/irc-logs / freenode / #whatwg / 2008-02-26 / end

Options:

  1. # Session Start: Tue Feb 26 00:00:00 2008
  2. # Session Ident: #whatwg
  3. # [00:01] <Lachy> oh, ok. I should have read the IRC log before posting.
  4. # [00:01] <Hixie> heh
  5. # [00:03] * jgraham_ still believes that making language features dependent on all major browsers changing their parsing in a largely untested way is not the greatest idea.
  6. # [00:04] <jgraham_> But I've said that before and I don't have any new arguments or anything
  7. # [00:05] <annevk> we can always revise our evil plans for world domination in two years
  8. # [00:05] <Lachy> jgraham_, we can always revisit the issue again if browsers come back after trying it and say it causes too many provlems
  9. # [00:09] <gsnedders> jgraham_: I'm currently leaving towards having supper together on Fri, FWIW
  10. # [00:11] * Joins: eseidel (n=eseidel@nat/google/x-7e4b89528cc9758c)
  11. # [00:19] <Hixie> jgraham_: i agree
  12. # [00:23] * mpt wonders whether there should be an element for captions of form elements
  13. # [00:24] <mpt> If someone tabs to an input field, the browser automatically scrolls to show as much of the field as possible, but doesn't scroll to show that field's caption, because it doesn't know what a caption is
  14. # [00:24] * Quits: tantek (n=tantek@adsl-63-195-114-133.dsl.snfc21.pacbell.net)
  15. # [00:26] <mpt> If there was a caption element that was tied to the input element, the browser would know to scroll to show it
  16. # [00:26] <Hixie> isn't that <label>?
  17. # [00:26] <mpt> no
  18. # [00:27] <mpt> <label> appears above or (in LTR scripts) to the left of an element
  19. # [00:27] <mpt> the caption appears underneath, and is usually in smaller print
  20. # [00:27] <Hixie> can you show me a page with the difference you are indicating?
  21. # [00:28] <mpt> <https://launchpad.net/projects/+new>, but it requires a Launchpad account
  22. # [00:29] * mpt looks for another
  23. # [00:30] <Hixie> annevk: would be cool if one could bookmark the settings on http://html5.org/tools/web-apps-tracker
  24. # [00:30] <Hixie> would also be cool to have an atom feed or something for that
  25. # [00:31] <Hixie> and finally, it would also be cool to adda Google Gears icon to the list of vendors, maybe with the letter 'r' as the way for me to trigger it
  26. # [00:31] <annevk> is anyone actually using that interface?
  27. # [00:31] <Hixie> they have apparently asked if they can be added since they are about to start doing more html5-y stuff
  28. # [00:31] * Joins: jruderman (n=jruderma@guest-228.mountainview.mozilla.com)
  29. # [00:31] <Hixie> i dunno if anyone else is, but i certainly use it
  30. # [00:31] <annevk> i guess i'm not representative
  31. # [00:32] <Hixie> from #webkit:
  32. # [00:32] <Hixie> 00:28 < aroben> Hixie: I find the tracker very handy, though
  33. # [00:32] <Hixie> (referring to the page)
  34. # [00:32] <annevk> sorry, i'm using the tracker *a lot*
  35. # [00:32] <annevk> i meant the UA-specific crap
  36. # [00:32] <annevk> all those options
  37. # [00:32] <annevk> (the colors and icons in the changelog lines are useful though, to me)
  38. # [00:33] <Hixie> i like the options, i would probably use it more to point to other people if i could bookmark them though
  39. # [00:33] * Joins: csarven (n=nevrasc@modemcable130.251-202-24.mc.videotron.ca)
  40. # [00:33] <Hixie> if you want an icon for gears, there's one at http://google-gears.googlecode.com/svn/trunk/gears/ui/common/icon_16x16.png
  41. # [00:35] <mpt> Hixie, in <http://login.live.com/login.srf?wa=wsignin1.0&rpsnv=10&ct=1203982238&rver=4.5.2130.0&wp=MBI&wreply=http:%2F%2Fmail.live.com%2Fdefault.aspx&id=64855> both the text fields have a label and a caption
  42. # [00:35] <mpt> (though in that page the scrolling isn't important, because you can probably see both without scrolling already)
  43. # [00:35] <Hixie> i disagree with calling that a caption
  44. # [00:35] <Hixie> but i agree they are related to the field
  45. # [00:35] <Hixie> and are not labels
  46. # [00:36] <Hixie> they're like the (?) part of the checkbox labels
  47. # [00:36] <Hixie> interesting cases
  48. # [00:36] <mpt> I don't know what to call them other than captions
  49. # [00:37] <mpt> <https://www.google.com/accounts/NewAccount?service=mail&continue=http%3A%2F%2Fmail.google.com%2Fmail%2Fe-11-10b567b3542c73d051854a3694480bd-27a8b709838b6e5bc593de69344785f3ae19edf1&type=2> has examples too
  50. # [00:38] <Hixie> yeah it's a common idionm
  51. # [00:38] <Hixie> idiom
  52. # [00:38] <Hixie> most are basically help strings
  53. # [00:38] <mpt> Tab to the "Secondary email" field, and you won't see its caption, because the browser doesn't know any better.
  54. # [00:38] <Hixie> in fact i think all of these are really help strings
  55. # [00:39] <Hixie> might make sense to have a <label> ... <input> <help>...</help> </label> or <label for="">...</label> <input id=""> <help for="">...</help> construct
  56. # [00:41] <mpt> But then, see that Gmail has "Learn More" links
  57. # [00:41] <mpt> *That* is help
  58. # [00:42] <Hixie> sure, you can have inline short help and external longer help
  59. # [00:42] <Hixie> i've noted this for wf3.
  60. # [00:42] <mpt> Right, it's a continuum
  61. # [00:42] <mpt> thanks
  62. # [00:42] <Hixie> thank _you_!
  63. # [00:43] <annevk> man, web-apps-tracker is a hack
  64. # [00:43] <annevk> i've to add google-gears to like five different places
  65. # [00:43] <annevk> scary
  66. # [00:43] <Hixie> hehe
  67. # [00:43] <Hixie> feel free to make a new one that's better :-)
  68. # [00:44] <Hixie> afk. i really do have to actually go to work... or i'll starve.
  69. # [00:45] * Joins: othermaciej (n=mjs@nat/apple/x-2d3747264849a62c)
  70. # [00:53] <annevk> ok, four different places
  71. # [00:53] <annevk> gears is now supported, all the fancy features will have to wait
  72. # [00:56] * Parts: SadEagle (n=maksim@cpe-69-202-89-106.twcny.res.rr.com) ("Konversation terminated!")
  73. # [00:56] <annevk> Hixie, "r" triggers Google Gears as I couldn't think of something better either
  74. # [00:58] * annevk -> bed
  75. # [01:10] * Quits: svl (n=me@ip565744a7.direct-adsl.nl) (Nick collision from services.)
  76. # [01:10] * Joins: svl (n=me@ip565744a7.direct-adsl.nl)
  77. # [01:12] * Quits: svl (n=me@ip565744a7.direct-adsl.nl) (Client Quit)
  78. # [01:12] * Quits: eseidel (n=eseidel@nat/google/x-7e4b89528cc9758c) (calvino.freenode.net irc.freenode.net)
  79. # [01:21] * Quits: tndH (i=Rob@adsl-87-102-35-237.karoo.KCOM.COM) (Read error: 104 (Connection reset by peer))
  80. # [01:37] * Quits: cgriego (n=cgriego@216.138.69.206)
  81. # [01:46] * Joins: eseidel (n=eseidel@nat/google/x-7e4b89528cc9758c)
  82. # [01:57] * Parts: hasather (n=hasather@90-231-107-133-no62.tbcn.telia.com)
  83. # [01:59] <Hixie> annevk: cool, thanks
  84. # [02:10] * Joins: weinig_ (n=weinig@17.255.110.169)
  85. # [02:16] * Quits: aroben (n=aroben@unaffiliated/aroben) (Read error: 104 (Connection reset by peer))
  86. # [02:27] * Quits: weinig (n=weinig@17.203.15.180) (Read error: 110 (Connection timed out))
  87. # [02:28] * Joins: weinig (n=weinig@17.203.15.180)
  88. # [02:34] * Joins: weinig__ (n=weinig@17.255.110.169)
  89. # [02:35] * Quits: weinig_ (n=weinig@17.255.110.169) (Read error: 104 (Connection reset by peer))
  90. # [02:36] * Joins: weinig_ (n=weinig@17.203.15.180)
  91. # [02:37] * Quits: weinig (n=weinig@17.203.15.180) (Read error: 104 (Connection reset by peer))
  92. # [02:46] * Joins: webben (n=benh@82.153.59.221)
  93. # [02:46] * Quits: webben (n=benh@82.153.59.221) (Client Quit)
  94. # [02:48] * Joins: tantek (n=tantek@m730f36d0.tmodns.net)
  95. # [02:48] * Hixie summons zcorpan
  96. # [02:49] <Hixie> anyone have an opinion on whether <ol> is appropriate for lists that happen to be ordered as opposed to lists where the order is significant?
