/irc-logs / freenode / #whatwg / 2009-09-06 / end

Options:

1. # Session Start: Sun Sep 06 00:00:00 2009
2. # Session Ident: #whatwg
3. # [00:05] * Quits: Unixmonkey (n=Unixmonk@ppp-69-208-150-144.dsl.ipltin.ameritech.net)
4. # [00:11] * Joins: Unixmonkey (n=Unixmonk@ppp-69-208-150-144.dsl.ipltin.ameritech.net)
5. # [00:13] * Quits: annevk3 (n=annevk@5355732C.cable.casema.nl)
6. # [00:14] * Joins: annevk2 (n=annevk@5355732C.cable.casema.nl)
7. # [00:17] * Quits: gsnedders (n=gsnedder@host86-167-19-36.range86-167.btcentralplus.com) (Read error: 104 (Connection reset by peer))
8. # [00:26] * Joins: yutak_home (n=kee@M006079.ppp.dion.ne.jp)
9. # [00:29] * Quits: svl_ (n=me@f051021207.adsl.alicedsl.de) ("And back he spurred like a madman, shrieking a curse to the sky.")
10. # [00:29] * Joins: dglazkov (n=dglazkov@c-67-188-0-62.hsd1.ca.comcast.net)
11. # [00:33] * Joins: cardona507 (n=cardona5@c-67-180-160-250.hsd1.ca.comcast.net)
12. # [00:35] * Joins: gsnedders (n=gsnedder@host86-167-19-36.range86-167.btcentralplus.com)
13. # [00:39] * Quits: Unixmonkey (n=Unixmonk@ppp-69-208-150-144.dsl.ipltin.ameritech.net)
14. # [00:40] * Joins: Unixmonkey (n=Unixmonk@ppp-69-208-150-144.dsl.ipltin.ameritech.net)
15. # [00:49] * Joins: SamerZ (n=SamerZ@CPE0024369ef3ab-CM001ac35cd4b4.cpe.net.cable.rogers.com)
16. # [00:51] * Joins: tantek (n=tantek@c-76-126-175-28.hsd1.ca.comcast.net)
17. # [00:54] * Quits: gsnedders (n=gsnedder@host86-167-19-36.range86-167.btcentralplus.com)
18. # [01:10] * Joins: murr4y` (n=murray@85.84-49-67.nextgentel.com)
19. # [01:11] * Quits: dglazkov (n=dglazkov@c-67-188-0-62.hsd1.ca.comcast.net)
20. # [01:19] * Quits: SamerZ (n=SamerZ@CPE0024369ef3ab-CM001ac35cd4b4.cpe.net.cable.rogers.com)
21. # [01:19] * Quits: Unixmonkey (n=Unixmonk@ppp-69-208-150-144.dsl.ipltin.ameritech.net)
22. # [01:21] * Quits: kristallpirat (n=kristall@c-base/crew/kristall) ("Wünsche weiterhin guten Flug")
23. # [01:25] <cardona507> no more name=""? a links to id now- nice
24. # [01:26] * Quits: murr4y (n=murray@85.84-49-67.nextgentel.com) (Read error: 110 (Connection timed out))
25. # [01:28] <Lachy> Updating this wiki page is proving to be a very time consuming task. The old version was significantly out of date. http://wiki.whatwg.org/wiki/HTML_vs._XHTML
26. # [01:29] <Lachy> but hopefully the new format I'm using is a lot more readable than it was before
27. # [01:31] <cardona507> that is a great resource. Thanks for your work on it.
28. # [01:33] * Quits: MikeSmith (n=MikeSmit@31-35-163.wireless.csail.mit.edu) ("Tomorrow to fresh woods, and pastures new.")
29. # [01:43] * Joins: SamerZ (n=SamerZ@CPE0024369ef3ab-CM001ac35cd4b4.cpe.net.cable.rogers.com)
30. # [01:51] * Joins: benward (n=benward@98.210.154.133)
31. # [02:33] * Quits: da3d (n=opera@h11n1fls34o986.telia.com)
32. # [02:49] * Quits: benward (n=benward@98.210.154.133)
33. # [02:49] * Joins: benward (n=benward@98.210.154.133)
34. # [03:14] * Joins: Camaban (n=alee@88-110-160-23.dynamic.dsl.as9105.com)
35. # [03:15] * Parts: Camaban (n=alee@88-110-160-23.dynamic.dsl.as9105.com) ("Ex-Chat")
36. # [03:16] * Joins: Unixmonkey (n=Unixmonk@adsl-69-208-129-202.dsl.ipltin.ameritech.net)
37. # [03:20] * Joins: Braxios (n=alee@88-110-160-23.dynamic.dsl.as9105.com)
38. # [03:21] <TabAtkins> I appreciate json more every day.
39. # [03:22] * Quits: Braxios (n=alee@88-110-160-23.dynamic.dsl.as9105.com) (Client Quit)
40. # [03:23] * Joins: nessy (n=nessy@ppp115-228.static.internode.on.net)
41. # [03:31] <Hixie> pity it doesn't define error handling, comments, or compatibility with JS Number types, but yes, other than that, it's a nice simple format
42. # [03:33] <inimino> I don't think it'd have caught on if it defined error handling
43. # [03:35] * Joins: rubys2 (n=rubys@cpe-065-190-139-141.nc.res.rr.com)
44. # [03:36] <inimino> (as anything other than "abort if the input doesn't match this grammar", that is)
45. # [03:36] <Hixie> that'd be fine too
46. # [03:36] <Hixie> i have nothing intrinsically against draconian error handling in general
47. # [03:37] <inimino> is that not implied?
48. # [03:38] <Hixie> the one thing that specs generally should try to avoid, is implication
49. # [03:38] <Hixie> well, s/the// :-)
50. # [03:40] <inimino> true enough :)
51. # [03:40] <inimino> the RFC does say "A JSON parser MAY accept non-JSON forms or extensions."
52. # [03:40] <Hixie> that's code for "there's no defined error handling"
53. # [03:42] <rubys2> I believe that the native JSON object coming in ECMAScript 5 is pretty draconian
54. # [03:42] <inimino> ES5's JSON.parse does define error handling, it throws a SyntaxError if the input does not conform to the grammar
55. # [03:42] <Hixie> yeah, es5's json is better
56. # [03:50] <cardona507> any good examples of <canvas> art?
57. # [03:50] <Hixie> http://www.whatwg.org/demos/2008-sept/color/color.html
58. # [03:52] <cardona507> that I saw (good video by the way)- That video is what got me interested in html5 - any other examples?
59. # [03:53] <inimino> http://www.bel.fi/~alankila/plasma.html
60. # [03:53] * Quits: tantek (n=tantek@c-76-126-175-28.hsd1.ca.comcast.net) (Read error: 104 (Connection reset by peer))
61. # [03:53] <cardona507> good one
62. # [03:54] <inimino> plenty more out there
63. # [03:58] * Quits: TabAtkins (n=chatzill@99-35-179-251.lightspeed.hstntx.sbcglobal.net) (Read error: 110 (Connection timed out))
64. # [04:00] * Joins: dglazkov (n=dglazkov@c-67-188-0-62.hsd1.ca.comcast.net)
65. # [04:11] * Joins: othermaciej (n=mjs@c-69-181-42-237.hsd1.ca.comcast.net)
66. # [04:20] * Joins: TabAtkins (n=chatzill@99-35-179-251.lightspeed.hstntx.sbcglobal.net)
67. # [04:21] <TabAtkins> Hixie: JSON purposely doesn't define comments. They did at first, but people were using them to embed proprietary parsing directives, so they took them out instead.
68. # [04:22] <Hixie> and made the proprietary directives valid instead? :-)
69. # [04:22] <TabAtkins> Nope. ^_^
70. # [04:22] <Hixie> "A JSON parser MAY accept non-JSON forms or extensions."
71. # [04:23] <TabAtkins> Ok, point.
72. # [04:24] <TabAtkins> But they're now explicitly proprietary, with no pretense of being able to be parsed safely by a generic parser.
73. # [04:24] <TabAtkins> Unlike the situation with comments, where the structure *would* be parsed by a generic parser, just incorrectly (because it didn't know how to obey the instructions embedded in the comments).
74. # [04:27] <Hixie> so what you're saying is that they removed both the possibility of comments and the safe extension mechanism at the same time
75. # [04:29] <TabAtkins> Yes, precisely.
76. # [04:30] <TabAtkins> That was the point - JSON wasn't supposed to ever be extended.
