [00:46] <Hixie> wait the numbers are weird
  48. # [00:47] <anne> goes wrong after lunch
  49. # [00:47] <Hixie> anyway the only talks that _don't_ come almost directly out of the whatwg or microformats work are the welcome, the lightning talks, and the discussions with the director.
  50. # [00:48] <Hixie> literally every other talk is related to whatwg or microformats work.
  51. # [00:48] <Hixie> jesus.
  52. # [00:48] <shepazu> I wouldn't say that video does
  53. # [00:49] <shepazu> that's more influenced by Adobe and Silverlight
  54. # [00:49] <shepazu> specifically, YouTube
  55. # [00:50] <Hixie> i guarentee that howcome's talk in that panel will include a demo of HTML5's <video>.
  56. # [00:50] * Quits: zcorpan_ (zcorpan@83.227.34.9) (Connection reset by peer)
  57. # [00:50] <mjs> regadless of the direction of influence, the agenda is well aligned with the goals of whatwg
  58. # [00:50] <Hixie> right
  59. # [00:50] * kingryan notes that the agenda still lists tantek as working at technorati
  60. # [00:50] <kingryan> (and spells his name wrong)
  61. # [00:51] <anne> ASCII without entities kingryan
  62. # [00:51] <anne> it's the W3C after all
  63. # [00:51] * anne hides
  64. # [00:51] <kingryan> it's not the IETF
  65. # [00:51] <anne> that'd be worse
  66. # [00:51] <anne> you'd get the same thing in plain text :)
  67. # [00:51] <kingryan> AFAIK UTF-8 works pretty well
  68. # [00:52] <anne> yeah, HTML does too
  69. # [00:52] * kingryan notices that the page *is* UTF-* ( <meta http-equiv="content-type" content="text/html; charset=UTF-8" />)
  70. # [00:52] <kingryan> I think people just don't know how to type Ç
  71. # [00:53] <kingryan> anytime, time to go. see ya guys later
  72. # [00:53] * Quits: kingryan (rking3@208.66.64.47) (Quit: kingryan)
  73. # [00:54] <anne> and I was wrong
  74. # [00:54] <anne> as it does say Håkon
  75. # [00:56] * Joins: zcorpan_ (zcorpan@83.227.34.9)
  76. # [00:57] <zcorpan_> shepazu: seems to me that session 8 is more or less the same as http://video.google.com/videoplay?docid=5545573096553082541&q=video-element&total=55&start=0&num=10&so=0&type=search&plindex=0
  77. # [00:58] <zcorpan_> or at least Håkon's slides, i'd expect :)
  78. # [00:58] * Quits: gavin_ (gavin@99.226.81.242) (Ping timeout)
  79. # [01:04] * Joins: gavin_ (gavin@99.226.81.242)
  80. # [01:09] <shepazu> zcorpan_, probably so... but to claim that as WHATWG's influence, rather than as a trend of the Web, is silly
  81. # [01:10] <shepazu> nobody is using the <html:video> tag in production code... though mobile SVG UAs are ;)
  82. # [01:10] <shepazu> that is, <svg:video>
  83. # [01:11] <shepazu> (which is taken from SMIL)
  84. # [01:12] * Joins: sbuluf (dysujgb@200.49.140.252)
  85. # [01:13] <mjs> Ç is easy to type on a Mac
  86. # [01:13] <mjs> not any harder than å
  87. # [01:13] * shepazu is eager for the next release of OSX... gonna buy a MacBook :)
  88. # [01:15] <jgraham> <AOL>me too</AOL> (probably)
  89. # [01:16] <mjs> all the cool kids are using 'em
  90. # [01:17] <anne> not all of them :p
  91. # [01:19] <zcorpan_> shepazu: mediawiki is using <html:video>