  97. # [02:49] <Hixie> e.g. should a list of place names, ordered alphabetically, be <ol> or <ul>?
  98. # [02:50] <tantek> Hixie, example of each?
  99. # [02:50] <tantek> <ol> - since "alphabetically" could mean any number of a different canonical sorting orders given variations in language/culture/punctuation treatment.
  100. # [02:51] <Hixie> hm, interesting
  101. # [02:51] <tantek> no way to really automatically discern how such ordering was done if in a <ul> and thus if it has significance or not.
  102. # [02:51] * Joins: webben (n=benh@82.153.59.221)
  103. # [02:51] <Hixie> you used the opposite argument to zcorpan in his feedback in 2006
  104. # [02:51] <Hixie> :-)
  105. # [02:51] <tantek> I've learned more about alphabetic sorting since then.
  106. # [02:52] <Hixie> no i mean he argued the opposite of you just now, in his mail from 2006
  107. # [02:52] <Hixie> he says:
  108. # [02:52] <Hixie> > I think <ol> is a list where the order is significant to the meaning;
  109. # [02:52] <Hixie> > where the order is emphasized. For lists that happen to be ordered but
  110. # [02:52] <Hixie> > the order isn't really of a big significance or isn't of higher
  111. # [02:52] <Hixie> > significance than the global order of the document, <ol> shouldn't be
  112. # [02:52] <Hixie> > used IMHO.
  113. # [02:52] <Hixie> > Otherwise people might use <ol> whenever a list happens to be in order, e.g.
  114. # [02:52] <Hixie> > an A-Z list...
  115. # [02:52] <tantek> One way to think about it is, if the author considers the ordering to be something they did deliberately (whether by hand in the markup), then the way to communicate "deliberate ordering" is by using an <ol>
  116. # [02:53] <tantek> (whether by hand in the markup, or by selection of some backend sorting option in a query)
  117. # [02:53] <Hixie> so <ul> elements must be in effectively random order?
  118. # [02:53] <tantek> not necessarily
  119. # [02:53] <tantek> randomness isn't necessary
  120. # [02:54] * Joins: webben_ (n=benh@dip5-fw.corp.ukl.yahoo.com)
  121. # [02:54] <Hixie> right, i mean effectively random, as opposed to necessarily truly random
  122. # [02:54] <tantek> If the author considers the order to be arbitrary (not necessarily random), or doesn't care about the order, then a <ul> makes sense.
  123. # [02:54] * Quits: weinig__ (n=weinig@17.255.110.169) (Read error: 110 (Connection timed out))
  124. # [02:54] <tantek> <ol> = author cared about the order. <ul> = author didn't care about the order. That simple. I think.
  125. # [02:54] <Hixie> i suppose a list of place names that were arbitrarily ordered alphabetically as opposed to ordered alphabetically for any particular reason would mean <ul> then
  126. # [02:56] <tantek> again, that assumes that a universal canonical alphabetic ordering exists that everyone implicitly understands, which is a false assumption.
  127. # [02:56] <Hixie> how do you mean?
  128. # [02:57] <tantek> alphabetic ordering is not just by characters per unicode code points for example
  129. # [02:58] <Hixie> sure. i mean, the ordering could be anything, i was just using "alphabetical" as an arbitrary ordering
  130. # [02:58] <tantek> e.g. depending on the context of the data, say it were English names of works of art (like movies or songs)
  131. # [02:59] <Hixie> for example the list could be in order of the population size per place, but if that was just an arbitrary order selected by the author simply for presentational reasons, that would be a <ul>
  132. # [02:59] <tantek> often a leading word "The" is omitted from the canonical sorting
  133. # [02:59] <tantek> there are numerous other such rules. and that's just for English. names of works of art.
  134. # [02:59] <tantek> so there are language specific rules. domain specific rules. etc.
  135. # [02:59] <Hixie> whereas if the list was a list of places, ordered by population size to show the differences in population sizes, <ol> would be correct
  136. # [03:00] <tantek> I think for whatever reason, if the author intends an order, then it should be marked up with <ol>
  137. # [03:00] <tantek> One thought experiment you could do is, if the browser scrambled the set of <li>s inside a <ul>, how upset would the author be when viewing the page?
  138. # [03:01] <Hixie> right
  139. # [03:01] <tantek> so a list of places, ordered by name (or population), that would be bad
  140. # [03:01] <tantek> but a grocery shopping list, not a problem
  141. # [03:01] <Hixie> so if i'm just listing places i've been, and i happen to have put them in order of population size, i don't care if the browser reorders them to be by order of how much rainfall they get per year
  142. # [03:02] <tantek> right, if the author didn't take care with the order, then a <ul> - hence my reasoning above
  143. # [03:02] <tantek> "for whatever reason, if the author intends an order, then it should be marked up with <ol>"
  144. # [03:02] <tantek> otherwise use <ul>
  145. # [03:02] <Hixie> but if i have a grocery list in order of importance, or in order of which aisle the stuff is found in, then i'd want <ol>
  146. # [03:02] <Hixie> well
  147. # [03:02] <Hixie> i'm saying the author might "intend an order" but not care about it at the same time
  148. # [03:02] <tantek> right, but I have yet to see such a grocery list ;)
  149. # [03:03] <tantek> then intend an order wins out
  150. # [03:03] <tantek> the intent is enough to justify an <ol>
  151. # [03:03] * Hixie often puts grocery lists in order of importance, so that he can make sure to get the important stuff first, since he is limited by volume when cycling from the shop :-)
  152. # [03:03] <Hixie> hmm
  153. # [03:03] <tantek> the intent doesn't have to be "big" or "significant" - because who decides that?
  154. # [03:03] <Hixie> so zcorpan is saying that intent doesn't win out, but that care does
  155. # [03:04] <Hixie> i wonder which is more useful
  156. # [03:04] <tantek> given that authors are the ones writing the markup, I find that going with respecting/modeling author intent usually produces better results. ;)
  157. # [03:04] <tantek> rather than some academic argument about how "significant" the ordering is
  158. # [03:05] <Hixie> well, the default rendering of <ul> is bullets, and i'd want bullets if i gave a list of places i'd been to, even if i ordered them (arbitrarily) by population size or lexically or by rainfall
  159. # [03:05] <Hixie> so that would argue against what you're saying :-)
  160. # [03:05] * tantek puts shopping lists into clusters by type of store. two levels of unordered lists.
  161. # [03:06] <tantek> the English "i'd want bullets if i gave a list of places i'd been to, even if i ordered them" translates to the CSS: ol.places { list-style:disc }
  162. # [03:06] <tantek> because that "want" is presentational
  163. # [03:06] <Hixie> sure, but authors aren't going to think that way
  164. # [03:06] <Hixie> they'll just use <ul>
  165. # [03:07] <Hixie> and pretend that they didn't order the items
  166. # [03:07] <tantek> that's fine, then they are saying the order doesn't matter
  167. # [03:07] <Hixie> and that they just _happen_ to be in alphabetical order
  168. # [03:07] <Hixie> right -- the order doesn't matter, even though there is one intended
  169. # [03:07] <Hixie> they don't care about the order, even though they intended one
  170. # [03:08] <tantek> no. if they are pretending they didn't order the items, then they didn't order them.
  171. # [03:08] <tantek> we can't really assume otherwise.
  172. # [03:08] * Quits: webben (n=benh@82.153.59.221) (Connection timed out)
  173. # [03:08] <Hixie> <ul> <li> Apple <li> Banana <li> Cherry </ul>
  174. # [03:08] <Hixie> i ordered that list.
  175. # [03:08] <tantek> in this case no care = no intent.
  176. # [03:08] <Hixie> that i'm using <ul> doesn't change the fact that i did indeed give an order
  177. # [03:09] <Hixie> now i might not care if the browser changes the order
  178. # [03:09] <Hixie> but i still intended one
  179. # [03:10] <tantek> so is the distinction then an intent of default presentational order vs. a semantic order?
  180. # [03:10] <tantek> because one could argue your ordering of that list was purely for presentational purposes. that is, you only intended it for presentational purposes, not for anything semantic.
  181. # [03:11] <Hixie> right
  182. # [03:11] <Hixie> i ordered it just to make it easier to scan
  183. # [03:12] <tantek> ok easy enough to modify my previous statement to take that into account:
  184. # [03:12] * Quits: phsiao (n=shawn@nat/ibm/x-467735d6b6dc2490)
  185. # [03:12] <tantek> "for whatever reason, if the author intends an order with some meaning behind it beyond just the presentation, then it should be marked up with <ol>"
  186. # [03:20] <Hixie> ok, i've tried to update the spec to explain this, along with some examples. http://www.whatwg.org/specs/web-apps/current-work/multipage/section-lists0.html
  187. # [03:21] * tantek clicks and hope his browser doesn't lock up.
  188. # [03:22] <tantek> BTW, really like the *zero* or more. About darn time.
  189. # [03:22] <tantek> do you need to say "change the meaning of the document"?
  190. # [03:22] <tantek> wouldn't just "change the meaning." be sufficient ?
  191. # [03:22] <Hixie> change the meaning of what?
