/irc-logs / w3c / #css / 2011-09-20 / end

Options:

  1. # Session Start: Tue Sep 20 00:00:00 2011
  2. # Session Ident: #css
  3. # [00:15] * Joins: dbaron (dbaron@63.245.220.240)
  4. # [00:39] * Quits: shepazu (shepazu@128.30.52.169) (Client exited)
  5. # [00:41] * Joins: karl (karlcow@128.30.54.58)
  6. # [01:19] <fantasai> TabAtkins: Wrt flex-flow... I think we should publish the next draft with all the values Alex came up with, and mark it as an issue / request for feedback whether we want to keep all of them.
  7. # [01:19] <TabAtkins> fantasai: Why?
  8. # [01:19] <fantasai> TabAtkins: Because whatever we choose, it'll be a subset of those values
  9. # [01:20] <fantasai> TabAtkins: So we might as well publish with the full set so it's out there.
  10. # [01:20] <fantasai> TabAtkins: And also take advantage of that publication to ask for use cases etc.
  11. # [01:20] <fantasai> TabAtkins: I don't like trimming things prematurely before they reach /TR
  12. # [01:20] <fantasai> TabAtkins: Because while they're archived in CVS, hardly anyone looks at CVS.
  13. # [01:21] <fantasai> TabAtkins: I'd rather have an archived proposal on /TR, since people can point to it later when the issue comes around again
  14. # [01:21] <TabAtkins> I prefer publishing what we actually expect to implement. Chrome came out strongly in favor of dropping, and Alex seemed ambivalent about keeping.
  15. # [01:22] <fantasai> TabAtkins: The spec isn't done yet, and you're not publishing LC yet
  16. # [01:22] <fantasai> TabAtkins: And this isn't about expectations of implementation, it's about designing the feature.
  17. # [01:23] <TabAtkins> Yes, to both. Neither has any bearing on what I said.
  18. # [01:24] <fantasai> TabAtkins: We don't have agreement on whether we want physical-only, or logical-only, or both.
  19. # [01:24] <fantasai> TabAtkins: And we have a proposal that addresses all options.
  20. # [01:24] <TabAtkins> Well, we definitely know we don't want physical-only. The question is between logical-only or physical+logical+directional
  21. # [01:25] <fantasai> TabAtkins: I don't want that proposal to be buried in the CVS archives when it's quite likely we'll need to dig it up again in the future.
  22. # [01:25] <TabAtkins> And feedback indicates the former is better. I can add an issue about the latter if you'd like.
  23. # [01:25] <fantasai> TabAtkins: Not sure what you mean by physical+logical+directional
  24. # [01:25] <fantasai> TabAtkins: I've heard people argue for just about everything at this point. :/
  25. # [01:26] <TabAtkins> The mixed version ("horizontal"/"horizont-reverse"/etc) are based on 'direction', but not writing-mode.
  26. # [01:26] <TabAtkins> I mean, every value is potentially useful.
  27. # [01:27] <fantasai> Right, that's why we added them :)
  28. # [01:27] <TabAtkins> The question is just if they're sufficiently useful to justify confusing authors with literally a dozen subtly different keywords.
  29. # [01:27] <fantasai> If they're each potentially useful, there's a decent chance that 10 years from now we will add some that we aren't adding next year
  30. # [01:27] <fantasai> s/next/this/
  31. # [01:28] <fantasai> So that we don't have to redesign this from scratch at that point, I would like the full proposal to appear on /TR
  32. # [01:28] <TabAtkins> I don't agree with using /TR solely as an archive of possible ideas.
  33. # [01:28] <fantasai> And if we have a good idea of what we'd like to limit it to for now, we can mark that in the draft.
  34. # [01:29] <fantasai> And ask if anyone objects to us cutting out the rest of the values, and if they do, what they want them for
  35. # [01:29] <TabAtkins> I already know what they're useful for. We established use-cases previously.
  36. # [01:30] <fantasai> TabAtkins: I don't like iterating only on dev.w3.org.
  37. # [01:30] <fantasai> TabAtkins: I think we should publish more often to /TR
  38. # [01:31] <fantasai> TabAtkins: I don't mind if things that are pretty obviously stupid or pretty obviously wrong or pretty obviously insufficient never getting to /TR
  39. # [01:31] <TabAtkins> I agree that we shoudl publish more. I don't interpret that as "publish anything that might be a good idea for later".
  40. # [01:31] <TabAtkins> We have mailing list archives and cvs history to preserve those things.
  41. # [01:32] <fantasai> TabAtkins: Yeah, and the main search interface we have for that seems to be me.