  92. # [01:19] <zcorpan_> or is it called wikimedia
  93. # [01:20] <anne> wikipedia
  94. # [01:21] <anne> although I guess that's not the software, but they're using it
  95. # [01:21] <zcorpan_> it seems wikipedia is part of wikimedia
  96. # [01:22] <zcorpan_> aha, i thought shepazu was talking about content
  97. # [01:27] <Philip> Wikimedia's Wikipedia uses MediaWiki plus an extension for handling Ogg
  98. # [01:27] <Philip> (so MediaWiki itself doesn't support video)
  99. # [01:41] * anne -> bed
  100. # [01:48] * Joins: polin8 (polin8@75.71.72.175)
  101. # [01:52] * Quits: zcorpan_ (zcorpan@83.227.34.9) (Ping timeout)
  102. # [02:20] * Quits: tH (Rob@87.102.2.12) (Quit: ChatZilla 0.9.78.1-rdmsoft [XULRunner 1.8.0.9/2006120508])
  103. # [03:00] * Quits: heycam (cam@203.214.114.92) (Quit: bye)
  104. # [03:07] * Quits: gavin_ (gavin@99.226.81.242) (Ping timeout)
  105. # [03:12] * Joins: gavin_ (gavin@99.226.81.242)
  106. # [03:39] * Quits: dbaron (dbaron@63.245.220.241) (Quit: 8403864 bytes have been tenured, next gc will be global.)
  107. # [04:28] * Joins: myakura (myakura@122.29.102.45)
  108. # [04:35] * Quits: mjs (mjs@17.255.103.143) (Quit: mjs)
  109. # [05:14] * Quits: gavin_ (gavin@99.226.81.242) (Ping timeout)
  110. # [05:19] * Joins: gavin_ (gavin@99.226.81.242)
  111. # [05:33] * Joins: mjs (mjs@64.81.48.145)
  112. # [06:40] * Joins: heycam (cam@130.194.72.84)
  113. # [06:58] * Quits: heycam (cam@130.194.72.84) (Quit: bye)
  114. # [07:21] * Quits: gavin_ (gavin@99.226.81.242) (Ping timeout)
  115. # [07:26] * Joins: gavin_ (gavin@99.226.81.242)
  116. # [08:50] * Quits: myakura (myakura@122.29.102.45) (Quit: Leaving...)
  117. # [09:24] * Quits: sbuluf (dysujgb@200.49.140.252) (Ping timeout)
  118. # [09:28] * Quits: gavin_ (gavin@99.226.81.242) (Ping timeout)
  119. # [09:33] * Joins: gavin_ (gavin@99.226.81.242)
  120. # [09:37] * Quits: mjs (mjs@64.81.48.145) (Connection reset by peer)
  121. # [09:37] * Joins: mjs_ (mjs@64.81.48.145)
  122. # [10:12] * Joins: myakura (myakura@122.29.102.45)
  123. # [10:22] * Joins: ROBOd (robod@89.122.216.38)
  124. # [11:35] * Quits: gavin_ (gavin@99.226.81.242) (Ping timeout)
  125. # [11:40] * Joins: gavin_ (gavin@99.226.81.242)
  126. # [13:05] * Quits: gsnedders (gsnedders@86.145.188.131) (Quit: gsnedders)
  127. # [13:32] * Joins: tH_ (Rob@87.102.2.12)
  128. # [13:32] * tH_ is now known as tH
  129. # [13:42] * Quits: gavin_ (gavin@99.226.81.242) (Ping timeout)
  130. # [13:47] * Joins: gavin_ (gavin@99.226.81.242)
  131. # [14:03] * Joins: zcorpan_ (zcorpan@83.227.34.9)
  132. # [14:16] * Joins: gsnedders (gsnedders@86.145.188.131)
  133. # [14:42] * Quits: tH (Rob@87.102.2.12) (Ping timeout)
  134. # [14:45] * Joins: tH (Rob@87.102.2.12)
  135. # [14:51] * Quits: polin8 (polin8@75.71.72.175) (Quit: polin8)
  136. # [14:52] * Quits: zcorpan_ (zcorpan@83.227.34.9) (Ping timeout)
  137. # [15:33] * Joins: hasather (hasather@90.227.221.48)
  138. # [15:48] * Quits: gavin_ (gavin@99.226.81.242) (Ping timeout)
  139. # [15:53] * Joins: gavin_ (gavin@99.226.81.242)
  140. # [15:58] * Joins: Sander (svl@86.87.68.167)