  192. # [03:23] <tantek> per English implied object resolution rules
  193. # [03:23] <tantek> the previous clause
  194. # [03:23] <tantek> "changing the order"
  195. # [03:23] <tantek> also implied object
  196. # [03:23] <tantek> see previous clause
  197. # [03:23] <tantek> "where the items have been intentionally ordered"
  198. # [03:23] <tantek> also implied object
  199. # [03:23] <tantek> see previous clause
  200. # [03:24] <tantek> "a list of items"
  201. # [03:24] <tantek> resolution complete
  202. # [03:24] <tantek> no need to bring "the document" into it.
  203. # [03:24] <Hixie> hmmm
  204. # [03:24] <Hixie> i _would_ like to remove the document from that setence
  205. # [03:24] <tantek> doesn't everyone read/write English like code?
  206. # [03:25] <tantek> ;)
  207. # [03:25] <othermaciej> the way I think of it is <ol> --> numbered list, <ul> --> bulleted list
  208. # [03:25] <othermaciej> I would expect that is the typical author's operational understanding as well
  209. # [03:25] * tantek smacks othermaciej for presentational-major thinking.
  210. # [03:26] <Hixie> othermaciej: i agree, but i think the current definition pretty much matches that
  211. # [03:26] <tantek> what do bullets sound like?
  212. # [03:26] <Hixie> about 100ms of silence
  213. # [03:26] <othermaciej> in normal English speech, nothing, they are just separators
  214. # [03:26] <othermaciej> to a screen reader, I dunno
  215. # [03:26] <tantek> othermaciej, that's why presentational-major thinking = FAIL.
  216. # [03:27] <tantek> Hixie, I suggested ditching the "... of the document." clause from both the <ul> and <ol> sections.
  217. # [03:27] <Hixie> "the meaning" on its own just doesn't sound right
  218. # [03:27] <tantek> Hixie, "sounding right" is irrelevant. It's a spec. The question is, does it parse?
  219. # [03:27] <othermaciej> tantek: I don't see how your argument addresses presentational thinking - it might be an argument against ever using bulleted lists, but I wouldn't buy that conclusion
  220. # [03:27] <Hixie> i'm leaving "of the document" until I (or someone else) can think of something better to replace it with
  221. # [03:27] <othermaciej> anyway, let me give a practical example
  222. # [03:27] <tantek> Hixie, empty space.
  223. # [03:27] <Hixie> tantek: actually i consider it sounding right to be of paramount importance to me :-)
  224. # [03:27] <othermaciej> sometimes, on the webkit blog, I post lists of new features or bug fixes
  225. # [03:28] <othermaciej> usually, I use <ul> for this since they are in no particular order, and mentally speaking, bullets are appropriate
  226. # [03:28] <othermaciej> on one occasion, I made a point of the fact that these were 10 features of particular interest
  227. # [03:28] <othermaciej> and in that case I used <ol>
  228. # [03:28] <tantek> Hixie, if you prefer to place a redundant object resolution place that's ok too, and English depends on redundancy for error correction.
  229. # [03:28] <tantek> e.g. s/of the document/of the list of items
  230. # [03:28] <othermaciej> I think those were both correct choices but I'd be hard-pressed to argue there is a deep semantic difference
  231. # [03:29] <tantek> since that "list of items" is what the object resolution demonstrated above
  232. # [03:29] <tantek> that should address the "sounding right" concern.
  233. # [03:29] <Hixie> othermaciej: well, if you have any suggested improvement to the text of the spec as it was just changed, i'm certainly open to it
  234. # [03:30] <Hixie> tantek: i'm not convinced people would understand what it meant if i did that
  235. # [03:30] <Hixie> it's not just the list that would change meaning
  236. # [03:30] <Hixie> it's the list plus any content referring to that list
  237. # [03:30] <Hixie> e.g. see the examples
  238. # [03:30] <Hixie> especially for <ol>
  239. # [03:30] <othermaciej> to be honest I didn't read the spec, I just wanted to point out that something based on subtle distinctions of how important or essential an order is would not match the mental model of the typical author
  240. # [03:31] <Hixie> othermaciej: i'm not sure defining them as being "when you want bullets" or "when you want numbers" would work either, though
  241. # [03:32] <othermaciej> I think the distinction between <ol> and <ul> is not intrinsic to the data, it's about whether the author wants to specify and emphasize the order
  242. # [03:32] <othermaciej> depending on authorial intent, you could meaningfully do either for the same list in the same context
  243. # [03:32] * Quits: tantek (n=tantek@m730f36d0.tmodns.net)
  244. # [03:32] <Hixie> othermaciej: in your example of numbering them to emphasise the number of items, you are not emphasising the order.
  245. # [03:32] <othermaciej> for example, "top 10" lists are pretty arbitrary in their order, many could just as easily be "10 things" lists
  246. # [03:33] <Hixie> indeed
  247. # [03:33] <Hixie> so that's not emphasising the order
  248. # [03:33] <othermaciej> fair enough
  249. # [03:33] <othermaciej> although if the spec says that in that situation you should use numbers in the text content or <ul> with CSS counter styling, I am not sure it will be followed
  250. # [03:34] <othermaciej> (also not sure that it would ever be an improvement to do that)
  251. # [03:34] <Hixie> i would say that in those contexts the current definition works better
  252. # [03:34] <Hixie> the current definition is that the order matters
  253. # [03:34] <Hixie> and if you refer to "number 5", clearly the order matters
  254. # [03:34] <Hixie> since changing the order would change what each one was numbered
  255. # [03:34] <othermaciej> that's true
  256. # [03:34] <Hixie> <ol> is now defined as "The ol element represents a list of items, where the items have been intentionally ordered, such that changing the order would change the meaning of the document."
  257. # [03:34] <othermaciej> sometimes the numbers aren't there for order, but to provide referents for external references
  258. # [03:34] <Hixie> right
  259. # [03:35] <othermaciej> (which might not even be in the same document)
  260. # [03:35] <othermaciej> that seems like a good enough definition
  261. # [03:35] <Hixie> the current definition just says they've been intentionally ordered, not that they are ordered according to any grand scheme
  262. # [03:35] <Hixie> i like it
  263. # [03:35] <Hixie> yeah
  264. # [03:35] <Hixie> it has withstood your attempt at disproving it :-P
  265. # [03:35] * Quits: eseidel (n=eseidel@nat/google/x-7e4b89528cc9758c)
  266. # [03:36] <othermaciej> I wasn't trying to disprove it
  267. # [03:36] <othermaciej> I hadn't even read it at the time I commented
  268. # [03:36] <othermaciej> I was just reading scrollback and thought it sounded a bit abstract
  269. # [03:36] <Hixie> i know, i'm just teasing :-)
  270. # [03:38] <othermaciej> if I were trying to disprove something, you'd know it :-)
  271. # [03:38] <Hixie> hah
  272. # [04:03] * Quits: Lachy (n=Lachlan@cm-84.215.54.100.getinternet.no) ("This computer has gone to sleep")
  273. # [04:19] * Joins: webben (n=benh@dip5-fw.corp.ukl.yahoo.com)
  274. # [04:22] * Quits: webben_ (n=benh@dip5-fw.corp.ukl.yahoo.com) (Read error: 110 (Connection timed out))
  275. # [04:33] * Hixie adds more examples and their hidden meanings to the spec
  276. # [05:02] * Joins: tantek (n=tantek@c-67-161-32-176.hsd1.ca.comcast.net)
  277. # [05:10] <Hixie> hm
  278. # [05:10] <Hixie> <ol reversed> is an interesting proposal
  279. # [05:10] <Hixie> i wonder how the browser vendors would react to it
  280. # [05:19] * Quits: weinig_ (n=weinig@17.203.15.180)
  281. # [05:32] * Hixie bcc's 50-odd people
  282. # [05:33] <Hixie> ok that was a reply to 100+ e-mails
  283. # [05:33] <Hixie> yay
  284. # [05:34] <Hixie> what's next... maybe section/p/hr stuff... maybe em/strong stuff...
  285. # [05:34] <Hixie> hm
  286. # [05:35] <tantek> Hixie, rather than reversed, why not allow a list-increment?
  287. # [05:35] <tantek> default: +1
  288. # [05:36] <tantek> -1 to achieve reverse
  289. # [05:37] <Hixie> reversed does more than just change the step increment
  290. # [05:37] <Hixie> it also changes the start value
  291. # [05:37] <Hixie> we can add step later if there's a use case for it
  292. # [05:37] <tantek> sure but there already is a start attr
  293. # [05:37] <Hixie> right but reversed changes the default start value
  294. # [05:37] <tantek> the use case is lists that are numbered in large increments
  295. # [05:37] <tantek> see patent docs
  296. # [05:37] <tantek> which often use increments of 2 or 10
  297. # [05:37] <Hixie> they can use value=""
  298. # [05:37] <tantek> for paragraphs etc
  299. # [05:38] <Hixie> decrementing to 1 is a much more common case
  300. # [05:38] <tantek> they can't often use "value" because that requires too much fix-up
  301. # [05:38] <tantek> for the same reason we don't ask all numbered lists to use "value"
  302. # [05:38] <tantek> on every list item
  303. # [05:38] <Hixie> the whole point of large numbers in those cases is that you don't need fixup to add an additional paragraph
  304. # [05:39] <Hixie> hence why value="" is the most appropriate
  305. # [05:39] <Hixie> (if you really want to use a list in those cases)
  306. # [05:40] <tantek> people will just use hacks then to get auto-numbered by +10 lists
  307. # [05:40] <tantek> like <li>...</li> then 9 empty <li> etc.
