/irc-logs / w3c / #html-wg / 2007-07-05 / end

Options:

  1. # Session Start: Thu Jul 05 00:00:00 2007
  2. # Session Ident: #html-wg
  3. # [00:10] * Quits: tH (Rob@83.100.252.160) (Quit: ChatZilla 0.9.78.1-rdmsoft [XULRunner 1.8.0.9/2006120508])
  4. # [00:14] * Quits: myakura (myakura@58.88.37.26) (Quit: Leaving...)
  5. # [00:25] * Joins: Lionhear1 (robin@66.57.69.65)
  6. # [00:26] * Quits: Lionheart (robin@66.57.69.65) (Ping timeout)
  7. # [00:30] * Joins: MikeSmith (MikeSmith@mcclure.w3.org)
  8. # [01:10] * Quits: gavin (gavin@74.103.208.221) (Ping timeout)
  9. # [01:11] * Joins: sbuluf (rxr@200.49.140.129)
  10. # [01:15] * Joins: gavin (gavin@74.103.208.221)
  11. # [02:02] * Joins: karl (karlcow@128.30.52.30)
  12. # [02:40] <karl> oooh I love days-off in USA. People don't send emails :)
  13. # [02:57] <MikeSmith> yeah
  14. # [02:57] <MikeSmith> thank you American founding fathers
  15. # [03:01] <karl> http://leftlogic.com/lounge/articles/entity-lookup/
  16. # [03:01] * Quits: MikeSmith (MikeSmith@mcclure.w3.org) (Client exited)
  17. # [03:01] <karl> HTML Entity Character Lookup
  18. # [03:01] <karl> Neat app
  19. # [03:02] * Joins: MikeSmith (MikeSmith@mcclure.w3.org)
  20. # [03:06] <karl> Example of Web site using the profile attribute http://2007.xtech.org/public/schedule/grid
  21. # [03:06] * Quits: Lachy (chatzilla@124.168.24.114) (Quit: ChatZilla 0.9.78.1 [Firefox 2.0.0.4/2007051502])
  22. # [03:06] * Joins: Lachy (chatzilla@124.168.24.114)
  23. # [03:18] * Quits: gavin (gavin@74.103.208.221) (Ping timeout)
  24. # [03:23] * Joins: gavin (gavin@74.103.208.221)
  25. # [03:25] * Quits: mjs (mjs@64.81.48.145) (Quit: mjs)
  26. # [03:34] * Joins: mjs (mjs@64.81.48.145)
  27. # [03:40] * Joins: olivier (ot@128.30.52.30)
  28. # [03:49] <karl> http://discobits.org/
  29. # [03:50] <karl> "A course grained but extensible ontology for modeling the kind of information we deal with on the web, in emails and filesystems. The aim is to model the data of the different systems in a way they share an ontology describing the generic concepts."
  30. # [04:01] * Quits: MikeSmith (MikeSmith@mcclure.w3.org) (Ping timeout)
  31. # [04:05] <Hixie> that http://2007.xtech.org/public/schedule/grid page is funny because it uses the wrong profile values for those microformats and misuses the hcalendar classes
  32. # [04:06] <karl> funny = more fun. cool!
  33. # [04:18] * Joins: MikeSmith (MikeSmith@mcclure.w3.org)
  34. # [05:07] * Quits: MikeSmith (MikeSmith@mcclure.w3.org) (Ping timeout)
  35. # [05:25] * Quits: gavin (gavin@74.103.208.221) (Ping timeout)
  36. # [05:30] * Joins: gavin (gavin@74.103.208.221)
  37. # [05:38] * Joins: MikeSmith (MikeSmith@mcclure.w3.org)
  38. # [05:46] * Quits: MikeSmith (MikeSmith@mcclure.w3.org) (Ping timeout)
  39. # [05:52] * Joins: MikeSmith (MikeSmith@mcclure.w3.org)
  40. # [07:32] * Quits: gavin (gavin@74.103.208.221) (Ping timeout)
  41. # [07:37] * Joins: gavin (gavin@74.103.208.221)
  42. # [09:06] <mjs> Dmitry Turin never ceases to entertain
  43. # [09:09] <anne> he posted something
  44. # [09:09] <anne> ?
  45. # [09:09] * anne missed it
  46. # [09:10] <anne> oh, on "Thursday, 28 June"
  47. # [09:10] <anne> why does he post in the past?!
  48. # [09:10] <anne> check http://lists.w3.org/Archives/Public/public-html/2007Jul/
  49. # [09:16] * Joins: Charl (charlvn@196.209.104.228)
  50. # [09:19] * Joins: tH (Rob@83.100.252.160)
  51. # [09:33] * Quits: karl (karlcow@128.30.52.30) (Quit: Where dwelt Ymir, or wherein did he find sustenance?)
  52. # [09:33] * Quits: olivier (ot@128.30.52.30) (Quit: This computer has gone to sleep)
  53. # [09:38] * Quits: gavin (gavin@74.103.208.221) (Ping timeout)
  54. # [09:43] * Joins: gavin (gavin@74.103.208.221)
  55. # [10:26] <Lachy> Dmitry's emails are contradictory. First he suggests to replace <object> and <link>, and presumably <img>, etc. too, unifying them as <link>.
  56. # [10:26] <Lachy> Then he proposes to replace the attributes href, action, etc. with src, where it's meaning is determined by the element. That won't work if all such elements are replaced with link.
  57. # [10:26] <Lachy> So he basically wants <link src=""> for all linking and embedding.
  58. # [10:37] <MikeSmith> I'm not yet convinced that Dmitri is real. I think he might be some kind of avant-garde surrealist performance artist
  59. # [10:40] <Lachy> have you seen the proposals he has on his website?
  60. # [10:41] <Lachy> they're a lot more nonsensical than the stuff he has posted to the list.