  42. # [01:32] <TabAtkins> If you really want to have it around in the source, I can insert a comment for it.
  43. # [01:32] <TabAtkins> fantasai: You're so good at it! ^_^
  44. # [01:33] <fantasai> TabAtkins: No, I want it visible for people to see and comment on.
  45. # [01:33] <fantasai> TabAtkins: Yeah, but I'm not going to be around forever.
  46. # [01:33] <fantasai> TabAtkins: So the more people know something exists, the more people can dig it up when it's needed.
  47. # [01:33] <fantasai> TabAtkins: It's not just us, there might be some team out there that needs values we don't decide to keep
  48. # [01:33] <fantasai> TabAtkins: And would otherwise wind up inventing their own syntax extensions
  49. # [01:34] <TabAtkins> That's true for almost anything.
  50. # [01:34] <fantasai> TabAtkins: Yeah, but we did the work this time.
  51. # [01:34] <fantasai> TabAtkins: So let's put that work where it's findable.
  52. # [01:35] <fantasai> TabAtkins: I know dev.w3.org is publicly accessible, but that doesn't make things on it easily accessible
  53. # [01:35] <fantasai> TabAtkins: Both dev.w3.org and www-style are very high traffic
  54. # [01:35] <fantasai> TabAtkins: /TR and the CSSWG blog are low traffic
  55. # [01:35] <Hixie> fantasai: re your mail to www-style just now, tab's e-mail linked to my e-mail on the whatwg list which summarised use cases
  56. # [01:36] <fantasai> TabAtkins: Which means more people are able to keep up with them.
  57. # [01:36] <fantasai> TabAtkins: I'm arguing for not excluding those people from this discussion.
  58. # [01:36] <TabAtkins> fantasai: I'm arguing that there's no need to include them. We know the use-cases already.
  59. # [01:38] <fantasai> Hixie: Must've scanned too fast
  60. # [01:38] * fantasai rereads
  61. # [01:40] <fantasai> Hixie: I guess my question is, are people actually running into this problem, or is it mostly theoretical?
  62. # [01:40] <fantasai> Hixie: I understand the current behavior of querySelector is confusing
  63. # [01:41] <Hixie> i gave a concrete example at the very top of my e-mail
  64. # [01:41] <TabAtkins> fantasai: They're running into the problem of #1-style selectors mostly being confusing. In #2, though, there's clear use in allowing an element to be styled based on its context.
  65. # [01:42] <fantasai> TabAtkins: I'm having some trouble understanding why you'd use a scoped style block in that case
  66. # [01:42] <fantasai> TabAtkins: Not that I object to the use case, I just don't understand it.
  67. # [01:42] <TabAtkins> fantasai: Because you're trying to scope the styles? In the example at the top of Hixie's email, you only want to style your own <h1>, not all the <h1>s in the page.
  68. # [01:43] <fantasai> TabAtkins: So, I understand the case of scoped styles being used so you can compartmentalize a piece of a web page
  69. # [01:44] <fantasai> TabAtkins: In that case scoped styles are handy, and you can avoid interactions with things outside your compartment
  70. # [01:44] <fantasai> TabAtkins: The part of responding to the context outside your compartment.. is I guess a case I haven't thought about before.
  71. # [01:45] <fantasai> TabAtkins: And I'm wondering if that actually happens?
  72. # [01:45] <TabAtkins> Yes, it absolutely happens. I've done exactly this (with normal selectors, not with <style scoped>) in production webpages.
  73. # [01:48] <fantasai> TabAtkins: But is that a case where you would have used <style scoped> if it were available?
  74. # [01:48] <TabAtkins> Yes.
  75. # [01:48] <fantasai> TabAtkins: Why use <style scoped> in that case?
  76. # [01:48] <TabAtkins> Because it scopes. ^_^
  77. # [01:49] <fantasai> TabAtkins: But it also requires you to define those styles inline, no?
  78. # [01:49] <TabAtkins> By "inline", do you mean in a <style> rather than a <link>?
  79. # [01:49] <fantasai> TabAtkins: yeah
  80. # [01:50] <TabAtkins> Then yes. That was the easiest way to do this.
  81. # [01:50] <fantasai> TabAtkins: You can scope by using a class selector, right? Which is presumably what you did.
  82. # [01:50] <fantasai> TabAtkins: Why is <style scoped> better for what you did than doing that?
  83. # [01:51] <TabAtkins> I used an id selector, but yes. <style scoped> is better because the only reason I gave the blocks ids was to scope them. They were otherwise insignificant.