77. # [04:32] <TabAtkins> After all, it very purposely is a lowest common denominator. It includes only things that *every* non-pathological language has and can understand. Allowing extensions would inevitably lead to complexity creep, and make it difficult/impossible to use as a truly universal data-interechange format.
78. # [04:38] * Quits: nessy (n=nessy@ppp115-228.static.internode.on.net) ("This computer has gone to sleep")
79. # [04:42] <Hixie> if it wasn't supposed to ever be extended, why say "A JSON parser MAY accept non-JSON forms or extensions."
80. # [04:45] <TabAtkins> Because that's precisely what existed in the wild - parsers that understood non-JSON extensions. Pretending they didn't exist wouldn't have served any point, but making it so that their extensions would cause a normal JSON parser to abort kept the infection from ever spreading far enough to be a de facto language extension.
81. # [04:47] <TabAtkins> Sorry, heading out now; wife wants to play video games.
82. # [04:51] <Hixie> later
83. # [04:56] <cardona507> I am trying to style my markup in columns - I am using <section> but for some reason I can't style it- do I wrap it in a <div> and style the div? Does that defeat the point of <div>?
84. # [04:56] <cardona507> I mean defeat the point of <section>?
85. # [04:57] * Quits: TabAtkins (n=chatzill@99-35-179-251.lightspeed.hstntx.sbcglobal.net) (Read error: 60 (Operation timed out))
86. # [05:00] * Parts: rubys2 (n=rubys@cpe-065-190-139-141.nc.res.rr.com)
87. # [05:15] * Quits: Unixmonkey (n=Unixmonk@adsl-69-208-129-202.dsl.ipltin.ameritech.net)
88. # [05:22] <inimino> I'm kind of curious why the RFC says that myself, seems highly un-Crockfordian
89. # [05:26] * Quits: cardona507 (n=cardona5@c-67-180-160-250.hsd1.ca.comcast.net)
90. # [05:28] * Joins: cardona507 (n=cardona5@c-67-180-160-250.hsd1.ca.comcast.net)
91. # [05:38] * Joins: htmlr (n=htmlr@203-214-78-100.dyn.iinet.net.au)
92. # [05:42] * Quits: htmlr (n=htmlr@203-214-78-100.dyn.iinet.net.au) (Remote closed the connection)
93. # [05:43] * Joins: htmlr (n=htmlr@203-214-78-100.dyn.iinet.net.au)
94. # [05:49] * Quits: SamerZ (n=SamerZ@CPE0024369ef3ab-CM001ac35cd4b4.cpe.net.cable.rogers.com)
95. # [06:09] * Joins: TabAtkins (n=chatzill@99.35.179.251)
96. # [06:10] * Quits: benward (n=benward@98.210.154.133)
97. # [06:20] * Joins: dglazkov_ (n=dglazkov@72.14.224.1)
98. # [06:21] * Quits: shepazu (n=schepers@31-33-83.wireless.csail.mit.edu)
99. # [06:30] * Joins: SamerZ (n=SamerZ@CPE0024369ef3ab-CM001ac35cd4b4.cpe.net.cable.rogers.com)
100. # [06:35] * Joins: bgalbraith (n=bgalbrai@71.202.109.116)
101. # [06:36] * Quits: bgalbraith (n=bgalbrai@71.202.109.116) (Client Quit)
102. # [06:37] * Quits: htmlr (n=htmlr@203-214-78-100.dyn.iinet.net.au)
103. # [06:38] * Quits: dglazkov (n=dglazkov@c-67-188-0-62.hsd1.ca.comcast.net) (Read error: 110 (Connection timed out))
104. # [06:38] * dglazkov_ is now known as dglazkov
105. # [06:43] * Joins: dglazkov_ (n=dglazkov@c-67-188-0-62.hsd1.ca.comcast.net)
106. # [07:03] * Quits: dglazkov (n=dglazkov@72.14.224.1) (Read error: 110 (Connection timed out))
107. # [07:03] * dglazkov_ is now known as dglazkov
108. # [07:18] * Quits: dglazkov (n=dglazkov@c-67-188-0-62.hsd1.ca.comcast.net)
109. # [07:20] * Quits: kinetik (n=kinetik@121.98.132.55) ("leaving")
110. # [07:23] * Quits: othermaciej (n=mjs@c-69-181-42-237.hsd1.ca.comcast.net)
111. # [07:28] * Joins: SamerZ_ (n=SamerZ@CPE0024369ef3ab-CM001ac35cd4b4.cpe.net.cable.rogers.com)
112. # [07:29] * Quits: cardona507 (n=cardona5@c-67-180-160-250.hsd1.ca.comcast.net)
113. # [07:30] * Joins: cardona507 (n=cardona5@c-67-180-160-250.hsd1.ca.comcast.net)
114. # [07:31] * Joins: mlpug (n=mlpug@a88-115-164-40.elisa-laajakaista.fi)
115. # [07:32] * Joins: myakura (n=myakura@p4115-ipbf3106marunouchi.tokyo.ocn.ne.jp)
116. # [07:36] * Joins: Unixmonkey (n=Unixmonk@ppp-69-208-150-144.dsl.ipltin.ameritech.net)
117. # [07:36] * Quits: TabAtkins (n=chatzill@99.35.179.251) (Read error: 145 (Connection timed out))
118. # [07:42] * Quits: SamerZ (n=SamerZ@CPE0024369ef3ab-CM001ac35cd4b4.cpe.net.cable.rogers.com) (Read error: 113 (No route to host))
119. # [07:42] * SamerZ_ is now known as SamerZ
120. # [07:47] * Joins: kinetik (n=kinetik@121.98.132.55)
121. # [07:48] * Quits: kinetik (n=kinetik@121.98.132.55) (Client Quit)
122. # [07:54] * Joins: kinetik (n=kinetik@121.98.132.55)
123. # [08:07] * Quits: cardona507 (n=cardona5@c-67-180-160-250.hsd1.ca.comcast.net)
124. # [08:08] * Quits: Unixmonkey (n=Unixmonk@ppp-69-208-150-144.dsl.ipltin.ameritech.net)
125. # [08:29] * Joins: shepazu (n=schepers@72.255.53.72)
126. # [08:29] * Quits: SamerZ (n=SamerZ@CPE0024369ef3ab-CM001ac35cd4b4.cpe.net.cable.rogers.com)
127. # [08:34] * Joins: lazni (n=lazni@118.71.60.182)
128. # [08:41] * Quits: number-6 (n=number-6@adsl-69-230-169-54.dsl.irvnca.pacbell.net)
129. # [08:48] * Parts: lazni (n=lazni@118.71.60.182)
130. # [08:54] * Joins: cardona507 (n=cardona5@c-67-180-160-250.hsd1.ca.comcast.net)
131. # [08:55] * Joins: lazni (n=lazni@118.71.60.182)
132. # [08:56] * Quits: cardona507 (n=cardona5@c-67-180-160-250.hsd1.ca.comcast.net) (Client Quit)
133. # [09:08] * Joins: cardona507 (n=cardona5@c-67-180-160-250.hsd1.ca.comcast.net)
134. # [09:10] * Quits: lazni (n=lazni@118.71.60.182) ("Leaving.")
135. # [09:11] * Joins: lazni (n=lazni@118.71.60.182)
136. # [09:12] * Quits: cardona507 (n=cardona5@c-67-180-160-250.hsd1.ca.comcast.net) (Client Quit)
137. # [09:36] * Joins: Rik` (n=Rik`@pha75-2-81-57-187-57.fbx.proxad.net)
138. # [09:52] * Joins: maikmerten (n=maikmert@Zaef1.z.pppool.de)
139. # [10:01] * Quits: myakura (n=myakura@p4115-ipbf3106marunouchi.tokyo.ocn.ne.jp) ("Leaving...")
140. # [10:06] <Dashiva> Something tells me most people won't get this "right": http://paulirish.com/2009/bulletproof-font-face-implementation-syntax/
141. # [10:07] * Quits: kinetik (n=kinetik@121.98.132.55) (Remote closed the connection)
142. # [10:07] * Joins: kinetik (n=kinetik@121.98.132.55)
143. # [10:16] * Quits: lazni (n=lazni@118.71.60.182) ("Leaving.")