  61. # [10:42] <MikeSmith> I've not done more than glanced at his site ... if it outdoes the stuff he posts, that's really saying something
  62. # [10:45] <MikeSmith> Lachy, you have any plans to attend Webdirections? David Storey from Opera will be there ... would be cool if you could meet him
  63. # [10:45] <Lachy> I'll probably meet him when I start working at Opera next month
  64. # [10:45] <MikeSmith> wow
  65. # [10:46] <MikeSmith> didn't know you'd been hired tehre
  66. # [10:46] <MikeSmith> there
  67. # [10:46] <Lachy> I am hoping to go to Web Directions, if I am able to get back from Oslo for it
  68. # [10:46] <MikeSmith> what group will you be working with?
  69. # [10:46] <Lachy> the QA dept.
  70. # [10:46] <MikeSmith> reporting to Snorre, or somebody else?
  71. # [10:46] <Lachy> I assumed everyone knew by now. I mentioned in IRC a few days ago and blogged it
  72. # [10:47] <Lachy> I think Chaals said it was Snorre
  73. # [10:47] <MikeSmith> I've kind of been preoccupied with things here in Japan for the last few weeks
  74. # [10:47] <MikeSmith> ... way behind on blog reading
  75. # [10:48] <MikeSmith> if you can find a good reason to visit the Linköping office while you are in Europe, you might want to try to do that
  76. # [10:49] <MikeSmith> there is a great product development team there
  77. # [10:49] <MikeSmith> developers and QA engineers
  78. # [10:49] <Lachy> is that where zcorpan is working at the moment?
  79. # [10:49] <MikeSmith> I think it is
  80. # [10:55] <hsivonen> I understood that zcorpan is working from home even though the Linköping office is the closest office
  81. # [11:20] * Joins: zcorpan (zcorpan@84.216.41.39)
  82. # [11:57] * Quits: MikeSmith (MikeSmith@mcclure.w3.org) (Ping timeout)
  83. # [12:04] * Joins: ROBOd (robod@86.34.246.154)
  84. # [12:49] * Quits: sbuluf (rxr@200.49.140.129) (Quit: sbuluf)
  85. # [12:54] * Joins: myakura (myakura@58.88.37.26)
  86. # [13:13] * Joins: edas (edaspet@88.191.34.123)
  87. # [13:22] * Joins: MikeSmith (MikeSmith@mcclure.w3.org)
  88. # [13:33] * Quits: gavin (gavin@74.103.208.221) (Ping timeout)
  89. # [13:38] * Joins: gavin (gavin@74.103.208.221)
  90. # [13:55] * Joins: olivier (ot@128.30.52.30)
  91. # [14:01] * Joins: karl (karlcow@128.30.52.30)
  92. # [14:02] * Quits: MikeSmith (MikeSmith@mcclure.w3.org) (Ping timeout)
  93. # [14:15] * Joins: MikeSmith (MikeSmith@mcclure.w3.org)
  94. # [14:49] * Quits: Lachy (chatzilla@124.168.24.114) (Ping timeout)
  95. # [14:59] * Quits: karl (karlcow@128.30.52.30) (Quit: Where dwelt Ymir, or wherein did he find sustenance?)
  96. # [15:03] * Quits: olivier (ot@128.30.52.30) (Quit: Leaving)
  97. # [15:21] * Joins: Lachy (chatzilla@124.168.24.114)
  98. # [15:40] * Quits: gavin (gavin@74.103.208.221) (Ping timeout)
  99. # [15:45] * Joins: gavin (gavin@74.103.208.221)
  100. # [16:28] * Quits: tH (Rob@83.100.252.160) (Ping timeout)
  101. # [16:31] * Joins: billmason (billmason@69.30.57.156)
  102. # [16:33] * Joins: tH (Rob@83.100.252.160)
  103. # [16:33] * Quits: myakura (myakura@58.88.37.26) (Quit: Leaving...)
  104. # [16:40] * Quits: zcorpan (zcorpan@84.216.41.39) (Ping timeout)
  105. # [17:22] * anne hopes DanC will chime in at some point
  106. # [17:22] * DanC waves, booting up a little slowly this morning
  107. # [17:22] <DanC> chime in on what?
  108. # [17:24] <anne> actually, I'm not sure it would help
  109. # [17:24] <anne> the long thread about still images...
  110. # [17:26] <DanC> I don't remember asking anybody to talk about still images. is there some particular reason I should encourage them to by responding?
  111. # [17:27] <gavin_> I don't really know what to make of Robert Burns latest messages in that thread
  112. # [17:27] <gavin_> it doesn't seem like further discussion will be useful, at this point
  113. # [17:27] <anne> "No, I don't"
  114. # [17:28] <anne> yeah; I'm going on holiday Saturday!
  115. # [17:28] <anne> you guys figure it out
  116. # [17:29] <MikeSmith> that seems like a thread that should maybe be taken up elsewhere if it all
  117. # [17:29] <MikeSmith> e.g., "Robert, let's move this over to www-html"
  118. # [17:29] <anne> dunno, I tried on www-archive
  119. # [17:29] <anne> wasn't productive either
  120. # [17:29] <anne> maybe it's me
  121. # [17:30] <MikeSmith> I suspect that the recent increase of traffic on public-html may be (re)scaring some people away from the list
  122. # [17:30] <anne> three people who've been involved for 3 years in HTML development have tried to explain it to him and he calls us irrational
  123. # [17:30] <anne> I'm not sure what to do
  124. # [17:30] <MikeSmith> people who might have something valuable to contribute
  125. # [17:32] <anne> I wonder why it gets so out of hand here where on the WHATWG list there's almost no flamewars; just people being productive (well, most of the time anyway)
  126. # [17:33] <MikeSmith> because maybe the WHATWG list has mostly been a self-selected group with a fairly high level of technical insight in the actual issues
  127. # [17:33] <MikeSmith> scope of WHATWG list discussion is well-bounded
  128. # [17:34] <anne> there's over 700 people on the WHATWG list
  129. # [17:34] <MikeSmith> yeah yeah
  130. # [17:34] <anne> I agree that it has a much clearer scope
  131. # [17:34] <anne> (also, we've been encouraging people all over the place to subscribe)
  132. # [17:34] <anne> (maybe we can do better, dunno)
  133. # [17:34] <MikeSmith> how many of the 700 people on the WHATWG list have ever actually posted to it?