  141. # [16:00] * Joins: tH_ (Rob@87.102.47.69)
  142. # [16:01] * Quits: tH (Rob@87.102.2.12) (Ping timeout)
  143. # [16:01] * tH_ is now known as tH
  144. # [16:13] * Quits: anne (annevk@81.68.67.12) (Client exited)
  145. # [16:17] * Joins: anne (annevk@81.68.67.12)
  146. # [16:36] * Joins: zcorpan_ (zcorpan@83.227.34.9)
  147. # [16:52] * Quits: zcorpan_ (zcorpan@83.227.34.9) (Ping timeout)
  148. # [17:01] * Joins: zcorpan_ (zcorpan@83.227.34.9)
  149. # [17:01] <zcorpan_> hsivonen: ping
  150. # [17:03] <hsivonen> zcorpan_: unable to properly pong right now, gotta run. sorry.
  151. # [17:04] <zcorpan_> k
  152. # [17:05] <zcorpan_> when you come back: i was wondering what your ideas are more specifically with cdata sections and the content model flag in svg scope
  153. # [17:08] <zcorpan_> on one hand, it's nice to have the same content model flag changes as legacy UAs, but that creates a problem with <textArea>. otoh, it's nice to have the same content model as in svg-in-xml, but that means we probably have to introduce cdata sections
  154. # [17:09] * Joins: marcos__ (chatzilla@131.181.148.226)
  155. # [17:10] <zcorpan_> authors are used to doing "<script>//<![CDATA[" when working with both html and xhtml
  156. # [17:10] * Quits: marcos (chatzilla@131.181.148.226) (Connection reset by peer)
  157. # [17:10] * marcos__ is now known as marcos
  158. # [17:16] <zcorpan_> also, using /> syntax on elements that are (r)cdata in legacy UAs will mess up in legacy UAs
  159. # [17:17] <anne> I don't think keeping the syntax indentical is realistic
  160. # [17:17] <zcorpan_> to the same extent as between html and xhtml is realistic
  161. # [17:18] <anne> that isn't either
  162. # [17:19] <zcorpan_> um, now i don't follow
  163. # [17:20] <hsivonen> zcorpan_: to be honest, I didn't think CDATA sections and legacy behavior of <script> and <style> through properly
  164. # [17:21] <anne> well, HTML has (R)CDATA elements, XHTML has namespace prefixes, etc.
  165. # [17:21] <hsivonen> zcorpan_: it would be good to raise the issues you see on list
  166. # [17:21] <zcorpan_> hsivonen: ok
  167. # [17:21] <anne> I mean, XHTML has namespaces
  168. # [17:22] <zcorpan_> anne: ah. right. i mean that it's possible to have a document that parses into roughly the same thing for both cases
  169. # [17:23] <zcorpan_> and, to the same extent as that is possible with html and xhtml, it is realistic to design the svg-in-text/html such that it would be possible with svg in html as well
  170. # [17:23] <zcorpan_> it seems possible with hsivonen's proposal, e.g.
  171. # [17:25] <anne> I think I'd do it in a way that you switch to something called the "SVG phase" in the parser
  172. # [17:25] <anne> and you go back to the "in body phase" (currently called body insertion mode iirc) the moment you hit <foreignobject>
  173. # [17:25] <anne> you switch to "SVG phase" when you hit <svg>
  174. # [17:26] <zcorpan_> yep, makes sense
  175. # [17:26] <anne> that would allow for <svg:script> to have different parsing rules even, if desirable
  176. # [17:26] <hsivonen> anne: yeah.