144. # [10:17] * Joins: Maurice (i=copyman@5ED548D4.cable.ziggo.nl)
145. # [10:20] * Joins: lazni (n=lazni@118.71.60.182)
146. # [10:20] * Quits: shepazu (n=schepers@72.255.53.72)
147. # [10:26] * Joins: virtuelv (n=virtuelv@125.175.251.212.customer.cdi.no)
148. # [10:35] * Joins: Chris_Wilson (n=cwilso@nat/microsoft/x-ysdyajwetocqfgfb)
149. # [10:38] * Quits: ChrisWilson (n=cwilso@131.107.0.101) (Read error: 145 (Connection timed out))
150. # [10:45] * Joins: annevk3 (n=annevk@83.85.115.44)
151. # [10:46] * Quits: kinetik (n=kinetik@121.98.132.55) (Remote closed the connection)
152. # [10:47] * Joins: kinetik (n=kinetik@121.98.132.55)
153. # [10:49] * Quits: Rik` (n=Rik`@pha75-2-81-57-187-57.fbx.proxad.net)
154. # [10:53] * Quits: kinetik (n=kinetik@121.98.132.55) (Remote closed the connection)
155. # [11:29] * Joins: kinetik (n=kinetik@121.98.132.55)
156. # [11:37] * Quits: kinetik (n=kinetik@121.98.132.55) (Remote closed the connection)
157. # [11:37] * Joins: kinetik (n=kinetik@121.98.132.55)
158. # [11:43] * Joins: myakura (n=myakura@p4115-ipbf3106marunouchi.tokyo.ocn.ne.jp)
159. # [11:45] * Quits: kinetik (n=kinetik@121.98.132.55) (Remote closed the connection)
160. # [11:55] * Joins: kinetik (n=kinetik@121.98.132.55)
161. # [11:57] * Quits: lazni (n=lazni@118.71.60.182) ("Leaving.")
162. # [12:04] <AryehGregor> "no final decisions on the content of the W3C spec will be made outside of public-html" So I guess not on the whatwg list, then, huh?
163. # [12:05] <Hixie> i guess we'll just have to have a "final decision" to do the same as the whatwg after the whatwg is done :-)
164. # [12:05] <Hixie> i guess that means the w3c can never be the first one to finish html5
165. # [12:06] * Quits: kinetik (n=kinetik@121.98.132.55) (Remote closed the connection)
166. # [12:06] * Joins: kinetik (n=kinetik@121.98.132.55)
167. # [12:13] * Quits: kinetik (n=kinetik@121.98.132.55) (Remote closed the connection)
168. # [12:14] * Joins: kinetik (n=kinetik@121.98.132.55)
169. # [12:18] <AryehGregor> Hixie, isn't specifying things like how to parse a nonnegative integer algorithmically going a little too far? Couldn't you just give it in some sort of BNF like <*whitespace "+"? 1*digit garbage> and say to return an error if it doesn't match the pattern, and otherwise return the value given by 1*digit parsed as a decimal number? Actually, I notice you don't seem to use formal grammars anywhere in the spec.
170. # [12:18] <Lachy> yeah, I always found it intriguing how the HTMLWG thinks it has total authority over the content of the spec and is entitled to such privliges, despite supposedly working with the WHATWG
171. # [12:19] <AryehGregor> Formal grammars are just as precise as algorithms, and typically easier to read, so unless they can't meet HTML 5's error-handling requirements, I don't see why algorithms would be preferred.
172. # [12:20] <Hixie> AryehGregor: a grammar doesn't tell you how to interpret the value, only how to check if it's valid or not.
173. # [12:20] <AryehGregor> I was actually trying to implement the real number parsing in PHP for MediaWiki (to ensure consistent client-/server-side parsing of <input type=number>), but gave up because the algorithm given is ridiculous in a language like PHP and the only sane thing to do was try to convert it to regex.
174. # [12:20] <AryehGregor> Hixie, it breaks it into pieces for you, and you can then assign meaning to the pieces.
175. # [12:21] <AryehGregor> In this case, "parse the 1*digit bit as a decimal integer, and ignore the rest".
176. # [12:22] <Hixie> you aren't supposed to implement the algorithms line for line
177. # [12:22] <Hixie> translate the algorithms into whatever is the most optimal implementation for your language
178. # [12:22] <Hixie> AryehGregor: then you're using algorithm _and_ a grammar
179. # [12:23] <Hixie> AryehGregor: and the spec becomes very inconsistent, sing grammars in the simple cases, and n the others
180. # [12:23] <AryehGregor> Part of the algorithm is just "parse this using BNF", which is already well-understood. Rely on existing standards and all that.
181. # [12:23] <AryehGregor> I think some letters got dropped from your last line.
182. # [12:24] <Hixie> hold on computer issues
183. # [12:24] <AryehGregor> In my case the usual way to do it would be regex, which is a lot closer to BNF than the C-style algorithm you give. I should think it would be easier to translate from BNF into a C-style algorithm than the reverse.
184. # [12:24] <annevk3> did you try writing something out?
185. # [12:24] <AryehGregor> In cases like the HTML parser you probably really can't use BNF because you need to specify error-handling and such in so much detail.
186. # [12:25] <annevk3> you need to specify error-handling in detail everywhere...
187. # [12:25] <Hixie> ok where was i
188. # [12:26] <AryehGregor> annevk3, I tried briefly. To be honest, in the end I mostly gave up because I didn't want to use it unless I had a test suite that would confirm I didn't botch things much. I'm sure I'll return to it and figure it out, but most specs use BNF and it seems both shorter and easier to read to me, and no less precise.
189. # [12:26] <Hixie> i was saying that if you use algorithms everywhere but some of them also use a grammar, you end up with an inconsistent style
190. # [12:26] <AryehGregor> annevk3, for things like parsing a non-negative integer there's no error handling except "return an error if it's invalid", which BNF is fine at.
191. # [12:26] <Hixie> and worse, maintenance becomes a nightmare
192. # [12:26] <Hixie> because sometimes (often!) things get more complicated
193. # [12:26] <Hixie> and so you then suddenly find yourself having to expand what was an algorithm that used a grammar into something without a grammar
194. # [12:27] <AryehGregor> Hmm.
195. # [12:27] <AryehGregor> Okay, that makes sense.
196. # [12:27] * Hixie isn't a big fan of formal grammars because for anything but the simplest of syntaxes, they simply aren't readable, nor do they actually define the semantics, nor do they make error handling even remotely sane to specify
197. # [12:28] <AryehGregor> Well, I was specifically referring to really simple syntaxes. :)
198. # [12:28] <AryehGregor> Anyway, Hixie, while we're chatting, seems the rules for unquoted attributes are too lax: http://code.google.com/p/html5lib/issues/detail?id=93
199. # [12:28] <AryehGregor> (I could file a bug report or post on the mailing list, but, why bother.)
200. # [12:28] <Hixie> if you want the spec to change, please file a bug or send a mail :-)
201. # [12:28] <Hixie> you can use the bug report tool in the spec
202. # [12:29] <AryehGregor> Okay.
203. # [12:29] <Hixie> it's as easy as typing on irc
204. # [12:29] <jgraham> Also, if you want to patch html5lib...
205. # [12:30] <AryehGregor> Well, I happened to implement this part of the spec myself in MediaWiki. I'm not specifically concerned about html5lib, I'm sure someone else can fix that.
206. # [12:32] * Quits: kinetik (n=kinetik@121.98.132.55) (Remote closed the connection)
207. # [12:32] <Lachy> Hixie, the spec seems to lack clear authoring requirements for using SVG and MathML in HTML. It seems that authors have to derive the requirements from the implementation requirements
208. # [12:32] * Joins: kinetik (n=kinetik@121.98.132.55)
209. # [12:32] <Hixie> Lachy: well that's not good. file a bug.
210. # [12:33] <Lachy> I will soon. Just need to work out everthing that's missing.
211. # [12:33] <Lachy> (I'll just note them in IRC for now, and then file the issues appropriately later)
212. # [12:34] <Hixie> thanks
213. # [12:34] <Lachy> the section about the case sensitivity of tag names isn't clear if it applies to HTML elements only, or all elements
214. # [12:35] * murr4y` is now known as murr4y
215. # [12:35] <Lachy> requirement to use "xlink" as the prefix for xlink attributes, rather than allowing anything like in XML
216. # [12:38] <Lachy> requirement on the values of xmlns:xlink attributes and other xmlns attributes to use the specific namespaces, and that other custom namespaces aren't permitted
217. # [12:44] * Joins: da3d (n=opera@h11n1fls34o986.telia.com)
218. # [12:44] <Lachy> (I must also remember to file the bug on hsivonen's HTML5 parser about handling LF after <textarea> start tag in Firefox.)