  134. # [17:34] <gavin_> despite those encouragements, I think the W3 lists are still "better known"
  135. # [17:34] <anne> I would actually have expected the W3C to attract 1000s of people
  136. # [17:34] <gavin_> and a lot of people see them as "autoritative"
  137. # [17:35] <MikeSmith> anyway, scope of public-html discussion is so far not particularly well-bounded
  138. # [17:35] * Quits: jgraham (jgraham@81.86.222.233) (Ping timeout)
  139. # [17:35] <MikeSmith> it seems to me at least
  140. # [17:35] <anne> scope of the HTML WG isn't either
  141. # [17:35] * Quits: tH (Rob@83.100.252.160) (Ping timeout)
  142. # [17:35] <MikeSmith> maybe we should try to set some better boundaries
  143. # [17:36] <MikeSmith> some clearer boundaries
  144. # [17:36] <anne> +1 :p
  145. # [17:36] <MikeSmith> based on what the goals really need to be
  146. # [17:37] <MikeSmith> unless we still want to be riding this same merry-go-round a year from now or two years from now
  147. # [17:37] <MikeSmith> as new people join with more, um,... ideas
  148. # [17:38] <anne> that seems mostly for the W3C to solve though
  149. # [17:38] <MikeSmith> no
  150. # [17:38] <anne> although please tell me if there's anything we can do
  151. # [17:39] <MikeSmith> it is for the working group to solve, as a group, if it wants to have hope of being successful as a group and meeting its objectives
  152. # [17:40] <DanC> MikeSmith, are you interested/available to help shape mailing list traffic?
  153. # [17:41] <anne> MikeSmith, Hixie has tried that several times as have several others; it doesn't seem to work very well
  154. # [17:41] <MikeSmith> yeah, I'm interested
  155. # [17:41] * Quits: edas (edaspet@88.191.34.123) (Quit: http://eric.daspet.name/ et l'édition 2007 de http://www.paris-web.fr/ )
  156. # [17:42] <MikeSmith> anything you can do: Think about where are response to a particular posting is likely to lead -- to a useful resolution or just off into the weeds
  157. # [17:42] <anne> various times I tried to explain my viewpoints and I get back "No, you don't"
  158. # [17:42] <MikeSmith> and don't respond to the ones that look likely to go off into the weeds
  159. # [17:42] <DanC> one big problem is that we have little precedent for "useful resolution"
  160. # [17:42] <MikeSmith> DanC - right
  161. # [17:43] <DanC> the latency for spec updates is high. I think that's reasonable, but I'm not sure how to explain it to WG members.
  162. # [17:43] <anne> spec updates in response to comments?
  163. # [17:43] <DanC> right
  164. # [17:43] <anne> yeah, there's a backlog of around 5000 comments aiui
  165. # [17:44] * DanC noodles on an "I'm just a bill..." cartoon ;-)
  166. # [17:45] <MikeSmith> some of the people posting to the list are not really well integrated into the community around Web technologies and especially around implementation realities
  167. # [17:46] <MikeSmith> It is nice to try to educate them some about that stuff
  168. # [17:46] <MikeSmith> but IMHO, public-html should not be the place to do it
  169. # [17:47] <DanC> I tried to explain the real cost of a feature once or twice. I'm only willing to repeat myself so often. I'd like other people to shape traffic by citing earlier instruction from the chair etc, perhaps promoting it to the wiki
  170. # [17:47] * Quits: gavin (gavin@74.103.208.221) (Ping timeout)
  171. # [17:47] * Joins: Sander (svl@71.57.109.108)
  172. # [17:48] <anne> MikeSmith, yes, it would be nice if everyone invested a few months in some browser project
  173. # [17:48] <MikeSmith> anne - not suggesting they need to go that far
  174. # [17:49] <MikeSmith> in some cases, I doubt they are ever going to come around to understanding it
  175. # [17:49] <MikeSmith> it is as if we have two separate planes of understanding here
  176. # [17:49] <DanC> on the 5000 comment backlog... we might preempt that in a month or so.
  177. # [17:49] <MikeSmith> or multiple planes
  178. # [17:49] <anne> preempt?
  179. # [17:49] <DanC> yes; redirect the energy of the editors to different priorities
  180. # [17:50] <anne> such as?
  181. # [17:50] <MikeSmith> I think some simple things like people setting Reply-to: www-html headers (not just Cc'ing www-html or www-archive) when replying to "let's take this elsewhere" messages would help a lot
  182. # [17:50] <DanC> I can only ask the WAI PF WG to wait so long for an answer to http://www.w3.org/mid/p06110406c28ca5f5160d@%5B192.168.1.100%5D
  183. # [17:51] <DanC> (sent June 6)
  184. # [17:51] <anne> oh
  185. # [17:51] <anne> Hixie is doing research into real world usage of longdesc and such
  186. # [17:51] <DanC> good.
  187. # [17:51] <anne> and headers=
  188. # [17:51] <DanC> yes, there seems to be some light in the discussion, as well as heat
  189. # [17:51] <anne> one of the early statements was that accessibility advocates might have made matters worse...