  84. # [01:51] <Hixie> <style scoped> supports @import, if you want out-of-band styles, btw
  85. # [01:53] <Hixie> fantasai: the use case for <style scoped> is syndication -- you can't add a class to the container (you don't create it) and you can't avoid clashes with other classes on the page (you don't control them)
  86. # [01:53] <fantasai> Hixie: Right, I got that. What I'm having trouble with is how that use case would use context outside the container.
  87. # [01:54] <TabAtkins> And it happens to be convenient for several other use-cases where you *do* control the markup outside.
  88. # [01:54] <fantasai> Hixie: Since in that use case, you presumably don't know anything about your container
  89. # [01:54] <fantasai> Hixie: Or at least that was my impression
  90. # [01:55] <fantasai> TabAtkins: That's like saying "the only reason I gave the blocks ids was so I could style them".
  91. # [01:56] <Hixie> fantasai: you can know something about your container. e.g. "the container is Google News and they have told us that we can distinguish whether we're being published on the front page or a dedicated page by examining the class of the <body>"
  92. # [01:57] <TabAtkins> fantasai: Prefixing every single style with the scoping id is rather frustrating.
  93. # [01:57] <fantasai> TabAtkins: Isn't that what your nested selector blocks proposal is supposed to solve? :)
  94. # [01:58] <TabAtkins> I fix my problems in several directions at once, yes. ^_^
  95. # [01:58] <fantasai> Hixie: Ok, that makes sense.
  96. # [01:58] <fantasai> Hixie: So then, would you always use :scope in a selector that's global?
  97. # [01:58] <fantasai> s/would/wouldn't/
  98. # [01:59] <fantasai> (Not arguing that we should distinguish that way, just trying to understand the constraints on the problem.)
  99. # [01:59] <TabAtkins> fantasai: I gave an example of why that's not equivalent. My example wouldn't apply in Hixie's use-case, though.
  100. # [02:00] <Hixie> fantasai: no, not necessarily
  101. # [02:00] <TabAtkins> fantasai: I also gave an example of why using :scope to distinguish the two cases isn't ideal - it means that you can't refer to the scoping element at all without :scope, which is slightly annoying.
  102. # [02:01] <Hixie> fantasai: e.g. if you jsut want to style all links on the front page one way, and all links on the article page another: body.front :link vs body.article :link
  103. # [02:01] <fantasai> Hixie: what's an example of not using :scope, or rather, where requiring :scope wouldn't work
  104. # [02:01] <Hixie> not to belabour the point, but the examples in the e-mail are sent already answer all these questions :-)
  105. # [02:02] <Hixie> http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2011-September/033222.html
  106. # [02:04] <fantasai> Hixie: Not seeing it.
  107. # [02:04] <fantasai> Hixie: I see cases where adding :scope seems excessively verbose
  108. # [02:04] <fantasai> Hixie: But not any where requiring it would cause a problem.
  109. # [02:04] <Hixie> none of the examples with @global use :scope
  110. # [02:05] <fantasai> Hixie: But they could.
  111. # [02:05] <Hixie> well sure, they could also use :root and universal selectors and :not() with bogus ID selectors
  112. # [02:06] <fantasai> Hixie: So is it fair to say that aside from the non-intuitiveness, the idea of triggering on the presence of :scope would solve the use cases?
  113. # [02:07] <TabAtkins> If you don't consider verbosity a problem, yes.
  114. # [02:07] <Hixie> it also affects teh specificity
  115. # [02:07] <Hixie> but yes, magic based on the presence or absence of :scope has been proposed as a solution in the past
  116. # [02:07] <fantasai> I'm not proposing that :)
  117. # [02:07] <fantasai> I just want to understand if a solution under the same technical constraints would solve the problem.
  118. # [02:09] <fantasai> ?
  119. # [02:09] <Hixie> i don't understand the question
  120. # [02:09] <TabAtkins> Yes, under the criteria of "matches the same elements", switching on the presence of :scope is equivalent to #2-style scoping + @global.
  121. # [02:10] <fantasai> TabAtkins: aside from intuitiveness, which is a problem, that solves the issue?
  122. # [02:10] <Hixie> the e-mail i cited above lists a whole bunch of different possible solutions
  123. # [02:10] <Hixie> including the magic-based-on-:scope thing
  124. # [02:10] <TabAtkins> Intuitiveness and verbosity are both problems with it.
  125. # [02:10] <fantasai> TabAtkins: But those are the only onese?
  126. # [02:10] <TabAtkins> I guess?