  308. # [05:40] <Hixie> i haven't seen that happen
  309. # [05:40] <tantek> then CSS to only have every 10th item show up
  310. # [05:40] <Hixie> i have seen peple use hacks for stepping down to -1, though
  311. # [05:40] <Hixie> (namely using scripts to do it)
  312. # [05:41] <tantek> yeah, i've seen them, but can't find any examples offhand
  313. # [05:41] <Hixie> see the e-mail i just sent whatwg for the research that was done to back up reversed="" -- it wasn't as extensive as with other cases like some of the research i've done, but it was eye-opening nonetheless
  314. # [05:41] <tantek> if i do i'll let you know
  315. # [05:41] <Hixie> cool, thanks
  316. # [05:43] <Hixie> lord, microsoft adcenter is trying to hire me
  317. # [05:43] * Hixie informs them politely that he's happy working for the company that's whipping their asses already
  318. # [05:45] <tantek> Hixie, I do also remember running into the need to do reverse lists in the past as well, and having to resort to "value" manually to do it in my own markup.
  319. # [05:50] * Quits: dbaron (n=dbaron@corp-241.mountainview.mozilla.com) ("8403864 bytes have been tenured, next gc will be global.")
  320. # [05:53] <othermaciej> Hixie: they probably got you from a list of people working for the company that's whipping their asses
  321. # [06:00] <tantek> question: is this a valid URL for this channel: irc://irc.freenode.net/whatwg ?
  322. # [06:03] * Quits: roc (n=roc@202.0.36.64)
  323. # [06:12] <takkaria> tantek: I believe so
  324. # [06:13] <tantek> I wasn't able to use that "company that's whipping asses" to find an irc: URL validator.
  325. # [06:17] * Joins: dbaron (n=dbaron@c-67-160-251-228.hsd1.ca.comcast.net)
  326. # [06:22] <othermaciej> tantek: there's two different expired drafts floating around
  327. # [06:22] <othermaciej> tantek: by the latest one, it should be irc://irc.freenode.net/#whatwg
  328. # [06:23] <jruderman> without a # works in more clients
  329. # [06:23] <othermaciej> or irc://irc.freenode.net/whatwg,ischannel
  330. # [06:24] <othermaciej> (not sure if anything implements either of those rules
  331. # [06:24] <othermaciej> )
  332. # [06:30] * Quits: mpt (n=mpt@canonical/launchpad/mpt) (Connection timed out)
  333. # [06:38] * Joins: othermaciej_ (n=mjs@17.255.99.190)
  334. # [06:41] * Quits: othermaciej (n=mjs@nat/apple/x-2d3747264849a62c) (Read error: 110 (Connection timed out))
  335. # [06:57] * Quits: csarven (n=nevrasc@modemcable130.251-202-24.mc.videotron.ca) ("http://www.csarven.ca/")
  336. # [06:59] * Quits: othermaciej_ (n=mjs@17.255.99.190) (Read error: 110 (Connection timed out))
  337. # [07:05] * Joins: weinig (n=weinig@c-71-198-176-23.hsd1.ca.comcast.net)
  338. # [07:07] * Quits: tantek (n=tantek@c-67-161-32-176.hsd1.ca.comcast.net)
  339. # [07:13] * Quits: MacDome (n=eric@c-69-181-78-198.hsd1.ca.comcast.net)
  340. # [07:13] * Joins: MacDome (n=eric@c-69-181-78-198.hsd1.ca.comcast.net)
  341. # [07:25] * Quits: Hixie (i=ianh@trivini.no) (Read error: 104 (Connection reset by peer))
  342. # [07:25] * Joins: Hixie (i=ianh@trivini.no)
  343. # [07:52] * weinig is now known as weinig|sad
  344. # [07:53] * weinig|sad is now known as weinig
  345. # [08:03] * Joins: othermaciej (n=mjs@64.81.48.145)
  346. # [08:11] * Quits: dbaron (n=dbaron@c-67-160-251-228.hsd1.ca.comcast.net) ("8403864 bytes have been tenured, next gc will be global.")
  347. # [08:27] * Joins: tantek (n=tantek@adsl-63-195-114-133.dsl.snfc21.pacbell.net)
  348. # [08:35] <Hixie> ok here goes
  349. # [08:35] <Hixie> 63 elements on sections
  350. # [08:35] <Hixie> and related subjects
  351. # [08:35] * Hixie beings replying
  352. # [08:38] * Joins: roc (n=roc@121.72.172.117)
  353. # [08:40] <othermaciej> 63 elements?
  354. # [08:40] <othermaciej> or do you mean 63 emails?
  355. # [08:40] <othermaciej> hello roc
  356. # [08:41] <Hixie> er yes
  357. # [08:41] <Hixie> e-mails
  358. # [08:41] <roc> hello
  359. # [08:42] <jruderman> roc!
  360. # [08:51] <Hixie> hey roc
  361. # [08:59] * Joins: maikmerten (n=merten@ls5laptop14.cs.uni-dortmund.de)
  362. # [09:14] * Quits: jruderman (n=jruderma@guest-228.mountainview.mozilla.com)
  363. # [09:21] * Joins: virtuelv (n=virtuelv@pat-tdc.opera.com)
  364. # [09:23] * Joins: tndH_ (i=Rob@adsl-83-100-138-77.karoo.KCOM.COM)
  365. # [09:23] * tndH_ is now known as tndH
  366. # [09:24] * Quits: roc (n=roc@121.72.172.117)
  367. # [09:36] <virtuelv> hsivonen: yt?
  368. # [09:37] <hsivonen> virtuelv: yes
  369. # [09:37] <virtuelv> do you have any plans on releasing the source for your validator?
  370. # [09:38] <hsivonen> virtuelv: it's already Free Software with a publicly readable svn repo
  371. # [09:38] <hsivonen> virtuelv: I don't have immediate plans to cut version numbered release packages
  372. # [09:38] <virtuelv> hsivonen: Hm. I totally missed that
  373. # [09:39] <hsivonen> virtuelv: http://about.validator.nu/#src
  374. # [09:39] <virtuelv> hsivonen: thanks
  375. # [09:39] * Joins: jruderman (n=jruderma@c-67-180-15-227.hsd1.ca.comcast.net)
  376. # [09:40] <virtuelv> Would be helpful if there was a ToC on the about page
  377. # [09:42] * Joins: Lachy (n=Lachlan@ti200710a340-4094.bb.online.no)
  378. # [09:44] <hsivonen> virtuelv: ok. I'll add one in the next update of that page that I have in preparation
  379. # [10:04] <Dashiva> Dmitry wants to hold votes on his proposals, this could get messy
  380. # [10:05] <Lachy> Hixie, the top 10 movie list example contains movies that really shouldn't be in the top 10, and is missing others that should. :-)
  381. # [10:05] <Lachy> also, there's no example of using the start attribute given
  382. # [10:05] <Hixie> pah
  383. # [10:05] <Hixie> (re the movies)
  384. # [10:05] <Hixie> and yeah
  385. # [10:05] <Hixie> i need to add more examples
  386. # [10:06] <Hixie> i just added examples for the things people asked about
  387. # [10:07] <Dashiva> Hixie: Maybe you should take a pointer from FORTRAN and start labeling your step sequences, that way it's easy to say which sequence to abort :)
  388. # [10:08] <Lachy> you need an example that shows <ol start=100 reversed>, and explain why long reversed lists should ideally include a start attribute to help with incremental rendering
  389. # [10:08] * Joins: roc (n=roc@121.72.172.117)
  390. # [10:11] * Quits: webben (n=benh@dip5-fw.corp.ukl.yahoo.com)
  391. # [10:19] <Hixie> Dashiva: maybe
  392. # [10:19] <Hixie> Lachy: if you really think so, send mail :-)
  393. # [10:20] * Hixie is well into a big e-mail on sections by now
  394. # [10:26] * MacDome is now known as MacDomeSleep
  395. # [10:36] * Joins: Camaban (n=adrianle@host81-133-60-253.in-addr.btopenworld.com)
  396. # [10:38] <Lachy> Hixie, I will later
  397. # [10:39] <Hixie> k
  398. # [10:42] <hendry> what is the (code) name for the IE7 engine? e.g. Firefox -> Gecko
  399. # [10:43] <hsivonen> hendry: Trident
  400. # [10:43] <hsivonen> hendry: since IE 4.0
  401. # [10:45] <hendry> though aren't there more codenames as the engine in 6 is different to 7, if you know what i mean. i just re-read all that stuff that came up last month.
  402. # [10:45] <hendry> i should just re-read :)
  403. # [10:45] <Hixie> the code name for the IE engine is trident
  404. # [10:45] <Hixie> as far as know it hasn't got version-specific names
  405. # [10:46] <madmoose> Maybe for IE8 they'll rename it "Tridents".