  190. # [17:52] <DanC> just one? ;-)
  191. # [17:52] * Joins: gavin (gavin@74.103.208.221)
  192. # [17:52] <anne> http://krijnhoetmer.nl/irc-logs/whatwg/20070704
  193. # [17:53] * Joins: tH (Rob@83.100.252.160)
  194. # [17:54] <DanC> MikeSmith, one mailing list moderation tactic that was suggested was that people shouldn't reply to threads unless the chair says that thread is OK. I don't think that can work exactly, but I have been noodling mixing in the browser project notion, a la: if you want to discuss a new feature, you have to get at least one browser project to say "yes, that's worth WG time"
  195. # [17:55] * DanC can't see at a glance how whatwg/20070704 is related to anne's previous msg
  196. # [17:56] <MikeSmith> DanC - criterion of having a least one brower-project expression of interest sounds like an idea worth exploring
  197. # [17:57] <MikeSmith> or implementor expression of interest
  198. # [17:58] <anne> oops
  199. # [17:58] <anne> http://krijnhoetmer.nl/irc-logs/whatwg/20070703#l-255
  200. # [17:58] <MikeSmith> given that we have implementations other than browsers to consider
  201. # [17:58] <MikeSmith> karl would point out authoring tools as one
  202. # [17:58] <anne> 04 was marked visited and started with longdesc so I thought that was the one...
  203. # [17:58] <DanC> I have been tracking the wiki only a little bit. A regular (weekly?) heartbeat of "let's focus on these 3[or 4] issue this week" could help a lot. I do a little bit of that sort of thing in preparation for Hypertext CG telcons every other week.
  204. # [17:59] * MikeSmith needs to wander off for a bit ... back in while
  205. # [18:00] <anne> someone hijacked the topic/HTML with a lot of empty pages
  206. # [18:00] <DanC> yes, anne, I've been following up on concerns of my own around accessibility advocacy
  207. # [18:00] <anne> well, template pages
  208. # [18:01] <DanC> empty/template pages are indeed a documented antipattern. http://esw.w3.org/topic/OnlyMakeInterestingPages .
  209. # [18:01] <DanC> let's see who made them...
  210. # [18:02] <DanC> LauraCarlson. ah. perhaps at my prompting.
  211. # [18:08] <DanC> everybody chip in and delete a few?
  212. # [18:08] * DanC sent mail about http://esw.w3.org/topic/OnlyMakeInterestingPages
  213. # [18:12] * Joins: hyatt (hyatt@24.238.171.189)
  214. # [18:17] <anne> hi hyatt
  215. # [18:17] <anne> hyatt, http://html5.org/parsing-tests/testrunner.htm runs html5lib tests against browsers
  216. # [18:19] <anne> it has about 200 tests
  217. # [18:22] <DanC> hmm... I just tried testrunner.htm ; I don't understand the output
  218. # [18:22] <DanC> (a title would be nice too)
  219. # [18:23] <anne> you can click on any of the lines to get detailed output
  220. # [18:24] <anne> and you can make patches: http://html5.googlecode.com/svn/trunk/parser-tests/ :)
  221. # [18:24] <anne> http://code.google.com/p/html5/source
  222. # [18:26] <DanC> anne, I hope you don't mind too much that we're not releasing html4-differences right away after all.
  223. # [18:26] <anne> it's already out there
  224. # [18:26] <DanC> right
  225. # [18:27] * DanC starts http://esw.w3.org/topic/HtmlTestMaterials to make http://html5.org/parsing-tests/testrunner.htm easier to find
  226. # [18:27] * DanC tries to remember where those mobile tests are...
  227. # [18:27] <anne> http://simon.html5.org/articles/mobile-results
  228. # [18:27] <DanC> ah. they have googlemark on "html5 mobile tests"
  229. # [18:28] <anne> there's also http://james.html5.org/parsetree.html which provides a web interface to html5lib
  230. # [18:28] * Quits: tH (Rob@83.100.252.160) (Ping timeout)
  231. # [18:29] * Joins: tH (Rob@83.100.252.160)
  232. # [18:33] * Quits: tH (Rob@83.100.252.160) (Ping timeout)
  233. # [18:33] <DanC> interesting... that gives "No parse errors", despite the fact that I don't keep to the html5 constraints about li content
  234. # [18:33] <DanC> (on my homepage, http://www.w3.org/People/Connolly/ )
  235. # [18:34] * Joins: Lachy_ (chatzilla@203.158.59.119)
  236. # [18:34] <anne> there's a difference between syntax and language
  237. # [18:34] * Quits: Lachy_ (chatzilla@203.158.59.119) (Quit: ChatZilla 0.9.78.1 [Firefox 2.0.0.4/2007051502])
  238. # [18:34] <anne> which is what you're experiencing
  239. # [18:34] * Quits: Lachy (chatzilla@124.168.24.114) (Ping timeout)
  240. # [18:34] <anne> html5lib so far only produces a DOM, it doesn't check that DOM
  241. # [18:35] <anne> hsivonen is working on that
  242. # [18:35] <DanC> interesting... does that task have a home in http://code.google.com/p/html5/issues/list ?
  243. # [18:35] <DanC> evidently not
  244. # [18:36] * DanC linked parsetree.html under "checking tool prototyping/development" in http://esw.w3.org/topic/HtmlTaskBrainstorm
  245. # [18:37] <anne> html5lib is http://code.google.com/p/html5lib/
  246. # [18:37] * Joins: hasather (hasather@80.203.71.22)
  247. # [18:37] <anne> But we're not planning on conformance checking afaict
  248. # [18:37] <MikeSmith> DanC, on second thought, I wonder whether the group should just decide to put a moratorium on any discuseion of new features -- meaning yet-unimplemented features, not features such as canvas that are already implemented in a least two browsers
  249. # [18:37] <MikeSmith> because the real expression of interest is whether it has actually been implemented yet or not
  250. # [18:38] <MikeSmith> wholly or partially
  251. # [18:39] <MikeSmith> With an agreement that discussion of all that will take place in "Phase II" (or whatever)
  252. # [18:39] * Quits: hyatt (hyatt@24.238.171.189) (Quit: hyatt)
  253. # [18:39] <Philip`> Does that include prototype implementations, like <video>?