  177. # [17:27] <zcorpan_> but is it?
  178. # [17:27] <anne> no
  179. # [17:27] <hsivonen> anne: I thought zcorpan_ was asking about <script> doing weird stuff in legacy browsers
  180. # [17:27] <anne> but it is for <textarea> within <svg>
  181. # [17:27] <anne> apparently
  182. # [17:27] <hsivonen> does SVG have its own textarea?
  183. # [17:27] <zcorpan_> textArea, yes
  184. # [17:28] <zcorpan_> it can contain elements
  185. # [17:28] <hsivonen> I didn't remember that
  186. # [17:28] <anne> SVG 1.2
  187. # [17:28] <hsivonen> ah, never had even made it on my radar to be forgotten, then
  188. # [17:29] <zcorpan_> it's a shape that contains text, aiui
  189. # [17:29] <anne> oh right, they have something else for editing
  190. # [17:30] * anne grmbls
  191. # [17:31] <hsivonen> (afk again)
  192. # [17:47] * Quits: myakura (myakura@122.29.102.45) (Quit: Leaving...)
  193. # [17:51] * Quits: zcorpan_ (zcorpan@83.227.34.9) (Ping timeout)
  194. # [17:56] * Quits: gavin_ (gavin@99.226.81.242) (Ping timeout)
  195. # [18:01] * Joins: gavin_ (gavin@99.226.81.242)
  196. # [18:18] * Quits: Lionheart (robin@66.57.69.65) (Ping timeout)
  197. # [18:26] * Joins: zcorpan_ (zcorpan@83.227.34.9)
  198. # [18:31] <Philip> Could <svg>...<svg:textArea>...</svg> be supported, for people who want to write SVG that works the same when deserialised as XML, HTML5 and legacy-HTML-with-scripted-fixups?
  199. # [18:41] <zcorpan_> Philip: perhaps, but then you would need double namespace declarations, and it seems very hacky
  200. # [18:43] <zcorpan_> Philip: then it seems cleaner to have prefixes for all svg elements
  201. # [18:43] <zcorpan_> (and pcdata content model for all of those)
  202. # [18:51] * Quits: zcorpan_ (zcorpan@83.227.34.9) (Ping timeout)
  203. # [18:56] * Joins: sbuluf (ompq@200.49.140.250)
  204. # [19:03] * Joins: zcorpan_ (zcorpan@83.227.34.9)
  205. # [19:18] * Quits: Sander (svl@86.87.68.167) (Quit: And back he spurred like a madman, shrieking a curse to the sky.)
  206. # [19:39] * anne might hack html5lib to add support for SVG
  207. # [19:39] <anne> although that's prolly more trouble than it's worth as we don't support namespaces or anything currently
  208. # [19:44] <anne> that, and case fixup which prolly requires a list of elements that are not all lowercase or something
  209. # [19:51] * Quits: zcorpan_ (zcorpan@83.227.34.9) (Ping timeout)
  210. # [20:02] * Quits: gavin_ (gavin@99.226.81.242) (Ping timeout)
  211. # [20:07] * Joins: gavin_ (gavin@99.226.81.242)
  212. # [20:09] * Joins: zcorpan_ (zcorpan@83.227.34.9)
  213. # [20:09] <zcorpan_> anne: that would not be forwards compatible, though
  214. # [20:10] <anne> depends on the new stuff
  215. # [20:11] <anne> doesn't really matter either if the feature isn't support anyway
  216. # [20:15] <zcorpan_> for html5lib it matters because you don't know what the app supports
  217. # [20:15] <anne> example of why it would matter?