  406. # [10:47] <Hixie> hah
  407. # [10:49] * jgraham_ wonders if they'll fulfil nominative-determinism and stop shipping new engines once they reach three
  408. # [10:50] <jgraham_> s/new/more/
  409. # [10:50] * Joins: ROBOd (n=robod@89.122.216.38)
  410. # [10:50] <Hixie> hm?
  411. # [10:50] <jgraham_> Trident == Three pronged, no?
  412. # [10:51] * Joins: webben (n=benh@nat/yahoo/x-9b424e4757a73b8a)
  413. # [10:53] <Hixie> oh, that name
  414. # [10:53] <Hixie> i thought you meant IE8
  415. # [10:53] <Hixie> and couldn't work out how 8 would mean 3
  416. # [11:01] * Quits: Lachy (n=Lachlan@ti200710a340-4094.bb.online.no) ("This computer has gone to sleep")
  417. # [11:03] <annevk> DOM travelsal is complicated :(
  418. # [11:03] <Hixie> yeah
  419. # [11:03] <annevk> traversal, even
  420. # [11:03] <Hixie> you wanna fix it?
  421. # [11:03] * Joins: mpt (n=mpt@canonical/launchpad/mpt)
  422. # [11:03] <annevk> heh
  423. # [11:04] <Hixie> it needs an editor
  424. # [11:04] <Hixie> with a big hammer and a bag full of nails
  425. # [11:04] <Hixie> and a lot of duct tape
  426. # [11:04] <annevk> i've got quite a list of specs already
  427. # [11:05] <annevk> can you tell if http://tc.labs.opera.com/dom/traversal/002.htm is valid or not?
  428. # [11:06] <annevk> opera had some minor failures, but i'm not sure if my minimized test is ok
  429. # [11:06] <Hixie> i dunno
  430. # [11:06] <Hixie> it took me days to do the acid3 test that does mutations in the handler
  431. # [11:06] <Hixie> and i'm still not sure it's correct
  432. # [11:06] <annevk> ouch
  433. # [11:06] <annevk> that's the test i'm talking about :)
  434. # [11:06] <Hixie> :-)
  435. # [11:07] <Hixie> someone should write up the traversal spec unambiguously
  436. # [11:07] <Hixie> then we'd know
  437. # [11:09] <jgraham_> Hixie: You should produce a list of specs that you _don't_ think need an editor, as it would be simpler to refer to that than a list of specs that do
  438. # [11:09] <jgraham_> ;)
  439. # [11:10] * Quits: MacDomeSleep (n=eric@c-69-181-78-198.hsd1.ca.comcast.net)
  440. # [11:12] <annevk> how did the moz call go?
  441. # [11:12] <annevk> cookies?
  442. # [11:15] * Joins: MacDome (n=eric@c-69-181-78-198.hsd1.ca.comcast.net)
  443. # [11:15] <Hixie> jgraham_: no, i have a list of specs that need an editor
  444. # [11:15] <Hixie> jgraham_: it's relatively short
  445. # [11:16] <Hixie> jgraham_: (companion specifications on the whatwg wiki)
  446. # [11:16] <jgraham_> Hixie: I wasn't quite being serious :)
  447. # [11:16] <Hixie> :-)
  448. # [11:18] * Joins: zcorpan (n=zcorpan@pat.se.opera.com)
  449. # [11:20] * Philip` wondered why his web server had TCP connections with apache2-hixie.hixie.dreamhost.com, before finding that that's just what whatwg.org reverse-resolves into
  450. # [11:21] * Joins: Lachy (n=Lachlan@pat-tdc.opera.com)
  451. # [11:22] <Philip`> By the way, http://status.whatwg.org/annotate-web-apps.php doesn't look so good
  452. # [11:25] * Philip` wonders if people who carefully make XHTML sites in PHP do something to disable/redirect warning/error messages like that, or if they just assume they're never going to occur
  453. # [11:26] * Quits: Lachy (n=Lachlan@pat-tdc.opera.com) (Client Quit)
  454. # [11:26] * Joins: Lachy (n=Lachlan@pat-tdc.opera.com)
  455. # [11:30] <zcorpan> Philip`: what? php error messages are supposed to be xhtml compatible!! they changed <br> to <br />!1
  456. # [11:32] * Quits: MacDome (n=eric@c-69-181-78-198.hsd1.ca.comcast.net)
  457. # [11:43] <Lachy> Hixie, yt?
  458. # [11:44] <Lachy> Hixie, in the first note in the dl elmement section, you're missing the word "using"
  459. # [11:45] * Quits: roc (n=roc@121.72.172.117)
  460. # [11:45] <Lachy> oh, actually, you wrote "using accidentally" instead of "accidentally using"
  461. # [12:04] <Hixie> thanks will fix
  462. # [12:06] * Joins: OZI^ (i=soonov@121.152.249.32)
  463. # [12:06] * Quits: OZI^ (i=soonov@121.152.249.32) (Client Quit)
  464. # [12:06] * Joins: OZI^ (i=soonov@121.152.249.32)
  465. # [12:06] * Quits: OZI^ (i=soonov@121.152.249.32) (Client Quit)
  466. # [12:14] <Lachy> Hixie, why do you refer to <dl> as an association list where it's defined, but then refer to description list everywhere else?
  467. # [12:19] <Hixie> carelessness, probably
  468. # [12:19] <Hixie> aw man, henri is asking for a rewrite of the outline algorithm
  469. # [12:19] <Hixie> and there i was thinking i'd get through all this feedback without having to do any work
  470. # [12:20] <Hixie> i guess i'll reply to this tomorrow
  471. # [12:20] <annevk> that must be a subset of his feedback then
  472. # [12:21] <annevk> lots of parsing questions stuff prolly requires work
  473. # [12:21] <annevk> s/stuff//
  474. # [12:22] <zcorpan> annevk: it seems you read "this" as "his"
  475. # [12:22] <annevk> yeah
  476. # [12:29] * Joins: hasather (n=hasather@90-231-107-133-no62.tbcn.telia.com)
  477. # [12:39] <zcorpan> the first example on <figure> reminds me of <listing>
  478. # [12:48] * zcorpan notes that the 3rd example has <p/><img/><p/>
  479. # [12:49] <annevk> that reminds me of a frontpage commercial
  480. # [12:49] <annevk> http://annevankesteren.nl/2004/02/microsoft-frontpage-and-valid-html
  481. # [12:50] <annevk> "</p>That's right. We said FrontPage.</p>"
  482. # [12:51] * Joins: myakura (n=myakura@p1215-ipbf3008marunouchi.tokyo.ocn.ne.jp)
  483. # [12:57] * Quits: maikmerten (n=merten@ls5laptop14.cs.uni-dortmund.de) ("Verlassend")
  484. # [13:05] * Quits: webben (n=benh@nat/yahoo/x-9b424e4757a73b8a)
  485. # [13:20] <zcorpan> Hixie: you got a 9 too much, so it would be 5000 people, not 500 :)
  486. # [13:23] <Philip`> <pre><code>
  487. # [13:23] <Philip`> interface PrimaryCore {
  488. # [13:23] <Philip`> seems sort of wrong, since it'll have a blank line at the top
  489. # [13:23] <Philip`> (since <pre> just deletes a directly following newline character token, which doesn't happen in this case)
  490. # [13:24] <Philip`> Also it says void sendData(in sequence&lt;byte> data); which looks wrong
  491. # [13:25] <zcorpan> Philip`: what wrong about it?
  492. # [13:25] <Philip`> Oh, whoops, nothing
  493. # [13:25] <Philip`> I forgot it was showing HTML code
  494. # [13:28] <annevk> the <aside> example that's also a self-reference contains a too early </pre>
  495. # [13:30] * Quits: mpt (n=mpt@canonical/launchpad/mpt) (Read error: 110 (Connection timed out))
  496. # [13:49] * Quits: MikeSmith (n=MikeSmit@58.157.21.205) ("Less talk, more pimp walk.")
  497. # [13:58] * Quits: Dashiva (i=Dashiva@wikia/Dashiva) (calvino.freenode.net irc.freenode.net)
  498. # [13:59] * Joins: webben (n=benh@nat/yahoo/x-c84e3b38173f2bb0)
  499. # [13:59] * Joins: Dashiva (i=Dashiva@wikia/Dashiva)
  500. # [14:15] * Joins: MikeSmith (n=MikeSmit@58.157.21.205)
  501. # [14:27] * Joins: Lachy_ (n=Lachlan@pat-tdc.opera.com)
  502. # [14:27] * Quits: Lachy (n=Lachlan@pat-tdc.opera.com) (Read error: 104 (Connection reset by peer))
  503. # [14:28] * Philip` wonders if there's a web server anywhere which sends Content-Encoding: deflate, so he can test his code on it
  504. # [14:38] * Joins: maikmerten (n=merten@ls5laptop14.cs.uni-dortmund.de)
  505. # [14:38] <hsivonen> Philip`: curiously, x-gzip (as opposed to gzip) happens
  506. # [14:40] * Quits: tantek (n=tantek@adsl-63-195-114-133.dsl.snfc21.pacbell.net)
  507. # [14:45] <hsivonen> Philip`: when I implemented gzip support for V.nu, I figured that clients support gzip so well that supporting deflate and x-gzip is not worthwhile.