  254. # [18:40] <MikeSmith> Philip` - my vote would be, if it's implemented in at least two browsers, yes
  255. # [18:40] <MikeSmith> if only in one, no
  256. # [18:40] <MikeSmith> because if it's only in one, it is simply a feature of that browser
  257. # [18:40] <MikeSmith> not something that needs to be standardized
  258. # [18:41] <anne> the intent to implement is more important I think
  259. # [18:41] <Philip`> In that case, I guess new features would be proposed in the WHATWG and people would play around with prototype implementations for a while, and then eventually they'd take it to the HTML WG who would say "oh, okay" and would be too late to influence the design
  260. # [18:41] <anne> if browsers want a certain feature, specifying it in a way that makes sense and keeps it implementable in an interoperable way should just be done...
  261. # [18:42] <anne> it should*
  262. # [18:42] <MikeSmith> yeah, it should be done. the question is when and at what priority
  263. # [18:42] <anne> depends on market pressure
  264. # [18:42] <MikeSmith> I thought the priority of the group was supposed to be on specifying interoperable handling of existing content
  265. # [18:43] <DanC> I have tried to discourage discussion of yet-unimplemented features, MikeSmith, to little effect. a formal "moratorium" raises questions of which browsers count and who speaks for them and such, which is likely more heat than light. But I'm happy to see traffic shaping by citing my earlier messages about the cost of new features.
  266. # [18:43] * Joins: zcorpan (zcorpan@84.216.41.39)
  267. # [18:43] <DanC> I think the model where prototype design happens elsewhere is a good one.
  268. # [18:44] <MikeSmith> yeah
  269. # [18:44] <MikeSmith> the reality is, that's the way it happens anyway
  270. # [18:45] <DanC> quite
  271. # [18:47] <anne> I thought the idea of the HTML WG was to involve MS in that process and have it all RF from the start
  272. # [18:47] <anne> but I may have missed something
  273. # [18:48] <MikeSmith> yeah, complete silence from Chris and other MS participants is not so encouraging
  274. # [18:49] <anne> one question from MS about the interaction of DOM attributes and content attributes; several answers; no answer back from MS
  275. # [18:49] <MikeSmith> which is another reason why it would be prudent to focus the discussions on what is relevant to them and what we most need for them to contribute to
  276. # [18:50] <MikeSmith> which is interoperability and implementation conformance
  277. # [18:50] <anne> so far MS hasn't really helped much; much of the parsing section has been build (also recently) through the joys of reverse engineering
  278. # [18:50] <MikeSmith> that's really unfortunate
  279. # [18:51] <MikeSmith> anyway, IMHO, what's far less productive to be doing right now even than spending cycles discussing unimplemented features is: Discussion of document conformance
  280. # [18:51] <MikeSmith> that is, discussing what existing elements should be conformant in HTML5 and which should not
  281. # [18:52] <anne> it would be better if the focus was more on interop and what does HTML need more in addition to what is already implemented/proposed to become a viable alternative to Flash, Flex, Silverlight etc.
  282. # [18:53] <anne> instead of things like redesigning the language to just use <link> or something
  283. # [18:53] <MikeSmith> right
  284. # [18:54] <anne> all of that seems to happen at whatwg@whatwg.org :(
  285. # [18:54] <anne> although hsivonen recently started sending parser comments to public-html
  286. # [18:55] <anne> in hope to encourage people to more productive stuff iirc
  287. # [18:55] <MikeSmith> but to me, it's a unnecessary and distracting at this point to have any discussion about whether or not an existing element should be conformance, when we still need to spec expected behavior (if any) of how implementations are supposed to handle that element
  288. # [18:55] <MikeSmith> should be conformant, I meant
  289. # [18:56] <MikeSmith> in other words, better to be restricting discussion to implementation conformance
  290. # [18:57] <MikeSmith> part of the problem is, we have the same word -- "conformance" -- for describing two different things: document conformance and implementation conformance
  291. # [18:58] <MikeSmith> might be a lot better to talk about document "compliance" instead
  292. # [18:58] <MikeSmith> to make the distinction clear
  293. # [18:58] <Philip`> Could bring back document "validity"
  294. # [18:59] <MikeSmith> but as hsivonen will tell you, validity is only part of document conformance
  295. # [18:59] <MikeSmith> table integrity for example
  296. # [19:00] <Philip`> or call it 'document awesomeness', so authors have a motivation to write awesome HTML because it sounds really good
  297. # [19:00] <MikeSmith> heh
  298. # [19:00] <MikeSmith> awesome +1 to that idea!! (^з^)-☆Chu!!
  299. # [19:01] <Philip`> The spec could just redefine what 'valid' means, because almost nobody cares about the actual details as long as the checker returns "OK" for their pages
  300. # [19:01] <anne> HTML 5 already did that
  301. # [19:01] <anne> does*
  302. # [19:01] <anne> Hixie recently introduced the term "validator"
  303. # [19:02] <MikeSmith> great, but is that what we really need right now?
  304. # [19:02] <MikeSmith> What is the main problem the group is trying to solve?
  305. # [19:03] <MikeSmith> I don't think it is the problem of authors not knowing whether their content is valid.
  306. # [19:03] <anne> fixing all of HTML and making sure we have a viable competitor to closed-source web platforms
  307. # [19:03] <MikeSmith> amen
  308. # [19:04] <anne> (at least the markup/script side of that competitor; styling is so far out of scope)
  309. # [19:04] <MikeSmith> so I wonder if we can somehow manage to set the bounds of the discussion on the mailing list to that
  310. # [19:05] <MikeSmith> say, we are here together right now to focus our attention on how to precisely, thoroughly, and rigorously specify conformant, interoperable handling of existing content
  311. # [19:06] <DanC> the "viable competitor" work does seem like an important role for this WG, but I find it overwhelming.
  312. # [19:06] <MikeSmith> the discussions about new features and whether particular elements should be conformant or not can continue to take place elsewhere, like on www-html
  313. # [19:07] <zcorpan> MikeSmith: why?
  314. # [19:07] <DanC> well, MikeSmith , having Sam Ruby post about silverlight and about the W3C HTML WG seems... useful somehow.