219. # [12:45] * Joins: Kalms (n=rasmuska@81.161.185.108)
220. # [12:49] * Quits: kinetik (n=kinetik@121.98.132.55) (Remote closed the connection)
221. # [12:50] * Joins: kinetik (n=kinetik@121.98.132.55)
222. # [12:51] <Lachy> also describe that the attributes xlink:actuate, xlink:arcrole, xlink:href, xlink:role, xlink:show, xlink:title, xlink:type, xml:base, xml:lang and xml:space can be used on foreign elements and must use those prefixes
223. # [12:51] <Lachy> and that declaring the xmlns and xmlns:xlink attributes on SVG elements is optional
224. # [12:55] * Quits: virtuelv (n=virtuelv@125.175.251.212.customer.cdi.no) (Remote closed the connection)
225. # [12:57] * Quits: kinetik (n=kinetik@121.98.132.55) (Remote closed the connection)
226. # [13:02] * Joins: kinetik (n=kinetik@121.98.132.55)
227. # [13:16] * Joins: virtuelv (n=virtuelv@125.175.251.212.customer.cdi.no)
228. # [13:36] * Quits: karlcow (n=karl@nerval.la-grange.net) ("This computer has gone to sleep")
229. # [13:38] * Joins: nessy (n=nessy@203-214-73-15.dyn.iinet.net.au)
230. # [13:43] * Joins: ROBOd (n=robod@89.122.216.38)
231. # [13:46] * Quits: nessy (n=nessy@203-214-73-15.dyn.iinet.net.au) ("This computer has gone to sleep")
232. # [13:49] * Joins: nessy (n=nessy@203-214-73-15.dyn.iinet.net.au)
233. # [13:56] <Hixie> hey, google's rdfa implementation can now be tested for bugs: http://www.google.com/webmasters/tools/richsnippets?url=http%3A%2F%2Fexamples.tobyinkster.co.uk%2Fhcard
234. # [14:11] * Joins: maikmerten_ (n=maikmert@BAE1a14.bae.pppool.de)
235. # [14:13] * Quits: maikmerten (n=maikmert@Zaef1.z.pppool.de) (Read error: 60 (Operation timed out))
236. # [14:14] <da3d> Why does <nav><ul/></nav> count as an untitled section in the outliner? It seems counter-intuitive to me...
237. # [14:15] <Hixie> why?
238. # [14:15] <Hixie> <nav> is like a special kind of <section>
239. # [14:17] <Lachy> Hixie, http://www.w3.org/Bugs/Public/show_bug.cgi?id=7483 can be marked RESOLVED FIXED now
240. # [14:18] <Hixie> go ahead
241. # [14:18] <Hixie> you all are as welcome to close bugs as i am
242. # [14:18] <Lachy> really? I thought you were meant to do it. But ok
243. # [14:18] <da3d> That seems odd to me, because I never thought of <nav> as a type of section. It confused me when I tried the outliner last night (I thought the untitled section had something to do with an <article>).
244. # [14:18] <Hixie> it's the wg bug system
245. # [14:18] <Hixie> not mine :-)
246. # [14:19] <Hixie> da3d: <nav>, <aside>, <section>, <article>
247. # [14:19] <Hixie> da3d: are all secitons
248. # [14:19] <Hixie> Lachy: if you feel like it, feel free to go and invalidate and invalid bugs, too
249. # [14:19] <Hixie> s/and/any/
250. # [14:19] <Hixie> or mark duplicates
251. # [14:19] <Hixie> or mark unclear bugs NEEDSINFO
252. # [14:19] <Hixie> or whatever
253. # [14:19] <Hixie> it would actually be a lot of help
254. # [14:20] <Lachy> ok
255. # [14:21] * Joins: Unixmonkey (n=Unixmonk@ppp-69-208-150-144.dsl.ipltin.ameritech.net)
256. # [14:22] * Quits: nessy (n=nessy@203-214-73-15.dyn.iinet.net.au) ("This computer has gone to sleep")
257. # [14:23] <da3d> Hixie: Might just be my odd expectations then, I just think <nav> is a bit odd to include in that group :)
258. # [14:23] <Hixie> why?
259. # [14:23] <Hixie> i mean, it's a section, no? how is <nav> different from <aside> or <section>?
260. # [14:24] <da3d> I don't expect navbars to generally need or contain headings.
261. # [14:25] * Joins: Rik` (n=Rik`@pha75-2-81-57-187-57.fbx.proxad.net)
262. # [14:25] <da3d> I very much expect all <section>, <article> and <aside> to have at least one heading.
263. # [14:26] <Hixie> i expect a lot of <section> and <aside>s to not have headings
264. # [14:27] <Hixie> e.g. consider gmail; the main inbox is a <section>, the search area is a <section>, the chat box is an <aside>...
265. # [14:27] <Hixie> no headers
266. # [14:30] <da3d> I'd consider that to be generally bad authoring (unless it's an <aside> pull-out quote or something). Web apps like Gmail might be a bit different, but I'd at least some of those to have headings.
267. # [14:32] <Hixie> some, sure
268. # [14:32] <Hixie> most, maybe, even
269. # [14:32] <Lachy> Maybe it would help if the outliner generated default implied headings for sectioning elements like <nav>. Something like "Untitled Navigation Section", or simply "Navigation". I don't think this really needs to be defined though, just an implementation feature.
270. # [14:34] <Hixie> yeah
271. # [14:34] <Hixie> good idea
272. # [14:36] <Lachy> maybe an informative note in the spec about that for implementations might be useful
273. # [14:36] <Hixie> file a bug :-)
274. # [14:39] <da3d> I agree with that. I think a whole lot of people will be scratching their head at those untitled sections otherwise (like I did last night) :-)
275. # [14:45] <da3d> Unrelated question: Would it be considered odd/wrong/bad (in terms of semantics) to have a <footer>, but no <header>?
276. # [14:46] <Lachy> da3d, no. that's fine
277. # [14:46] <da3d> ok, thanks :)
278. # [14:52] <Lachy> Hixie, what is the outline supposed to look like for this case:
279. # [14:52] <Lachy> <body>
280. # [14:52] <Lachy> <h1>Heading A</h1>
281. # [14:52] <Lachy> <section>
282. # [14:52] <Lachy> No heading here.
283. # [14:52] <Lachy> <section>
284. # [14:52] <Lachy> <h1>Heading B</h1>
285. # [14:52] <Lachy> </section>
286. # [14:52] <Lachy> <h1>Heading C</h1>
287. # [14:52] <Lachy> </section>
288. # [14:52] <Lachy> http://james.html5.org/outliner.html is giving a strange result
289. # [14:55] <Hixie> um
290. # [14:55] * Hixie looks at the spec
291. # [14:56] <Lachy> Hixie, this step doesn't make sense:
292. # [14:56] <Lachy> If the top of the stack is a heading content element
293. # [14:56] <Lachy> Do nothing.
294. # [14:56] <Lachy> and since the spec says to execute the first relevant step, once a heading content element is at the top of the stack, then the rest of the steps are irrelevant
295. # [14:57] <Lachy> unless I'm misunderstanding something
296. # [14:58] <Hixie> it appears that the result is this:
297. # [14:58] <Hixie> Document
298. # [14:58] <Hixie> |
299. # [14:58] <Hixie> '-- Heading A
300. # [14:58] <Hixie> |
301. # [14:58] <Hixie> +-- Heading C
302. # [14:58] <Hixie> |
303. # [14:58] <Hixie> '-- Heading B
304. # [14:59] <Lachy> right, that's what I would intuitively expect. So it's a bug in jgraham's outliner
305. # [14:59] <Hixie> what does it do?
306. # [14:59] <Hixie> i couldn't get it to work for me at all
307. # [14:59] <Hixie> and you're misusnderstanding that step -- the step before it will apply when exiting the heading content element
308. # [15:00] <Lachy> #body
309. # [15:00] <Lachy> | Heading A
310. # [15:00] <Lachy> | Heading C
311. # [15:00] <Lachy> | Heading B
312. # [15:00] <Hixie> that's what i meant
313. # [15:01] <Hixie> that is correct
314. # [15:01] <Hixie> ignore what i pasted, i am too tired to draw ascii art apparently
315. # [15:02] <Hixie> it's nearly 6am
316. # [15:02] <Hixie> i should so totally sleep
317. # [15:02] <Hixie> nn
318. # [15:02] <Lachy> that doesn't make sense. Heading B is clearly after the Heading C section, so it shouldn't be a subsection of it
319. # [15:02] <Hixie> heading c's section is its parent "section> element
320. # [15:02] <Hixie> which comes bofer heading b's <section>
321. # [15:03] <Hixie> ok i'm really leaving now. good night :-)
322. # [15:03] <Lachy> The result should be the same as:
323. # [15:03] <Lachy> <body>
324. # [15:03] <Lachy> <h1>Heading A</h1>
325. # [15:03] <Lachy> <section>
326. # [15:03] <Lachy> test
327. # [15:03] <Lachy> <section>
328. # [15:03] <Lachy> <h1>Heading B</h1>
329. # [15:03] <Lachy> </section>
330. # [15:03] <Lachy> <section>
331. # [15:03] <Lachy> <h1>Heading C</h1>
332. # [15:03] <Lachy> </section>
333. # [15:03] <Lachy> </section>
334. # [15:03] <Lachy> ah, except I got Heading B and C switched
335. # [15:03] <Lachy> I will look into it and file a bug
336. # [15:05] <Lachy> (ah, wait, ignore my comment about them being switched.)