  508. # [14:45] <hsivonen> Philip`: it seems to me that deflate is mostly a dead letter in the spec
  509. # [14:46] <hsivonen> Philip`: so my guess is that your software doesn't need to support deflate
  510. # [14:59] * Joins: SadEagle (n=maksim@cpe-69-202-89-106.twcny.res.rr.com)
  511. # [15:13] <Philip`> Hmm, I suppose I might as well keep untested deflate support since it's only two lines of code and it'll let me see how many servers attempt to send deflate responses
  512. # [15:14] <Philip`> It looks like only Opera sends accept-encoding:x-gzip
  513. # [15:15] <Philip`> (and nothing sends 'compress')
  514. # [15:16] <SadEagle> konq does x-gzip as well. But no compress.
  515. # [15:16] * Philip` adds x-gzip to see what happens
  516. # [15:17] <Philip`> SadEagle: Aha, okay
  517. # [15:17] <Philip`> Konqueror 3.5.something sends "x-gzip, x-deflate, gzip, deflate"
  518. # [15:18] <Philip`> but x-deflate seems a bit peculiar since it e.g. isn't mentioned in RFC2616
  519. # [15:20] * Philip` wonders how many people handle accept-encoding:gzip,identity;q=0
  520. # [15:34] <annevk> If I have to boxes A and B. A is located at 0,0 and B at 10,10. I can now speak about the distance between the top edge of A and the top edge of B (being 10). Now if B is located at -10,-10 this distance should be -10. Is saying "The distance between the top edge of A and the top edge B" enough or should I say "downward distance"
  521. # [15:34] <annevk> suggestions?
  522. # [15:36] <zcorpan> if it can be negative then it's not a distance
  523. # [15:36] <Philip`> "The y coordinate of the top edge of A minus the y coordinate of the top edge of B"?
  524. # [15:36] <annevk> i guess that's best, yes
  525. # [15:37] <Philip`> Or skip the annoyingly verbose English and say A<sub>T</sub><sub>y</sub> - B<sub>T</sub><sub>y</sub>
  526. # [15:38] <Philip`> Actually, skip the annoyingly verbose HTML too and say $A_T_y - B_T_y$
  527. # [15:38] <annevk> heh
  528. # [15:38] * Quits: myakura (n=myakura@p1215-ipbf3008marunouchi.tokyo.ocn.ne.jp) ("Leaving...")
  529. # [15:39] <zcorpan> Hixie: it should be attribute boolean reversed, not attribute long reversed
  530. # [15:40] <zcorpan> Hixie: s/is present/is absent/ in the paragraph defining reversed=''
  531. # [15:57] * Joins: aroben (n=aroben@unaffiliated/aroben)
  532. # [15:58] <Philip`> Oh, it turns out there's at least one site which sends 'delate' content, and it makes my code die with "unknown compression method"
  533. # [16:00] * Quits: MikeSmith (n=MikeSmit@58.157.21.205) ("Less talk, more pimp walk.")
  534. # [16:01] <zcorpan> 'delate'?
  535. # [16:01] <Philip`> Uh
  536. # [16:02] <Philip`> 'deflate'
  537. # [16:04] <Philip`> Aha, it works when I add a second parameter and do new InflaterInputStream(in, new Inflater(true));
  538. # [16:08] * Joins: MikeSmith (n=MikeSmit@58.157.21.205)
  539. # [16:12] <Philip`> In 1024 pages, I see 2 deflate and 141 gzip
  540. # [16:12] <Philip`> (when sending accept-encoding:gzip,deflate)
  541. # [16:25] <Philip`> In 8192, I see 12 deflate and 1195 gzip
  542. # [16:29] <Philip`> In 16384, 16 and 2439
  543. # [16:29] <Philip`> deflate is not entirely negligible
  544. # [16:31] * Parts: zcorpan (n=zcorpan@pat.se.opera.com)
  545. # [16:31] * Joins: zcorpan (n=zcorpan@pat.se.opera.com)
  546. # [16:31] * Quits: annevk (n=annevk@77.163.243.203) (Remote closed the connection)
  547. # [16:32] * Joins: phsiao (n=shawn@nat/ibm/x-e347e3db4e6bc714)
  548. # [16:32] * Joins: annevk (n=annevk@77.163.243.203)
  549. # [16:34] * Parts: zcorpan (n=zcorpan@pat.se.opera.com)
  550. # [16:34] <Philip`> http://www.toua-u.ac.jp/ - "Video/X-Flv: .flv" - is that meant to do anything?
  551. # [16:34] * Joins: zcorpan (n=zcorpan@pat.se.opera.com)
  552. # [16:38] <Philip`> http://www.superexpressgonzalez.com/ - P3P: policyref="http://www.tiendavirtual.ws/w3c/p3p.xml" CP="NOI DSP COR NID PUB NOR" """"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""
  553. # [16:56] * Quits: maikmerten (n=merten@ls5laptop14.cs.uni-dortmund.de) ("Verlassend")
  554. # [17:00] <Philip`> <!doctype php public "-//w3c//dtd xphp 1.0 transitional//en" "http://www.w3.org/tr/xphp1/dtd/xphp1-transitional.dtd">
  555. # [17:01] <Philip`> <!doctype html public "guest house,jersey,haven,guest house,accommodation,jersey,channel islands,saint helier,british ">
  556. # [17:01] <Philip`> <!doctype html public "-//w3c//dtd xhtml 1.0 transitional//en" "http://www.w3.org/tr/xhtml1/dtd/xhtml1-transitional.dtd" "http://www.w3.org/tr/html4/loose.dtd">
  557. # [17:01] <Philip`> The world is a crazy place :-(
  558. # [17:02] <annevk> fortunately we have unambigious rules to interpret it
  559. # [17:03] <zcorpan> i've seen "php" doctypes before
  560. # [17:03] <Philip`> http://www.thermaglaze.com/ - <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd" [ <!ATTLIST a target CDATA #IMPLIED> ]> - that's another way to work around the nonconformance of <a target>
  561. # [17:03] <zcorpan> people change all their files from .html to .php and do a global s/html/php/
  562. # [17:04] <zcorpan> (which suggests there are some <php> elements out there, too)
  563. # [17:04] <Philip`> http://www.pervasive.com/ - <!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 5.0//EN" />
  564. # [17:05] <Philip`> <!doctype html public "&#150;//w3c//dtd html 3.2//en"> - hmm, bad copy-and-paste
  565. # [17:05] <Philip`> (from Word, presumably)
  566. # [17:06] <zcorpan> or search-replace
  567. # [17:06] <Philip`> Oh, I suppose that's possible
  568. # [17:08] * Quits: Lachy_ (n=Lachlan@pat-tdc.opera.com) ("This computer has gone to sleep")
  569. # [17:21] <Philip`> Hmm, IE6 and HTML5 (or non-IE browsers) seem to differ in doctype moding for ~1% of pages
  570. # [17:21] * Joins: Lachy (n=Lachlan@ti200710a340-4094.bb.online.no)
  571. # [17:23] <Philip`> The most significant difference is <!doctype html public "-//w3c//dtd html 4.0 transitional//en" "http://www.w3.org/tr/rec-html40/loose.dtd"> (on ~0.5% of pages)
  572. # [17:25] <zcorpan> Philip`: do you have pointers to the remaining ~0.5%?
  573. # [17:31] <Philip`> 276 – <!doctype html public "-//w3c//dtd html 4.0 transitional//en" "http://www.w3.org/tr/rec-html40/loose.dtd">
  574. # [17:32] <Philip`> 69 – <!doctype html public "-//"aol hometown//html 3.0 transitional//en">
  575. # [17:32] <Philip`> 45 – <!doctype html public "-//w3c//dtd html 4.0 transitional//en" "http://www.w3.org/tr/html4/loose.dtd">
  576. # [17:32] <Philip`> 34 – <!doctype html public "-//w3c//dtd xhtml 1.0 transitional//en" system "http://www.w3.org/tr/xhtml1/dtd/xhtml1-transitional.dtd"/>
  577. # [17:32] <Philip`> 22 – <!doctype html public "-//w3c//dtd html 4.0 transitional//en" "http://www.w3.org/tr/html40/loose.dtd">
  578. # [17:32] <Philip`> 19 – <!doctype html public "-//w3c//dtd html 4.0 frameset//en" "http://www.w3.org/tr/rec-html40/frameset.dtd">
  579. # [17:32] <Philip`> zcorpan: ^ Those are the counts (out of 62592 pages) of the top doctypes that are IE standards and HTML5 quirks
  580. # [17:32] * Philip` will upload some page with all the details later
  581. # [17:32] <zcorpan> Philip`: thanks!
  582. # [17:33] <Philip`> (These are all pages from dmoz.org, so there's still an uncorrected bias towards certain sites/domains)
  583. # [17:34] <Philip`> (Also, by "HTML5" I mean HTML5 plus http://lists.w3.org/Archives/Public/public-html/2008Jan/0006.html )
  584. # [17:42] * Philip` wonders how many pages he should download
  585. # [17:47] <gsnedders> how many implementations are there of RDF?