  315. # [19:08] <MikeSmith> zcorpan - because we have to have some bounds for discussion in order to have any hope of getting any real work done and producing a Candidate Rec in a reasonable time frame
  316. # [19:08] <MikeSmith> instead of spinning our wheels
  317. # [19:09] <MikeSmith> and by real work, I mean production of test cases, for example
  318. # [19:09] * Joins: Lachy (chatzilla@203.158.59.119)
  319. # [19:09] <MikeSmith> which can't happen for a particular feature until there is a complete spec for that feature
  320. # [19:10] <zcorpan> MikeSmith: if something is implemented but not specced, you can still write tests to figure out what is implemented
  321. # [19:10] <anne> browser vendors would disagree
  322. # [19:10] <anne> (with MikeSmith)
  323. # [19:11] <anne> although I agree that having a spec is worth a lot
  324. # [19:11] <zcorpan> certainly helps getting interop for a feature :)
  325. # [19:11] <MikeSmith> anyway, my point is that I would think that should be the priority right now
  326. # [19:12] <zcorpan> agreed
  327. # [19:12] <MikeSmith> to the exclusion of everything else, for the time being
  328. # [19:12] <Philip`> Writing test cases seems like a very helpful way to find where the spec is incomplete and how it should be fixed, so the feature's spec doesn't need to be complete before that stage
  329. # [19:13] <anne> yeah, see <canvas>, parsing, etc.
  330. # [19:13] <MikeSmith> true
  331. # [19:13] <MikeSmith> the spec-writing process is always iterative
  332. # [19:13] <MikeSmith> or should be
  333. # [19:13] <MikeSmith> by complete spec I did not mean final spec
  334. # [19:13] <Philip`> (I guess implementors don't care so much about edge cases because they'll just implement what's easy, whereas people writing tests are specifically trying to poke all the edge cases, and so they find the cases where the spec is insufficient)
  335. # [19:13] <anne> lol, someone wasted a day on making empty pages
  336. # [19:13] <anne> talking about good use of your time...
  337. # [19:13] <MikeSmith> I mean a spec that aims to be complete as it can be
  338. # [19:15] <MikeSmith> question while I'm thinking of it: Do any current desktop browsers actually do anything with the content of table summary attribute?
  339. # [19:15] <MikeSmith> do they do any handling with it in any way?
  340. # [19:15] <DanC> somebody put some table test data in the wiki...
  341. # [19:15] <MikeSmith> anything other than just putting it into the DOM I mean
  342. # [19:16] <anne> Firefox has some handling
  343. # [19:16] * DanC wishes for a link from http://esw.w3.org/topic/HtmlTestMaterials to <canvas> test materials
  344. # [19:16] <anne> right click, properties, etc.
  345. # [19:16] <Philip`> Do you count desktop browsers with extensions?
  346. # [19:16] <anne> DanC, http://canvex.lazyilluminati.com/tests/tests/
  347. # [19:16] <MikeSmith> anne - why, I wonder
  348. # [19:16] <anne> DanC, http://tc.labs.opera.com/html/canvas/
  349. # [19:16] <MikeSmith> Philip` - anything
  350. # [19:16] * DanC waits a long time for esw.w3.org :-/
  351. # [19:16] <anne> MikeSmith, just because
  352. # [19:17] <anne> they expose other metadata attributes in a similar way
  353. # [19:17] <MikeSmith> anne - right
  354. # [19:17] <MikeSmith> they shouldn't
  355. # [19:17] <Philip`> They show the longdesc URL in the image properties window, though it's rather useless since it's not even a clickable link
  356. # [19:17] <anne> I think it was part of some HTML 4 conformance plan that was eventually dropped because HTML 4 is buggy
  357. # [19:17] <MikeSmith> except in Firebug or whatever
  358. # [19:17] <MikeSmith> Philip` - yeah, another good example
  359. # [19:18] <Philip`> DanC: http://hixie.ch/tests/adhoc/html/canvas/ too
  360. # [19:18] <MikeSmith> (example of uselessness, I meant)
  361. # [19:18] <Philip`> (Those three collections are the only ones I know of)
  362. # [19:18] * Quits: schepers (schepers@69.134.24.226) (Quit: Free at last!)
  363. # [19:19] * Philip` should probably find somewhere better to host his tests
  364. # [19:19] <anne> I can get you philip.html5.org but I'm not sure if that's better
  365. # [19:19] <anne> there's also quite a lot in http://simon.html5.org/test/html/
  366. # [19:20] <anne> but not all <canvas> related
  367. # [19:20] <anne> not <canvas> related*
  368. # [19:20] <DanC> there's some test materials in/near http://esw.w3.org/topic/HTML/TableAccessibility
  369. # [19:20] * DanC just linked it from http://esw.w3.org/topic/HtmlTestMaterials , along with canvas stuff
  370. # [19:21] <DanC> ugh... esw.w3.org is slow
  371. # [19:23] <anne> most of your questions there are covered by parser tests btw
  372. # [19:23] <anne> a very useful tool for browser parsing debugging is http://software.hixie.ch/utilities/js/live-dom-viewer/
  373. # [19:23] <Philip`> I expect I'll finish writing all the canvas tests some time soonish (at least until the spec gets updated and the tests have to change), so I may try to sort something out then about not hosting them on the dust-infested box on the floor next to me
  374. # [19:24] <DanC> in the tasks survey, quite a few people indicated interest in manual test reporting for various browsers.
  375. # [19:24] <DanC> it would be good to get a feedback loop with them going
  376. # [19:26] <Philip`> I can easily run the canvas tests myself in all the Windows/Linux browsers (so Safari-OSX is about the only interesting one missing), and they only take a couple of minutes each, so that part probably wouldn't benefit from other people helping
  377. # [19:27] * Joins: schepers (schepers@69.134.24.226)
  378. # [19:27] <anne> same for parsing tests
  379. # [19:27] <anne> although I've yet to set up some reporting mechanism
  380. # [19:27] <anne> maybe a few python scripts that compare some numbers...