337. # [15:12] * Parts: da3d (n=opera@h11n1fls34o986.telia.com)
338. # [15:12] * Joins: da3d (n=opera@h11n1fls34o986.telia.com)
339. # [15:13] * Quits: Kalms (n=rasmuska@81.161.185.108)
340. # [15:13] * Joins: Kalms (n=rasmuska@81.161.185.108)
341. # [15:31] * Joins: erlehmann (n=erlehman@tmo-104-56.customers.d1-online.com)
342. # [15:39] * Quits: Amorphous (i=jan@unaffiliated/amorphous) (Read error: 60 (Operation timed out))
343. # [15:46] * Joins: kristallpirat (n=kristall@c-base/crew/kristall)
344. # [15:57] * Joins: Amorphous (i=jan@unaffiliated/amorphous)
345. # [16:18] * Quits: erlehmann (n=erlehman@tmo-104-56.customers.d1-online.com) ("Ex-Chat")
346. # [16:19] * Joins: erlehmann (n=erlehman@tmo-104-56.customers.d1-online.com)
347. # [16:21] * Joins: lazni (n=lazni@118.71.60.182)
348. # [16:32] * Quits: lazni (n=lazni@118.71.60.182) (Read error: 60 (Operation timed out))
349. # [16:40] * Joins: dglazkov (n=dglazkov@67.188.0.62)
350. # [16:41] * Joins: TabAtkins (n=chatzill@99.35.179.251)
351. # [16:44] * Joins: lazni (n=lazni@118.71.60.182)
352. # [16:50] * Joins: shepazu (n=schepers@31-33-83.wireless.csail.mit.edu)
353. # [16:56] * Quits: lazni (n=lazni@118.71.60.182) ("Leaving.")
354. # [16:59] * Joins: jacobolus (n=jacobolu@dhcp-0059871802-99-6d.client.student.harvard.edu)
355. # [17:06] * Joins: lazni (n=lazni@118.71.60.182)
356. # [17:07] * Disconnected
357. # [17:08] * Attempting to rejoin channel #whatwg
358. # [17:08] * Rejoined channel #whatwg
359. # [17:08] * Topic is 'WHATWG (HTML5) -- http://www.whatwg.org/ -- Logs: http://krijnhoetmer.nl/irc-logs/ -- Vennligst legg igjen din følelse av logikken ved døren, takk!'
360. # [17:08] * Set by Lachy on Wed Aug 26 20:57:24
361. # [17:09] <AryehGregor> Unfortunately, I enter my Ph.D. program on Tuesday, so my time might become extremely limited.
362. # [17:09] <AryehGregor> I'll see, though.
363. # [17:11] <TabAtkins> Congrats, then. ^_^
364. # [17:11] <TabAtkins> What are you going to be Professor of?
365. # [17:12] <AryehGregor> Pure mathematics.
366. # [17:12] <AryehGregor> Being a professor is a few years off yet, though.
367. # [17:12] <AryehGregor> . . . like eight or something. :/
368. # [17:13] <AryehGregor> (depending on how long I take to finish my thesis and postdocs)
369. # [17:13] <TabAtkins> Awesome! I'm sad that higher math doesn't fit my brain. -_-
370. # [17:14] * Joins: cardona507 (n=cardona5@c-67-180-160-250.hsd1.ca.comcast.net)
371. # [17:14] <TabAtkins> I can understand up to roughly the point where a math major would end you, but beyond that all I can do is smile and nod.
372. # [17:15] <AryehGregor> You know things like group theory and topology?
373. # [17:15] <TabAtkins> I know *of* them, and I like reading about them. But I don't understand them.
374. # [17:15] <AryehGregor> Heh.
375. # [17:15] <AryehGregor> That puts you well above most people, then. :)
376. # [17:15] <AryehGregor> Higher math is more of the same, just ever more specialized.
377. # [17:15] <TabAtkins> Yeah, I know, but "mathematician" was one of my goals as a kid. ^_^
378. # [17:16] <AryehGregor> Cool.
379. # [17:17] <TabAtkins> The funny thing is, I *know* there's a lot of group theory behind Haskell. I *can* understand that, as long as you keep it in programming terms. The moment an explanation veers off into actual math I get lost until I can say "Oh, so it's just function x being lifted and applying to two argument" or whatever.
380. # [17:17] <TabAtkins> And then I go write it into my lisp utilities, because it's usually useful.
381. # [17:18] <AryehGregor> If it's usually useful, then either you aren't learning pure mathematics or you do very strange kinds of programming. :)
382. # [17:18] <TabAtkins> There's no such thing as math that can't be applied. ^_^
383. # [17:19] * Joins: dglazkov_ (n=dglazkov@72.14.224.1)
384. # [17:19] <TabAtkins> But yeah, heavily functional programming is usually very strange too. ^_^
385. # [17:20] <TabAtkins> ... I've used smilies 5 times on this screenful. I have a problem.
386. # [17:20] <Dashiva> If you vary the smiley, it's less obvious
387. # [17:20] <AryehGregor> All math can be applied, if publishing a paper on it counts as applying it.
388. # [17:21] <TabAtkins> That doesn't count, Aryeh. ^_^
389. # [17:21] <TabAtkins> goddammit
390. # [17:21] <TabAtkins> Also: I'm not sure why the background-position syntax is so restrictive. There are several 2 and 3 token cases that are disallowed by the grammar but that are perfectly unambiguous.
391. # [17:22] <AryehGregor> Then I think you'd be pretty hard pressed to find many direct applications of, say, . . . I don't want to name anything lest someone come up with an application.
392. # [17:22] <AryehGregor> But I guess I'll try anyway. Any applications for Galois theory?
393. # [17:22] <TabAtkins> Pretty sure I've heard of something, yeah. Give me a sec to look it up.
394. # [17:22] <Dashiva> AryehGregor: Applications of it in other math, which is applied in other math, and at the end of the chain is a practical application
395. # [17:23] <AryehGregor> Yeah, probably.
396. # [17:23] <AryehGregor> I'm sure Galois theory is used in some branch of applied math.
397. # [17:23] <AryehGregor> Who cares, though, it's cool regardless.
398. # [17:24] <Dashiva> Yes, for a definition of cool most people will never comprehend :)
399. # [17:25] * Quits: krijnh (n=krijnhoe@ktk.xs4all.nl) (Read error: 110 (Connection timed out))
400. # [17:26] <AryehGregor> That just makes it cooler.
401. # [17:26] * AryehGregor permutes Dashiva and TabAtkins while leaving himself fixed
402. # [17:26] <TabAtkins> Hey man, I'm married.
403. # [17:26] <TabAtkins> I don't need to be permuted with anyone else.
404. # [17:27] * Quits: dglazkov (n=dglazkov@67.188.0.62) (Read error: 145 (Connection timed out))
405. # [17:27] * dglazkov_ is now known as dglazkov
406. # [17:27] <Dashiva> And I don't have a tab interface
407. # [17:27] <TabAtkins> Gah, so I can't find what I remember right now, but being as it can be used to address the solubility of polynomials, I'm going to guess that it's been applied to error-correcting codes.
408. # [17:28] * Joins: Kalms_ (n=rasmuska@81.161.185.108)
409. # [17:28] * Quits: Kalms (n=rasmuska@81.161.185.108) (Read error: 104 (Connection reset by peer))
410. # [17:28] * Kalms_ is now known as Kalms
411. # [17:29] <AryehGregor> CRC uses some group theory, but I don't think it uses any Galois theory.
412. # [17:29] <AryehGregor> CRC apparently uses the two-element field, so Galois theory is pretty much irrelevant.