  127. # [02:10] <Hixie> it also affects teh specificity
  128. # [02:10] <fantasai> Is that a bug or a feature?
  129. # [02:11] <fantasai> (I can see arguments both ways)
  130. # [02:11] <TabAtkins> Bug. ^_^ (I've many times run into problems caused by the required "scope it with an id" technique.)
  131. # [02:11] <fantasai> Ok
  132. # [02:11] <fantasai> I'll try to draft a solution, and we'll see where it goes.
  133. # [02:12] <TabAtkins> If it's based on switching based on the presence of :scope, I'll yell at you.
  134. # [02:12] <Hixie> other proposals are of course always welcome
  135. # [02:12] <fantasai> LOL
  136. # [02:12] <Hixie> but if you do want to come up with another solution, i encourage you to actually read the whatwg thread first
  137. # [02:12] <Hixie> (and not just skim it)
  138. # [02:12] <Hixie> since a whole host of them were proposed there
  139. # [02:12] <fantasai> Hixie: ok
  140. # [02:13] <Hixie> just out of interest, do you see any issues with the @global idea?
  141. # [02:13] <TabAtkins> Note that Hixie's email that I linked in the summary is not the start of the thread. I link the start a bit further down.
  142. # [02:13] <fantasai> TabAtkins: Yeah, I know. Hixie's email is a summary of the thread. He's good at that :)
  143. # [02:13] <fantasai> Hixie: Yeah, I think it's kindof awkward.
  144. # [02:14] <Hixie> awkward?
  145. # [02:14] <fantasai> Hixie: I suppose that's not a well-articulated issue though.
  146. # [02:14] <Hixie> as in "not smooth or graceful; ungainly"?
  147. # [02:14] <fantasai> Hixie: I don't think it captures the structure of what we're trying to express very well.
  148. # [02:14] <Hixie> ah
  149. # [02:15] <Hixie> well i'm not sure what we're trying to express other than "change how this selector is matched to be global rather than scoped"
  150. # [02:15] <fantasai> :)
  151. # [02:15] <Hixie> for which @rules seemed emminently appropriate :-)
  152. # [02:15] <fantasai> Actually, what we're trying to express isn't that
  153. # [02:15] <Hixie> i could see a desire for @global { rule { } } rather than @global rule { }
  154. # [02:15] <Hixie> but which is more verbose probably depends on how many rules depend on the global scope
  155. # [02:16] <fantasai> but rather, "I want this selector to be scoped within my containment element's scope, but I only want it to take effect within this context."
  156. # [02:16] <fantasai> s/within this context/when my containment is within this context/
  157. # [02:16] <Hixie> not sure how that differs, but ok
  158. # [02:16] <Hixie> either way it's a mode flag
  159. # [02:16] <fantasai> You're not suddenly thinking globally when you switch to @global.
  160. # [02:17] <fantasai> You're still thinking of your little scoped-style box
  161. # [02:17] <fantasai> But you want to key off something outside the context of the box.
  162. # [02:17] <Hixie> xbl2 called it something like "allow-selectors-through"
  163. # [02:17] * Joins: shepazu (shepazu@128.30.52.169)
  164. # [02:17] <Hixie> iirc
  165. # [02:19] <TabAtkins> fantasai: I think of it as their being two scopes - a matching scope and a filtering scope. This switches the matching scope to be the root, while leaving the filtering scope alone.
  166. # [02:19] * Quits: dsinger (dsinger@216.178.108.111) (Quit: dsinger)
  167. # [02:20] <fantasai> TabAtkins: Yeah, I don't see it that way. I think of there being two parts to consider: the part inside my box, and the part outside my box
  168. # [02:20] <fantasai> TabAtkins: Normally we want to only think about/consider/match within the part inside the box
  169. # [02:20] <fantasai> TabAtkins: But sometimes we want to select against components of the context as well.
  170. # [02:24] <TabAtkins> I think that's what I said, too. ^_^
  171. # [02:24] <fantasai> TabAtkins: It's equivalent in functionality, but not in construction. :)
  172. # [02:30] * Joins: myakura (myakura@203.136.181.177)
  173. # [02:31] <stearns> fantasai: Should I assume that the "rules for pagination" would apply to regions as well?
  174. # [02:31] <TabAtkins> Pretty sure that's the intent - that regions act like mini-pages.