  381. # [19:28] * Joins: tH (Rob@83.100.252.160)
  382. # [19:30] <Philip`> If somebody wants to implement getImageData in WebKit, that would make the canvas tests even quicker to run :-)
  383. # [19:30] <Philip`> (though it'd be nice to implement it correctly for once, since I already have to work around two differently buggy implementations of the function :-p )
  384. # [19:40] <DanC> Philip`, even if you can run all the tests yourself, I think it helps a lot for other WG members to learn to do it, and to learn to read the relevant parts of the spec while they're at it, and to help fix both the tests and the spec as a result.
  385. # [19:41] <DanC> in fact, I think that would be a good use of telcon time... presenting test materials.
  386. # [19:41] * DanC finds it somewhat surreal to see "wiki page for..." used in subject lines
  387. # [19:44] * Quits: Lionhear1 (robin@66.57.69.65) (Ping timeout)
  388. # [19:44] <zcorpan> http://wiki.whatwg.org/wiki/Test_cases
  389. # [19:55] * Quits: gavin (gavin@74.103.208.221) (Ping timeout)
  390. # [20:00] <zcorpan> hmm, perhaps i should do a detailed review of section 8.3
  391. # [20:00] * Joins: gavin (gavin@74.103.208.221)
  392. # [20:00] <zcorpan> oh wait, i already have: http://lists.whatwg.org/pipermail/whatwg-whatwg.org/2006-January/005361.html
  393. # [20:00] * Quits: schepers (schepers@69.134.24.226) (Quit: Free at last!)
  394. # [20:01] * Joins: schepers (schepers@128.30.52.30)
  395. # [20:04] * Quits: schepers (schepers@128.30.52.30) (Quit: Free at last!)
  396. # [20:05] * Joins: schepers (schepers@128.30.52.30)
  397. # [20:18] * Quits: MikeSmith (MikeSmith@mcclure.w3.org) (Ping timeout)
  398. # [20:30] * Joins: hyatt (hyatt@24.238.171.189)
  399. # [20:37] * Joins: dbaron (dbaron@63.245.220.242)
  400. # [20:58] * Quits: hyatt (hyatt@24.238.171.189) (Quit: hyatt)
  401. # [21:03] * Quits: Charl (charlvn@196.209.104.228) (Quit: Leaving)
  402. # [21:48] <DanC> ok, I just updated http://www.w3.org/html/wg/ in preparation for Hypertext CG telcon tomorrow.
  403. # [21:48] <DanC> some have suggested I should blog when I do this.
  404. # [21:49] <DanC> a blog articlet that parallels the CVS commit message for http://www.w3.org/html/wg/ would be more boring than watching paint dry: "removed 9 May decision from current events"
  405. # [21:49] <anne> html4-differences has a short link too now: http://www.w3.org/html/wg/html5/diff/
  406. # [21:50] <DanC> ah...
  407. # [21:50] <anne> although the URL seems slightly confusing
  408. # [21:50] <Hixie> if you can give me an API that my script can blog to i'd be happy to add it to the pile
  409. # [21:50] <Hixie> right now each commit sends an e-mail, twitters, and commits to svn and cvs
  410. # [21:51] <DanC> yeah... /html5/diff/ is kinda odd... hmm... I'm going with it for nwo
  411. # [21:53] * Joins: jgraham (jgraham@81.86.213.61)
  412. # [21:54] <DanC> hmm... the /html/ blog has comments... http://www.w3.org/QA/2007/07/html-elements-list.html#comments
  413. # [21:54] <DanC> that suggests it's using MT, which has an XMLRPC API, I think
  414. # [21:55] <DanC> it seems to be a category in the QA blog. interesting. http://www.w3.org/QA/archive/technology/html/
  415. # [21:55] <anne> yeah, it's MT
  416. # [21:55] <DanC> ah... good... karl (I assume) categorized entries back to Oct 2006 and beyond
  417. # [21:56] <anne> with nice encoding errors on which XML would fail :)
  418. # [21:56] <anne> fortunately all is text/html
  419. # [21:56] * DanC doesn't think encoding errors are nice
  420. # [21:57] <anne> if browsers started failing on encoding errors in XML including feed parsing I wonder how much stuff would still work
  421. # [21:57] <DanC> if browsers had failed all along, the encoding errors would never have seen the light of day
  422. # [21:58] <anne> besides browser interop issues encoding is probably one of the major pains of the web
  423. # [21:58] <Hixie> the browsers not failing on encoding errors was unintentional
  424. # [21:58] <Hixie> in fact i believe it was a bug in expat, though i may be wrong
  425. # [21:58] <anne> for Firefox, yeah
  426. # [21:58] <DanC> yeah, we all live with the bugs now.
  427. # [21:58] <Hixie> unfortunately once the bug was out there, it was too late to change
  428. # [21:58] <DanC> encoding bugs are great fodder for all sorts of "XML has failed, utterly" articles.
  429. # [21:59] <Hixie> people like to make extreme sensational statements :-)
  430. # [21:59] <anne> encoding bugs and RFC3023
  431. # [21:59] <DanC> a more reasonable headline would be "international text is really hard to get right", but editors aren't much for reasonable headlines.
  432. # [21:59] <anne> which is not really realistic either
  433. # [21:59] <anne> not treating text/xml as application/xml for instance...
  434. # [21:59] <anne> (which is what every sane impl does)
  435. # [22:01] * jgraham believe the web would have been a whole lot less successful if browsers were strict from the start
  436. # [22:02] * Quits: xover (xover@193.157.66.5) (Ping timeout)
  437. # [22:03] * Quits: gavin (gavin@74.103.208.221) (Ping timeout)
  438. # [22:03] <anne> if someone can define strict for me first I'll point out some flaws 8-)
  439. # [22:03] * Quits: laplink (link@193.157.66.214) (Ping timeout)
  440. # [22:05] <Philip`> Ooh, actual data on the list
  441. # [22:05] <jgraham> Philip`: Yeah, it's wonderful. Seems to be someone's master's thesis :)
  442. # [22:06] <anne> that's from long ago btw
  443. # [22:06] <DanC> hmm... a few people said the'd dot heir spec section reviews by 30 June... Debi Orton, Marco Neumann, Robert Burns, Henrik Dvergsdal . I don't recall seeing those.