413. # [17:29] * jgraham guesses solubility of polynomials is generally rather useful
414. # [17:29] <TabAtkins> Reed-Solomon?
415. # [17:30] <TabAtkins> I just saw a reference to that.
416. # [17:30] * Joins: gsnedders (n=gsnedder@86.167.19.36)
417. # [17:30] <TabAtkins> evening, gsnedders
418. # [17:30] <jgraham> Like I imagine there are physics applications where it is important to know whether there exist a closed form solution to certain polynomials or not
419. # [17:30] <AryehGregor> It's true that Galois theory can be used to prove facts about the solvability of polynomials over Q, so I grant it probably does have some indirect application.
420. # [17:30] <jgraham> But I can't think of any right now
421. # [17:30] <gsnedders> afternoon
422. # [17:30] <AryehGregor> Computational if nothing else.
423. # [17:30] <TabAtkins> Physics has, like, 2 closed form solutions.
424. # [17:31] <AryehGregor> TabAtkins, unless you count integrals as closed-form! Then it has at least 5.
425. # [17:31] <AryehGregor> If you count systems of differential equations as closed-form, *then* you're in business. :P
426. # [17:31] <TabAtkins> That's iffy, but okay. diffeqs are *not* closed form. ^_^
427. # [17:32] <AryehGregor> Well, it's mathematics. You're allowed to use whatever bizarre and confusing definitions you like.
428. # [17:32] <TabAtkins> Ah, but this is physics. And at the end of the day you need to shut up and compute.
429. # [17:32] <TabAtkins> And all your fancy-pants abstract structures fall to the simple expedient of iteration and ever-more computing power.
430. # [17:33] * gsnedders wonders how many to meet up with next week back at home
431. # [17:33] <Dashiva> Numerical methods win the day again
432. # [17:34] <TabAtkins> Moore is the patron saint of numerical methods.
433. # [17:34] <jgraham> Numerical methods suck. always.
434. # [17:34] <AryehGregor> Yes.
435. # [17:34] <AryehGregor> Anyone who says numerical methods beat analytic methods has no idea what they're talking about.
436. # [17:34] <TabAtkins> Go solve the 3-body problem, then. ^_^
437. # [17:35] <AryehGregor> Numerical methods are what you use when you have no other options.
438. # [17:35] <TabAtkins> Well, yeah. Which is usually the case.
439. # [17:35] <TabAtkins> If you have a cheap analytic method, of course you use it. But they're often rare.
440. # [17:35] <AryehGregor> TabAtkins, numerical methods are useless on the general three-body problem. Some solutions are chaotic. Any finite-precision numerical method will give you useless results for those cases.
441. # [17:35] <TabAtkins> On the other hand, there's totally a supercomputer in the other room...
442. # [17:35] <AryehGregor> (AFAIK)
443. # [17:36] <AryehGregor> I have access to supercomputers at Courant, I think, if I need them.
444. # [17:36] <TabAtkins> AryehGregor: Indeed, when attempting to forecast out too far.
445. # [17:36] <jgraham> Indeed, you often have to use them. But they suck to implement are generally have problems that you don't want to deal with
446. # [17:36] <AryehGregor> TabAtkins, "too far" possibly meaning "any useful distance"
447. # [17:36] <AryehGregor> .
448. # [17:36] <TabAtkins> (though knowing when 'too far' is is probably itself uncomputable)
449. # [17:36] <AryehGregor> No, it's not uncomputable.
450. # [17:36] <AryehGregor> You can compute the doubling time for errors.
451. # [17:37] <AryehGregor> Of course, in the chaotic cases, your input data is worthless, so no solution is useful.
452. # [17:37] <TabAtkins> Hm, that's true. I bluffed my way through numerical methods, actually.
453. # [17:37] <AryehGregor> I never took any of that applied nonsense.
454. # [17:37] <jgraham> TabAtkins: You did a maths degree?
455. # [17:37] <AryehGregor> But I've heard a bit about chaos theory.
456. # [17:37] <TabAtkins> jgraham: at my university, a comp sci degree is a single class short of a math minor.
457. # [17:38] <TabAtkins> so of course i took that class and got myself another feather in the cap.
458. # [17:38] <jgraham> TabAtkins: Ah
459. # [17:38] <TabAtkins> I think the class was Calc3. Dunno why. That stuff was easy, and useful for comp sci.
460. # [17:39] * gsnedders sighs at unis
461. # [17:41] <TabAtkins> Yeah, good luck with that, dude. You planning on attending anywhere?
462. # [17:42] <gsnedders> Somewhere, somehow.
463. # [17:42] <TabAtkins> Just trick them into paying for you. It worked for me.
464. # [17:42] <gsnedders> That's not the problem.
465. # [17:43] <gsnedders> The problem is having been ill a lot and having only a C in maths, where I need at least a B for most good places.
466. # [17:43] * jgraham notes that the UK university system is not yet insane in the amount of money it charges
467. # [17:44] <TabAtkins> Oh, right, you guys have to deal with grades from high school actually mattering.
468. # [17:44] <TabAtkins> Come to the US instead. ^_^
469. # [17:44] <gsnedders> Then I hit the money issue. :P
470. # [17:44] <TabAtkins> Heh.
471. # [17:45] * Quits: lazni (n=lazni@118.71.60.182) ("Leaving.")
472. # [17:46] * Quits: gsnedders (n=gsnedder@86.167.19.36) (Read error: 131 (Connection reset by peer))
473. # [17:46] * Joins: gsnedders_ (n=gsnedder@host86-167-19-36.range86-167.btcentralplus.com)
474. # [17:46] <TabAtkins> I'm lucky that grades matter less in most places over here. I just don't mesh well with a lot of classes, so I ended up with bad grades anywhere that placed more emphasis on homework than tests.
475. # [17:46] * gsnedders_ is now known as gsnedders
476. # [17:46] <AryehGregor> jgraham, I went to the City College of New York. It was like $2000 a term.
477. # [17:46] <AryehGregor> . . . only for state residents, of course.
478. # [17:46] <TabAtkins> Yeah, about to say. Lucky you.
479. # [17:47] <gsnedders> AryehGregor: English unis are £3,145 per year
480. # [17:47] <AryehGregor> All of them?
481. # [17:47] <gsnedders> AryehGregor: yes
482. # [17:47] <TabAtkins> That's about what I paid after conversion, gsnedders.
483. # [17:47] <TabAtkins> Wait, per *year*?
484. # [17:47] <AryehGregor> Crazy socialists. :P
485. # [17:47] <AryehGregor> Here they're all different prices.
486. # [17:47] <jgraham> Strictly that is just an upper limit
487. # [17:47] <gsnedders> AryehGregor: Scottish unis, for Scottish and EU (but not England, Wales, and NI) citizens are £0 per year
488. # [17:47] <AryehGregor> Harvard is free for undergraduates these days, I think, unless you're rich.
489. # [17:48] <TabAtkins> I was probably 3k-4k per semester.
490. # [17:48] <gsnedders> jgraham: Oh, is it? I never knew. Everywhere charges it though. :P
491. # [17:48] <AryehGregor> gsnedders, if you go to a public university and you're needy and you're a good student, you can usually get enough scholarships to go for free and make a profit, AFAICT.
492. # [17:48] <AryehGregor> (in America)
493. # [17:48] <AryehGregor> At least I had a friend who did that.
494. # [17:48] <TabAtkins> I was doing that for 2 years until I lost the scholarship.
495. # [17:49] <gsnedders> AryehGregor: Then I have to go and sit SATs (and arrange to sit them) then fill out x application forms…
496. # [17:49] <TabAtkins> SATs is easy.
497. # [17:49] <jgraham> Scholarship based systems are generally not good. They create too much pressure
498. # [17:49] <gsnedders> TabAtkins: I still need to arrange and do them, which is effort. :P
499. # [17:49] <AryehGregor> Not for good students who can keep up the terms reliably.
500. # [17:50] <jgraham> Which doesn't always have a positive impact
501. # [17:50] <gsnedders> AryehGregor: It's still extra pressure on them
502. # [17:50] * gsnedders waits to lose wifi again
503. # [17:50] <gsnedders> *loose
504. # [17:50] <gsnedders> *lose
505. # [17:50] * gsnedders can spk English, honest
506. # [17:50] <gsnedders> :P
507. # [17:51] <AryehGregor> Well, I don't know. The alternative is to be given money/services for free regardless of whether you're actually attending class and doing your work?