  586. # [17:48] * Joins: cgriego (n=cgriego@216.138.69.206)
  587. # [17:49] <Philip`> I get quite a few "Corrupt GZIP trailer" errors
  588. # [17:53] <Philip`> <button onlick="..."> doesn't quite do what I want
  589. # [17:56] * Quits: Camaban (n=adrianle@host81-133-60-253.in-addr.btopenworld.com) (Read error: 104 (Connection reset by peer))
  590. # [17:56] * Joins: Camaban (n=adrianle@host81-133-60-253.in-addr.btopenworld.com)
  591. # [18:11] <Philip`> <!doctype html public "-//w3c//dtd xhtml 1.0 transitional//en" "http://www.w3.org/tr/xhtml1/dtd/xhtml1-transitional.dtd" />
  592. # [18:12] <Philip`> HTML5 say quirks, Firefox/Opera say limited-quirks
  593. # [18:12] <Philip`> *says
  594. # [18:13] <Philip`> (IE says limited-quirks too, and I can't be bothered to test Safari)
  595. # [18:18] <zcorpan> um... how come safari doesn't do quirks mode in the live dom viewer?
  596. # [18:19] <Philip`> How are you testing that?
  597. # [18:19] <zcorpan> <style>body{background:f00}</style>
  598. # [18:21] <zcorpan> safari seems to not care about garbage after the system identifier
  599. # [18:22] <zcorpan> so html5 is in disagreement with browsers about trailing /> in the doctype
  600. # [18:22] <Philip`> Safari seems to think "foo<!doctype html>" is non-quirks
  601. # [18:22] <zcorpan> mozilla too
  602. # [18:23] <aroben> zcorpan: Philip`: what version of Safari are you testing?
  603. # [18:23] <zcorpan> probably an old one
  604. # [18:24] <zcorpan> 3.0.2 (522.13.1)
  605. # [18:24] <aroben> ok
  606. # [18:24] <aroben> that's good
  607. # [18:24] <aroben> WebKit trunk just recently changed to match HTML 5's doctype parsing much more closely
  608. # [18:24] <aroben> I'm glad you're testing the behavior of Safari 3
  609. # [18:24] <Philip`> I see 88 pages with <!doctype .../> out of 109434
  610. # [18:24] <zcorpan> aroben: would be interesting to know if your changes breaks anything
  611. # [18:26] <aroben> zcorpan: yes, it would ;-)
  612. # [18:26] <aroben> zcorpan: I'm sure we'll find out eventually
  613. # [18:26] <zcorpan> i would also like to know if pages that use something other than //en in the FPI are rendered correctly in firefox/safari
  614. # [18:27] * gsnedders is confused. how is the billion laughs attack that many? if 0 is 2 chars, then is repeated twice, it becomes four, then that is repeated twice you get eight, etc. only doubling.
  615. # [18:28] <zcorpan> gsnedders: double it enough times and it will become a billion pretty quickly :)
  616. # [18:28] <gsnedders> but with only 30 it gets to only 120, I thinks
  617. # [18:28] <gsnedders> that can't be right
  618. # [18:28] * gsnedders is being dumb
  619. # [18:30] <gsnedders> doubling starting at two goes up in powers of two.
  620. # [18:30] <gsnedders> duh.
  621. # [18:31] <gsnedders> which means laugh30 is $2^{30+1}$ characters
  622. # [18:32] <Philip`> I see 8 pages with <!doctype html> out of 109434
  623. # [18:38] * Quits: virtuelv (n=virtuelv@pat-tdc.opera.com) (Read error: 110 (Connection timed out))
  624. # [18:39] <annevk> mine mine mine
  625. # [18:42] <Philip`> 135 pages have Server "ZX_Spectrum/1997 (Sinclair_BASIC)"
  626. # [18:43] <SadEagle> coolness.
  627. # [18:45] * aroben is now known as aroben|lunch
  628. # [18:47] <Philip`> Oh, they're all subdomains of narod.ru
  629. # [18:47] <Philip`> so there's only one web server running on a Spectrum :-(
  630. # [18:52] * Joins: maikmerten (n=maikmert@L82e3.l.pppool.de)
  631. # [18:52] * Quits: webben (n=benh@nat/yahoo/x-c84e3b38173f2bb0) (Read error: 110 (Connection timed out))
  632. # [18:57] * Joins: svl (n=me@ip565744a7.direct-adsl.nl)
  633. # [19:02] * aroben|lunch is now known as aroben
  634. # [19:04] * Quits: weinig (n=weinig@c-71-198-176-23.hsd1.ca.comcast.net)
  635. # [19:06] <Philip`> http://philip.html5.org/data/doctypes-2.html
  636. # [19:06] <Philip`> (The "Test" thing probably doesn't work in WebKit)
  637. # [19:08] <gsnedders> yeah, it just reportd non-quirks for all
  638. # [19:09] <gsnedders> Philip`: <http://bugs.webkit.org/show_bug.cgi?id=15062>
  639. # [19:09] * Parts: Camaban (n=adrianle@host81-133-60-253.in-addr.btopenworld.com)
  640. # [19:22] <hsivonen> hrm. Google docs reject Firefox 3 and tells me to get Firefox 1.5.0.12 or newer
  641. # [19:23] <svl> Google does lots of bad sniffing for "Firefox" instead of looking at "Gecko" :(
  642. # [19:24] * Joins: jwalden (n=waldo@STRATTON-TWO-EIGHTY-SEVEN.MIT.EDU)
  643. # [19:25] <svl> (And Firefox 3 at present is still "minefield" (I'm pretty certain, though didn't check))
  644. # [19:27] <aroben> Philip`: here's a test for what mode the page is in that works in WebKit: http://trac.webkit.org/projects/webkit/browser/trunk/LayoutTests/fast/doctypes/resources/TestDoctype.js?rev=30431
  645. # [19:28] * Quits: phsiao (n=shawn@nat/ibm/x-e347e3db4e6bc714) (Read error: 110 (Connection timed out))
  646. # [19:34] * Joins: MacDome (n=eric@c-69-181-78-198.hsd1.ca.comcast.net)
  647. # [19:37] <gavin> svl: nightly builds are "Minefield"
  648. # [19:37] <gavin> betas are "Firefox"
  649. # [19:37] <hsivonen> I tried with a beta, fwiw
  650. # [19:39] * Joins: jgraham (n=james@81-86-215-67.dsl.pipex.com)
  651. # [19:39] * Quits: jruderman (n=jruderma@c-67-180-15-227.hsd1.ca.comcast.net)
  652. # [19:39] <gsnedders> svl: Firefox is a registered trademark of the Mozilla Foundation, and its use is massively limited
  653. # [19:40] <gavin> "massively limited"? limited to "Firefox", sure :)
  654. # [19:43] * Joins: Lachy_ (n=Lachlan@cm-84.215.54.100.getinternet.no)
  655. # [19:45] <gsnedders> gavin: limited to any official build
  656. # [19:45] <zcorpan> Philip`: i think the regular expression could have been <!doctype[^>]*>
  657. # [19:47] <gsnedders> Philip`: seems to have been merged into trunk by now
  658. # [19:48] <zcorpan> <!doctype html public "-//sq//dtd html 2.0 + all extensions//en" "hmpro3.dtd">
  659. # [19:48] <zcorpan> hmm
  660. # [19:50] * Quits: cgriego (n=cgriego@216.138.69.206) (Read error: 113 (No route to host))
  661. # [19:51] <gsnedders> Philip`: only diff in WebKit from HTML 5 is <!doctype html public "-//w3c//dtd html 4.01 transitional//en" ""> is quirks
  662. # [19:51] * Quits: Lachy (n=Lachlan@ti200710a340-4094.bb.online.no) (Read error: 110 (Connection timed out))
  663. # [19:54] <zcorpan> opera seems to mostly match html5 on commonly used doctypes, but mostly match ie on uncommonly used doctypes
  664. # [19:55] <zcorpan> what i want to know is if changing our impl will break pages with the uncommonly used doctypes :)
  665. # [19:56] <zcorpan> if they already are broken in safari 3 and firefox, then it might be a good idea to revise html5
  666. # [19:57] <zcorpan> (i.e. broken because they get the wrong mode)
  667. # [20:05] * Joins: dbaron (n=dbaron@guest-228.mountainview.mozilla.com)
  668. # [20:06] * Joins: jruderman (n=jruderma@guest-228.mountainview.mozilla.com)
  669. # [20:08] * MacDome is now known as MacDomeOut
  670. # [20:08] * Joins: roc (n=roc@121.72.172.117)
  671. # [20:28] * Joins: cgriego (n=cgriego@216.138.69.206)
  672. # [20:32] * Quits: Lachy_ (n=Lachlan@cm-84.215.54.100.getinternet.no) (Read error: 104 (Connection reset by peer))
  673. # [20:33] * Joins: Lachy (n=Lachlan@cm-84.215.54.100.getinternet.no)
  674. # [20:36] * Quits: roc (n=roc@121.72.172.117)
  675. # [20:36] <Philip`> zcorpan: <!doctype[^>]*> probably would be entirely sensible, given that that's what HTML5 now expects
  676. # [20:37] * Philip` will re-run it with that
  677. # [20:39] * Philip` saw 39 application/xhtml+xml pages out of 100K (when sending the same Accept header as FF3)
  678. # [20:45] * Quits: jruderman (n=jruderma@guest-228.mountainview.mozilla.com) (Remote closed the connection)
  679. # [20:46] <Philip`> gsnedders: Do you mean the 'Test' button thing works correctly in WebKit now, so I don't need to change anything (like copying aroben's suggestion)?