  218. # [20:19] <zcorpan_> you'd have to change html5lib every time a new camelCase element is introduced, and everyone would have to upgrade
  219. # [20:19] <anne> yeah, that's fairly trivial
  220. # [20:19] <anne> especially given that the deployment of a new element typically takes a few years
  221. # [20:20] <zcorpan_> by the same reasoning you could have a list of void svg elements, btw
  222. # [20:21] <anne> yeah
  223. # [20:21] <zcorpan_> and treat /> the same as >
  224. # [20:21] <anne> yeah, maybe
  225. # [20:21] <anne> although in case of SVG /> can save quite a few characters
  226. # [20:24] <zcorpan_> i mean that you'd treat <circle> etc. as void elements in the parser
  227. # [20:24] <anne> that wouldn't work for SVG animation
  228. # [20:25] <zcorpan_> right
  229. # [20:28] <anne> I believe html5lib actually lowercases at the parser level currently
  230. # [20:31] <Philip> At least attribute names usually need to be lowercased in the tokeniser, to handle <a b=1 B=2> correctly
  231. # [20:36] <anne> seems that someone hacked a flag into our tokenizer for that
  232. # [20:37] <anne> lowercaseAttrName
  233. # [20:37] <anne> there's also lowercaseElementName
  234. # [20:37] * Joins: billmason (billmason@69.30.57.156)
  235. # [20:38] <anne> so I guess that I said before was based on something that's now obsoleted
  236. # [20:39] * Quits: billmason (billmason@69.30.57.156) (Quit: .)
  237. # [20:51] * Quits: zcorpan_ (zcorpan@83.227.34.9) (Quit: zcorpan_)
  238. # [20:51] * Joins: zcorpan_ (zcorpan@83.227.34.9)
  239. # [21:24] <zcorpan_> hsivonen: now that validator.nu uses a "proper" html5 parser, can't the full xhtml5 schema be used for checking html5 documents? what about the xhtml 1.0 frameset schema? (not that the latter is useful, just curius)
  240. # [21:24] <zcorpan_> s/curius/curious/
  241. # [21:39] * Quits: zcorpan_ (zcorpan@83.227.34.9) (Ping timeout)
  242. # [21:39] * Joins: zcorpan_ (zcorpan@83.227.34.9)
  243. # [21:39] * zcorpan_ is now known as zcorpan
  244. # [22:10] * Quits: gavin_ (gavin@99.226.81.242) (Ping timeout)
  245. # [22:15] * Joins: gavin_ (gavin@99.226.81.242)
  246. # [22:20] * Quits: ROBOd (robod@89.122.216.38) (Quit: http://www.robodesign.ro )
  247. # [22:38] * Quits: zcorpan (zcorpan@83.227.34.9) (Ping timeout)
  248. # [22:55] * Joins: zcorpan (zcorpan@83.227.34.9)
  249. # [23:20] <jgraham> IIRC the current state of html5lib is that we lowercase at the parser level because San wanted liberal XML parsing that doesn't case fold. However the tokenizer needs to know whether lowercasing is occurring to get the attributes right in all cases
  250. # [23:23] <jgraham> I would generally prefer SVG in HTML to special-case as few names as possible; specifically I would rather have an XML-mode flag that changed the behaviour of the tokenizer wrt case folding + etc. and a parser mode that dealt with empty elements, CDATA, etc.
  251. # [23:23] <anne> tokenizer.py doesn't look like that
  252. # [23:23] <jgraham> s/San/Sam/
  253. # [23:24] <jgraham> Ah, I think we had to move it to get an error in the right place to pass a test
  254. # [23:24] <jgraham> That wasn't the greatest reason in the world...
  255. # [23:31] <Philip> It wasn't just for the error (at least for the attribute case case) - it was for tests like "<a a A>" where HTML5 says there should be a token with one attribute, and so the tests should test that behaviour (instead of testing html5lib's different behaviour, and conflicting with any other implementation that does follow HTML5)
  256. # [23:38] * Quits: zcorpan (zcorpan@83.227.34.9) (Ping timeout)
  257. # Session Close: Sun Oct 14 00:00:00 2007

The end :)