508. # [17:52] <jgraham> Yep. You loose out in the long term by getting a shitty degree and not being able to get a job
509. # [17:53] <jgraham> (actually the only person I knew at university who dropped out had previously been on a scholarship to Eton. He was bright but burned out by 18)
510. # [17:54] <gsnedders> (For those keeping track, basic answer from Southampton is to phone up the new CS admissions tutor once the post has been appointed after the current guy retiring this month, and try and work out what comprimise can be made for the maths)
511. # [17:55] * Joins: dglazkov_ (n=dglazkov@c-67-188-0-62.hsd1.ca.comcast.net)
512. # [17:56] * TabAtkins can't understand how he could possibly have written this function to take up over two screens, when he just rewrote it to be about half a screen.
513. # [17:57] * AryehGregor likes the use of "screen" as a unit of measurement
514. # [17:57] <jgraham> gsnedders: You are planning on S'hampton + G'gow + E'brugh?
515. # [17:58] <TabAtkins> That's the most important measure of code. Once a function gets longer than a screen, it's time to refactor. You can't really *understand* a function that you can't see all at once.
516. # [17:58] <gsnedders> jgraham: + 2 other places
517. # [17:58] <gsnedders> jgraham: But yeah, that's my current rough plan
518. # [17:58] <AryehGregor> TabAtkins, yeah, but . . . different people have totally different screen sizes.
519. # [17:59] <TabAtkins> AryehGregor: Yeah, that's why you get as large a screen as possible. That way you're not as limited.
520. # [17:59] <jgraham> gsnedders: Do you know which two?
521. # [17:59] <gsnedders> jgraham: no
522. # [18:01] <TabAtkins> If you would like, you can use "head" as a unit of measurement. If your code is bigger than your head (pressed against the screen), it needs to be refactored. It amounts to roughly the same thing.
523. # [18:01] * gsnedders hears ship horns from the docks
524. # [18:02] * jgraham wishes for keybindings on OS-X for "make this window take up the full height and left two thirds of my screen"
525. # [18:02] * gsnedders does too
526. # [18:07] <gsnedders> jgraham: I was talking to several people and several of them were still saying I ought to apply for Oxbridge, if I manage to do something to get my maths up to scratch
527. # [18:09] <gsnedders> Basically, I'm not sure how much point there is in applying to many places in England as unless I got into either Oxbridge or Southampton, I'd rather go to Edinburgh (which I should get into on the basis of Highers alone)
528. # [18:09] <AryehGregor> TabAtkins, I have a big head.
529. # [18:09] <AryehGregor> I prefer to use LOC, personally. ;)
530. # [18:09] <TabAtkins> Then you get to write bigger functions!
531. # [18:09] <gsnedders> Seeming I have five courses to apply for, in some ways it seems that I may as well in some ways
532. # [18:10] <TabAtkins> I think LOC has been poisoned by management techniques that imply that more LOC = more betta'.
533. # [18:10] * Quits: gsnedders (n=gsnedder@host86-167-19-36.range86-167.btcentralplus.com) (Remote closed the connection)
534. # [18:10] * Joins: gsnedders (n=gsnedder@host86-167-19-36.range86-167.btcentralplus.com)
535. # [18:11] <TabAtkins> By using a unit of measurement that is inherently personal, it personalizes the rule.
536. # [18:11] <TabAtkins> It also doesn't require counting.
537. # [18:12] <jmb> gsnedders: who did you talk to at soton, ooi?
538. # [18:12] <TabAtkins> You can't really talk about how many screens long a program is, after all. It just doesn't make much sense. But saying "each function should fit on the screen" promotes the ideal of refactoring and splitting out subfunctions rather than duplicating code.
539. # [18:12] <gsnedders> jmb: mw and Hugh… Davis
540. # [18:13] <jmb> right
541. # [18:13] * gsnedders wonders who jmb is
542. # [18:13] <jmb> mw is/was my phd supervisor
543. # [18:13] <gsnedders> ah
544. # [18:14] <jmb> and hugh's head of that research group :)
545. # [18:14] * Quits: dglazkov (n=dglazkov@72.14.224.1) (Read error: 110 (Connection timed out))
546. # [18:14] * dglazkov_ is now known as dglazkov
547. # [18:18] * TabAtkins has finally gotten his homepage's cpu usage under control.
548. # [18:19] * gsnedders grumbles about his intention of getting a bus at 8:10 on Wednesday
549. # [18:19] <TabAtkins> I can now add more elements to it, confident that I'll actually be able to test the thing without locking up ff
550. # [18:20] <AryehGregor> Use Chrome and you don't have to worry so much about locking up, in my experience. :P
551. # [18:21] <TabAtkins> For some strange reason, Chrome's not doing it right.
552. # [18:21] * Joins: karlcow (n=karl@nerval.la-grange.net)
553. # [18:21] <TabAtkins> Though generally, yeah.
554. # [18:21] <TabAtkins> View www.xanthir.com in FF, then in Chrome.
555. # [18:22] <AryehGregor> Haha.
556. # [18:22] <AryehGregor> Cute.
557. # [18:22] <AryehGregor> Very video-gamey.
558. # [18:22] <TabAtkins> Well, they're the mario clouds, so yeah.
559. # [18:23] <TabAtkins> I'm thinking of making them do a pixellated move, rather than a smooth transition.
560. # [18:23] <TabAtkins> but anyway! Chrome, for some reason, won't adjust their top.
561. # [18:23] <AryehGregor> Hmm, I see.
562. # [18:24] <TabAtkins> I wonder if it doesn't like receiving an untyped value for top?
563. # [18:24] <AryehGregor> The top rule isn't showing up.
564. # [18:24] <AryehGregor> jQuery, you mean?
565. # [18:24] <Philip`> Hixie: That rich snippets thing seems a pretty unhelpful tool for testing the RDFa parser - I'd want something that simply dumps triples from arbitrary RDFa, so I can add its output to http://philip.html5.org/demos/rdfa/results.html
566. # [18:24] * Joins: svl_ (n=me@g227035173.adsl.alicedsl.de)
567. # [18:25] <gsnedders> hmm, TfL website says to take Bakerloo/Victoria to get from Waterloo to King's Cross. Odd.
568. # [18:25] <TabAtkins> Yeah, I set the top in the jQuery.
569. # [18:25] <TabAtkins> I'm wondering if Chrome just doesn't like it for some reason. But I send an untyped value to left as well.
570. # [18:26] <AryehGregor> I'd expect it's a jQuery bug.
571. # [18:26] <AryehGregor> . . . I never played much Mario.
572. # [18:26] <AryehGregor> Except Super Mario RPG.
573. # [18:26] <AryehGregor> On the SNES.
574. # [18:26] <AryehGregor> <3
575. # [18:27] <TabAtkins> <3 indeed. I've bought that game 3 times now. Twice on SNES, once on Wii.
576. # [18:27] <AryehGregor> var scale = 1+Math.random();
577. # [18:27] <AryehGregor> top = random(0, screenheight/2 - 96),
578. # [18:27] <AryehGregor> Do you want the ";" to be a ","?
579. # [18:27] <TabAtkins> ...
580. # [18:27] <TabAtkins> yes i do.
581. # [18:28] <TabAtkins> Ah, all fixed.
582. # [18:28] <AryehGregor> I wonder what the underlying JS interop problem was in V8 vs. TraceMonkey.
583. # [18:29] <TabAtkins> I dunno. I wonder which one that was a bug in?
584. # [18:29] <Philip`> Hixie: (This thing just complains I don't have typeof, and if I do have typeof then it doesn't say what namespace any predicates are in)
585. # [18:36] <jgraham> gsnedders: That sounds sensible to me
586. # [18:36] <jgraham> (the TfL advice)
587. # [18:37] <jgraham> IIRC it is the fewest stops (so probably the quickest) but looks longer on the tube map
588. # [18:37] <jgraham> Also, it avoids the northern line which is always good
589. # [18:42] <gsnedders> jgraham: It's geographically longer too
590. # [18:43] <gsnedders> I'm just somewhat surprised that the geographically longer route is quicker even if it does have fewer stops
591. # [18:43] <gsnedders> and IIRC it'll take longer to walk at KGX
592. # [18:44] <gsnedders> As I think Victoria is out on a limb there
593. # [18:45] <jgraham> gsnedders: It won't take longer to walk
594. # [18:45] <jgraham> At least not substantially longer
595. # [18:45] <gsnedders> jgraham: It's been a while since I've been on the Victoria
596. # [18:46] <jgraham> (unless I am confusing it with the picadilly line at KGX which is possible)
597. # [18:46] <gsnedders> Piccadilly is very close to Northern, which is no distance at all
598. # [18:47] <gsnedders> There's extra, long, tunnels to walk down to get to Victoria that'll take a min or two
599. # [18:47] <gsnedders> But I have over an hour to make it across :P
600. # [18:47] <jgraham> Anyway the point is that that is a route that I often use to go Waterloo to Kings Cross and I believe it to be one of the quicker options
601. # [18:48] * gsnedders generally doesn't go to Waterloo :P
602. # [18:48] <jgraham> As quick as any other option within the uncertainty limit of train arrival times
603. # [18:48] <jcranmer> there's, extra, commas, in, that, sentence
604. # [18:48] <gsnedders> jcranmer: In what I wrote? Removing them changes the meaning of it.