  175. # [02:32] <stearns> that's my assumption as well - just wanted to verify
  176. # [02:59] * Quits: stearns (anonymous@192.150.22.5) (Ping timeout)
  177. # [03:30] * Quits: dbaron (dbaron@63.245.220.240) (Ping timeout)
  178. # [03:53] * Joins: stearns (anonymous@50.132.9.217)
  179. # [04:07] * Joins: Florian (Florian@58.1.224.28)
  180. # [04:12] * Joins: dbaron (dbaron@173.228.28.227)
  181. # [04:30] * Quits: dbaron (dbaron@173.228.28.227) (Ping timeout)
  182. # [05:02] * Joins: pjrm (pmoulder@130.194.69.10)
  183. # [05:45] * Joins: dbaron (dbaron@173.228.28.227)
  184. # [06:48] * Joins: dsinger (dsinger@12.201.85.2)
  185. # [07:07] * Quits: stearns (anonymous@50.132.9.217) (Quit: stearns)
  186. # [07:21] * Joins: homata (homata@58.158.182.50)
  187. # [08:44] * Quits: dbaron (dbaron@173.228.28.227) (Ping timeout)
  188. # [09:46] * Quits: nimbupani (Adium@24.18.47.160) (Client exited)
  189. # [11:49] * Quits: myakura (myakura@203.136.181.177) (Client exited)
  190. # [11:49] * Joins: myakura (myakura@203.136.181.177)
  191. # [11:52] * Quits: myakura (myakura@203.136.181.177) (Ping timeout)
  192. # [12:52] * Quits: homata (homata@58.158.182.50) (Ping timeout)
  193. # [12:56] * Quits: karl (karlcow@128.30.54.58) (Quit: Freedom - to walk free and own no superior.)
  194. # [13:27] * Joins: homata (homata@58.158.182.50)
  195. # [14:20] * Joins: myakura (myakura@203.136.181.177)
  196. # [15:17] * Joins: stearns (anonymous@50.132.9.217)
  197. # [15:21] * Quits: homata (homata@58.158.182.50) (Ping timeout)
  198. # [16:08] * Joins: Ms2ger (Ms2ger@91.181.17.37)
  199. # [16:09] * Joins: PHPar (bca1ef7a@109.169.29.95)
  200. # [16:12] * Joins: nimbupani (Adium@24.18.47.160)
  201. # [17:03] * Parts: PHPar (bca1ef7a@109.169.29.95)
  202. # [17:06] * Quits: dsinger (dsinger@12.201.85.2) (Quit: dsinger)
  203. # [17:22] * Joins: dsinger (dsinger@216.178.108.111)
  204. # [17:27] * Joins: dbaron (dbaron@173.228.28.227)
  205. # [17:47] * Joins: arronei (arronei@131.107.0.112)
  206. # [17:47] * Quits: arronei_ (arronei@131.107.0.76) (Ping timeout)
  207. # [18:14] * Quits: stearns (anonymous@50.132.9.217) (Quit: stearns)
  208. # [18:34] * Quits: myakura (myakura@203.136.181.177) (Client exited)
  209. # [18:34] * Joins: myakura (myakura@203.136.181.177)
  210. # [18:36] * Quits: plinss (plinss@68.107.116.23) (Quit: plinss)
  211. # [18:37] * Quits: myakura (myakura@203.136.181.177) (Ping timeout)
  212. # [18:39] * Joins: plinss (plinss@68.107.116.23)
  213. # [18:43] * Joins: stearns (anonymous@192.150.22.5)
  214. # [19:41] * Quits: dbaron (dbaron@173.228.28.227) (Ping timeout)
  215. # [19:50] * Joins: appelsinol (appelsinol@129.241.206.180)
  216. # [20:11] * Joins: dbaron (dbaron@63.245.220.224)
  217. # [20:16] * Quits: appelsinol (appelsinol@129.241.206.180) (Quit: appelsinol)
  218. # [20:32] * Quits: dbaron (dbaron@63.245.220.224) (Ping timeout)
  219. # [20:44] * Joins: dbaron (dbaron@63.245.220.240)
  220. # [20:51] * Joins: plinss_ (plinss@192.6.114.30)
  221. # [21:48] * Quits: Ms2ger (Ms2ger@91.181.17.37) (Quit: nn)
  222. # [22:07] * Joins: shepazutoo (shepazu@128.30.52.169)
  223. # [22:11] * Joins: arno (arno@192.150.10.200)
  224. # [22:11] * Quits: shepazu (shepazu@128.30.52.169) (Ping timeout)
  225. # [22:11] * shepazutoo is now known as shepazu
  226. # [23:40] * Quits: plinss_ (plinss@192.6.114.30) (Quit: plinss_)
  227. # Session Close: Wed Sep 21 00:00:00 2011

The end :)