  680. # [20:46] <gsnedders> Philip`: yeah
  681. # [20:46] <Philip`> gsnedders: Okay, thanks, that makes it easy for me :-)
  682. # [20:47] <annevk> http://tc.labs.opera.com/html/parsing/doctype/001.htm also tests DOCTYPE handling
  683. # [20:47] <annevk> fwiw
  684. # [20:51] * Philip` wonders if he should make each doctype link to a list of all the pages which were found to use it
  685. # [20:54] * Quits: zcorpan (n=zcorpan@pat.se.opera.com) (Read error: 110 (Connection timed out))
  686. # [20:55] * Quits: jgraham (n=james@81-86-215-67.dsl.pipex.com) ("I get eaten by the worms")
  687. # [21:21] * Joins: jruderman (n=jruderma@guest-228.mountainview.mozilla.com)
  688. # [21:21] * Joins: roc (n=roc@202.0.36.64)
  689. # [21:25] * Quits: dbaron (n=dbaron@guest-228.mountainview.mozilla.com) ("8403864 bytes have been tenured, next gc will be global.")
  690. # [21:26] * Joins: dbaron (n=dbaron@corp-241.mountainview.mozilla.com)
  691. # [21:29] * Joins: jgraham (n=james@81-86-215-67.dsl.pipex.com)
  692. # [21:29] <Hixie> Philip`: how is <!doctype html public "-//"aol hometown//html 3.0 transitional//en"> triggering HTML5 quirks?
  693. # [21:32] <Philip`> Hixie: After the second '"' it's in 'after DOCTYPE public identifier state', and the 'a' puts it into 'bogus DOCTYPE state', which ends up setting the correctness flag to 'incorrect', so it gets treated as quirks
  694. # [21:33] * Quits: Lachy (n=Lachlan@cm-84.215.54.100.getinternet.no) ("This computer has gone to sleep")
  695. # [21:35] * Quits: syp_ (n=syp@lasigpc9.epfl.ch) (Read error: 104 (Connection reset by peer))
  696. # [21:36] <Hixie> Philip`: oh, yeah, good point
  697. # [21:36] * Joins: syp_ (n=syp@lasigpc9.epfl.ch)
  698. # [21:38] <Philip`> Oh, I just realised why my downloader was failing to terminate for one particular site
  699. # [21:38] <Philip`> That site is an infinitely long radio stream
  700. # [21:39] <Hixie> hah
  701. # [21:39] <Hixie> yeah
  702. # [21:39] <Hixie> i've come across infinite pages before too
  703. # [21:39] <Hixie> that's how i know there's an infinite number of <font> elements on the web
  704. # [21:40] <gsnedders> :D
  705. # [21:42] * Quits: ROBOd (n=robod@89.122.216.38) ("http://www.robodesign.ro")
  706. # [21:47] * Quits: maikmerten (n=maikmert@L82e3.l.pppool.de) (Remote closed the connection)
  707. # [21:52] <Philip`> Why has my /bin/grep become thousands of times slower than it should be?
  708. # [21:55] <Philip`> and only on my XML file of extracted doctypes, not on any other file I've tested?
  709. # [21:56] <Philip`> Oh, it goes thousands of times faster again if I set LANG=en_GB without the .UTF-8
  710. # [21:58] * Joins: weinig (n=weinig@17.203.15.180)
  711. # [21:58] <Philip`> even though it's perfectly fast on a different computer with LANG=en_GB.UTF-8
  712. # [21:59] <gsnedders> computers are odd.
  713. # [22:03] <Dashiva> Probably has to do with matching stuff like \w
  714. # [22:05] <Philip`> I was only doing "grep processed foo.xml"
  715. # [22:08] <Dashiva> Maybe it does some conversion to an internal representation or some other premature optimization anyway?
  716. # [22:09] <Philip`> <!DOCTYPE "="&quot;" PUBLIC="PUBLIC" HTML="html" HTTP://WWW.W3.ORG/TR/XHTML1/DTD/XHTML1-STRICT.DTD="http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd" 1.0="1.0" -//W3C//DTD="-//W3C//DTD" STRICT//EN="Strict//EN"
  717. # [22:09] <Philip`> ="
  718. # [22:09] <Philip`> " XHTML="XHTML">
  719. # [22:09] <Philip`> Hmm, I can guess how that happened
  720. # [22:11] <Dashiva> haha
  721. # [22:12] <Hixie> wow, what a mess
  722. # [22:12] <Dashiva> Whitespace is our friend, right?
  723. # [22:13] * Joins: eseidel (n=eseidel@nat/google/x-e7a64af12e4d126a)
  724. # [22:14] * Quits: jwalden (n=waldo@STRATTON-TWO-EIGHTY-SEVEN.MIT.EDU) (Remote closed the connection)
  725. # [22:21] <Hixie> hsivonen: re your request for rewriting the outlining algorithm
  726. # [22:21] <Hixie> you say
  727. # [22:21] <Hixie> > This sweep should probably define what checkpoint data to
  728. # [22:21] <Hixie> > store on each element that implements the HTMLHeaderElement interface to allow
  729. # [22:21] <Hixie> > local recomputation.
  730. # [22:22] <Hixie> can't i just say what i usually say, namely that at any time, the outline must match the results you would get from running the algorithm over the whole tree?
  731. # [22:22] <Hixie> i don't want to be the one defining how you are to do optimisations...
  732. # [22:23] <hsivonen> Hixie: does stating that actually make the said optimization reasonably implementable?
  733. # [22:23] <Hixie> depends what the algorithm is
  734. # [22:24] <Philip`> http://philip.html5.org/data/doctypes-2.html#%3C!doctype_html%3E - now with links to pages, at least for the obscurer doctypes
  735. # [22:25] <hsivonen> Hixie: I don't see how it would hurt to define the outline in terms of forward sweeps from a hn state
  736. # [22:25] <Hixie> oh it wouldn't, and that's just a subset of defining it as part of a forward sweep in general
  737. # [22:25] <Hixie> i'm just saying the spec shouldn't define ny algorithm twice
  738. # [22:25] <Hixie> since the spec shouldn't be optimising
  739. # [22:26] <Hixie> or rather, it should be optimising for lack of ambiguity
  740. # [22:26] <Hixie> and adding two algorithms defeats that :-)
  741. # [22:26] <gsnedders> the current section confuses me
  742. # [22:26] <Hixie> hmm... i guess each sectioning element is effectively atomic in the algorithm...
  743. # [22:26] <Hixie> that is, anything outside a section can't affect what's inside it
  744. # [22:26] <Hixie> and what's inside a section can only affect the relative position of further nested sections...
  745. # [22:27] <Hixie> that's useful
  746. # [22:27] <hsivonen> Hixie: surely a forward sweep algorith can be the only definition
  747. # [22:27] <Hixie> hsivonen: yes
  748. # [22:27] <Hixie> hsivonen: that's what i plan to do
  749. # [22:27] <hsivonen> Hixie: excellent. thanks
  750. # [22:28] <Hixie> but your e-mail asked for one or two other things too, which i'm not sure i should necessarily do
  751. # [22:28] <Hixie> actually i think what i quoted above is a non-issue, and your second requirement might already be met
  752. # [22:34] * aroben is now known as aroben|meeting
  753. # [22:37] * Quits: othermaciej (n=mjs@64.81.48.145)
  754. # [22:38] * Joins: phsiao (n=shawn@146.195.217.216.transedge.com)
  755. # [22:39] * Quits: phsiao (n=shawn@146.195.217.216.transedge.com) (Client Quit)
  756. # [22:39] * Joins: phsiao (n=shawn@146.195.217.216.transedge.com)
  757. # [22:54] * Quits: gsnedders (n=gsnedder@host86-151-228-75.range86-151.btcentralplus.com) ("Partying in teh intarwebs")
  758. # [23:16] * Joins: kingryan (n=ryan@dsl092-219-050.sfo1.dsl.speakeasy.net)
  759. # [23:29] * Joins: Lachy (n=Lachlan@cm-84.215.54.100.getinternet.no)
  760. # [23:33] * Quits: jgraham (n=james@81-86-215-67.dsl.pipex.com) ("I get eaten by the worms")
  761. # [23:36] * Joins: othermaciej (n=mjs@nat/apple/x-5b1ccb77073d5c7e)
  762. # [23:37] * Quits: Lachy (n=Lachlan@cm-84.215.54.100.getinternet.no) ("Leaving")
  763. # [23:37] * Joins: Lachy (n=Lachlan@cm-84.215.54.100.getinternet.no)
  764. # Session Close: Wed Feb 27 00:00:00 2008

The end :)