  444. # [22:06] <jgraham> It seems to be dated 12th June 2006
  445. # [22:06] <Philip`> It's interesting how there's so much more valid CSS than valid HTML out there
  446. # [22:07] <Philip`> (I think I saw that site a long time, but didn't read it in any detail)
  447. # [22:08] * Joins: gavin (gavin@74.103.208.221)
  448. # [22:08] * Philip` wonders how much (syntactically) valid JavaScript there is
  449. # [22:08] * jgraham wonders how good a sample one gets from the open directory project
  450. # [22:09] <Philip`> First, define "good" :-)
  451. # [22:09] <jgraham> Representatice of the parent population :)
  452. # [22:09] <jgraham> s/Representatice/Representative/
  453. # [22:10] <Philip`> Wouldn't it be better to weight the results towards pages that people look at most frequently, because those are the ones that people actually care about?
  454. # [22:10] <DanC> hmm... is http://triin.net/2006/06/12/CSS the middle of something? it starts "67.20% of pages uses CSS in one way or another." without saying where it gets the list of pages or whatever.
  455. # [22:11] * Quits: mjs (mjs@64.81.48.145) (Ping timeout)
  456. # [22:11] <Philip`> (like, you wouldn't want to break one of the top 100 sites, but you wouldn't care about breaking all of the bottom 100 sites because nobody's even going to look at them)
  457. # [22:11] <Philip`> DanC: The links on the left point to the rest of it
  458. # [22:11] <jgraham> Maybe. It depends what you're interested in I guess. Maybe the "population" that's relevant is the population of visited pages
  459. # [22:12] <jgraham> where by visited I mean "pages that people come across in their browser"
  460. # [22:12] <jgraham> Which I think is a long-winded way of saying "yes"
  461. # [22:12] <jgraham> But I don't think Hixie's stats do that either
  462. # [22:13] <Philip`> I remember there was some press release about the kinds of pages people look at in Opera Mini, so I guess they're logging all the traffic, so maybe if they could just release that data... :-)
  463. # [22:13] <DanC> ah... now I see, Philip` . Meanwhile I sent mail. I suppose it makes me look a little dumb, but probably the answer will help others.
  464. # [22:15] <jgraham> Philip`: Did you ever make your survey tool avaliable?
  465. # [22:15] * Joins: mjs (mjs@64.81.48.145)
  466. # [22:15] <DanC> this is hardly a scientific representative survey of the whole web, but it's hard data nonetheless.
  467. # [22:17] <Philip`> http://www.opera.com/pressreleases/en/2006/08/08/ - oh, that was it
  468. # [22:17] <Philip`> ("The most popular sites are Google search, community and dating sites, e-mail and domains for more grown-up entertainment.")
  469. # [22:17] <Philip`> (so we should research all of those in order to not break them)
  470. # [22:17] * DanC wanders off for lunch and such...
  471. # [22:18] <Philip`> jgraham: I didn't, because it was kind of rubbish and it's easier to write a new one than fix it
  472. # [22:18] <jgraham> Philip`: OK
  473. # [22:18] <Philip`> (and it really needs a fast HTML parser before it's usable for large-scale surveys)
  474. # [22:19] <jgraham> I wonder how much slower html5lib is than the perl thing this guy used
  475. # [22:20] <Philip`> The one he used is implemented in C
  476. # [22:21] <Philip`> (with a Perl interface)
  477. # [22:21] * anne has some hopes for the Java thing from hsivonen
  478. # [22:23] <jgraham> Oh OK
  479. # [22:24] * Philip` wonders if it'd be possible to use his university's distributed computation system to do stuff
  480. # [22:38] * Joins: xover (xover@193.157.66.5)
  481. # [22:39] <Philip`> DanC: Looks like Asbj<thing>rn's name got mangled on http://www.w3.org/html/wg/il16 again
  482. # [22:40] <DanC> yup. :-/
  483. # [22:41] <anne> fwiw, can't make the telcon
  484. # [22:41] * anne -> greece
  485. # [22:41] <Hixie> just so everyone is aware and doesn't wonder if i died or something, i'm going to be on vacation for 3 weeks starting sunday
  486. # [22:41] <anne> :p
  487. # [22:52] * Quits: ROBOd (robod@86.34.246.154) (Quit: http://www.robodesign.ro )
  488. # [22:58] * Quits: mjs (mjs@64.81.48.145) (Quit: mjs)
  489. # [23:02] * Quits: anne (annevk@213.236.208.22) (Ping timeout)
  490. # [23:26] * Joins: MikeSmith (MikeSmith@mcclure.w3.org)
  491. # [23:27] <hsivonen> DanC: Re: what Anne was saying about my activities: see http://hsivonen.iki.fi/validator-about/#src for the current parser code in Java (nu.validator.htmlparser in the htmlparser module). I'll announce it when there's a runnable library to announce.
  492. # [23:28] <hsivonen> I've implemented the tokenizer in Java and am now implementing the tree builder
  493. # [23:28] <hsivonen> Also, I implemented June 22 version of encoding sniffing
  494. # [23:54] * Quits: MikeSmith (MikeSmith@mcclure.w3.org) (Client exited)
  495. # [23:54] * Joins: mjs (mjs@17.255.106.198)
  496. # [23:56] <mjs> hi everyone
  497. # [23:58] * Parts: hasather (hasather@80.203.71.22)
  498. # [23:59] * Joins: MikeSmith (MikeSmith@mcclure.w3.org)
  499. # Session Close: Fri Jul 06 00:00:00 2007

The end :)