605. # [18:49] <gsnedders> jcranmer: Removing them implies the "extra" describes the "long"
606. # [18:49] <jcranmer> that's what I thought you had intended
607. # [18:49] <gsnedders> Not that they are both extra and they are long
608. # [18:49] <gsnedders> I said what I intended :)
609. # [18:50] <AryehGregor> Tra-la-la.
610. # [18:51] <AryehGregor> $ wget -qO- --user-agent=Simetrical http://www.wikipedia.org | head -n 1
611. # [18:51] <AryehGregor> <!doctype html>
612. # [18:51] <AryehGregor> We need bugs 640 and 641 fixed in the validator . . . http://validator.nu/?doc=http://www.wikipedia.org
613. # [18:52] <Philip`> How ugly :-(
614. # [18:52] * Philip` prefers <!DOCTYPE html>
615. # [18:52] * gsnedders uses the lowercase form, because it's less effort to type
616. # [18:52] <AryehGregor> nonsense. uppercase letters are the devil.
617. # [18:54] <Philip`> gsnedders: How is it less effort? You have to hold down the shift key for <!, and you just keep it down a bit longer
618. # [18:55] <AryehGregor> Of course, www.wikipedia.org is kind of trivial, since it's just a static HTML page.
619. # [18:55] * Quits: cardona507 (n=cardona5@c-67-180-160-250.hsd1.ca.comcast.net)
620. # [18:55] <AryehGregor> Converting en.wikipedia.org is more interesting. I have no idea when the next software deployment will be.
621. # [18:56] * Joins: cardona507 (n=cardona5@c-67-180-160-250.hsd1.ca.comcast.net)
622. # [18:57] <Philip`> AryehGregor: Is there any change on www. other than the doctype?
623. # [18:57] <AryehGregor> $ gzip < /tmp/wp.html > /tmp/wp1.html.gz
624. # [18:57] <AryehGregor> $ sed 's/<!doctype/<!DOCTYPE/' < /tmp/wp.html | gzip > /tmp/wp2.html.gz
625. # [18:57] <AryehGregor> $ ls -l /tmp/wp*
626. # [18:57] <AryehGregor> -rw-r--r-- 1 aryeh aryeh 12004 2009-09-06 12:51 /tmp/wp1.html.gz
627. # [18:57] <AryehGregor> -rw-r--r-- 1 aryeh aryeh 12007 2009-09-06 12:51 /tmp/wp2.html.gz
628. # [18:57] <AryehGregor> -rw-r--r-- 1 aryeh aryeh 46418 2009-09-06 12:50 /tmp/wp.html
629. # [18:57] <AryehGregor> <!doctype is three bytes shorter than <!DOCTYPE when gzipped. I rest my case.
630. # [18:57] <AryehGregor> Philip`, yeah, a few trivial things like removing <big>.
631. # [18:57] <AryehGregor> And dropping type="text/css" and such.
632. # [18:58] <AryehGregor> The wiki itself will be deploying a few actual features, like HTML 5 <input>.
633. # [18:58] <AryehGregor> But not many. There aren't many new HTML 5 features that are usable in practice yet for authors.
634. # [18:58] <AryehGregor> Suggestions are welcome.
635. # [18:58] <AryehGregor> http://www.mediawiki.org/wiki/HTML_5 <-- Add here if you like.
636. # [18:59] * Joins: zdobersek (n=zan@cpe-92-37-68-231.dynamic.amis.net)
637. # [18:59] <Dashiva> AryehGregor: Is it still shorter if you have a title containing DOCTYPE again right afterwards?
638. # [18:59] <AryehGregor> Dashiva, doubtful.
639. # [18:59] <AryehGregor> Really doubtful if you have a few.
640. # [18:59] <AryehGregor> But maybe.
641. # [18:59] <Philip`> What if you use <!Doctype>?
642. # [18:59] <AryehGregor> The reason is that there tend to be a lot more lowercase letters in English text than uppercase.
643. # [19:00] <AryehGregor> So lowercase strings get compressed better.
644. # [19:00] <TabAtkins> I'll be happy when the openid patch gets integratd.
645. # [19:00] <AryehGregor> Philip`, 12005, so still a byte larger.
646. # [19:00] * Quits: gsnedders (n=gsnedder@host86-167-19-36.range86-167.btcentralplus.com) (Read error: 54 (Connection reset by peer))
647. # [19:00] <AryehGregor> TabAtkins, there is no OpenID patch. There's an extension, but nobody's really proposed it for deployment by Wikimedia.
648. # [19:00] <AryehGregor> It would need some careful thought and has gotten none.
649. # [19:01] <Philip`> AryehGregor: I don't think English text is really an important part of www.wikipedia.org, given that most of it is markup and JS and CSS, and most of the rest is non-English :-p
650. # [19:01] <TabAtkins> Bah. I want to use my openid in more places.
651. # [19:01] <AryehGregor> Wikimedia is really understaffed in the tech department, sadly, when it comes to senior people who can review code.
652. # [19:01] <TabAtkins> I spent like a whole hour setting it up. ^_^
653. # [19:01] * Joins: gsnedders (n=gsnedder@host86-167-19-36.range86-167.btcentralplus.com)
654. # [19:01] <AryehGregor> Philip`, the markup *is* the English text. :) Most pages end up being more markup (ergo, mostly-lowercase English-oid) than actual content.
655. # [19:02] <Philip`> AryehGregor: It should follow the Wikipedia approach and let anybody commit new code that gets automatically deployed
656. # [19:02] <AryehGregor> TabAtkins, realistically, almost the only thing OpenID lets you do is avoid having to choose a password, right?
657. # [19:02] <AryehGregor> Philip`, uh-huh.
658. # [19:02] <AryehGregor> We could at least use a DVCS, though.
659. # [19:02] <AryehGregor> We'll probably switch sometime.
660. # [19:02] <TabAtkins> Aryeh, yeah.
661. # [19:03] <TabAtkins> The site queries my openid server, which requests my login and then reports back to the site.
662. # [19:03] <TabAtkins> Though I also transmit a goodly bit of identifying information with my openid, so most profiles can be autofilled.
663. # [19:04] <AryehGregor> Wikipedia only requires username and password for new registrations.
664. # [19:04] <AryehGregor> Even e-mail is optional.
665. # [19:04] <AryehGregor> (although recommended)
666. # [19:04] <TabAtkins> The CSSWG wiki automatically knew my name, which was fun. ^_^ Then I had to sign up for a normal account anyway because the openid accounts weren't properly integrated into the editor page.
667. # [19:04] <AryehGregor> The editor page?
668. # [19:05] <AryehGregor> Passwords of any kind are terrible anyway. Why hasn't PKA caught on in the wider world as a substitute?
669. # [19:05] <AryehGregor> Like nobody uses it. :(
670. # [19:05] <Dashiva> Because it's a mess
671. # [19:06] <TabAtkins> Also: keys are annoying to lug around.
672. # [19:06] <AryehGregor> In what way is it a mess?
673. # [19:06] <TabAtkins> AryehGregor: fantasai couldn't find my account to give me editting privileges.
674. # [19:06] <AryehGregor> TabAtkins, only because nobody's marketing some standard USB-based thing you could stick in a wallet.
675. # [19:06] <Dashiva> AryehGregor: In what way isn't it?
676. # [19:07] <AryehGregor> Dashiva, . . . that's a kind of broad question.
677. # [19:07] <Dashiva> Take your USB thing
678. # [19:07] <Dashiva> First you'd have to password protect it, so that's a password to remember
679. # [19:07] <Dashiva> What do you do if you lose it?
680. # [19:07] <TabAtkins> AryehGregor: but then you're relying on a fob containing information that you can't possibly memorize as an authent mechanism.
681. # [19:07] <Dashiva> How do you protect against a malicious application sniffing all your keys when you use an unknown computer?
682. # [19:07] <AryehGregor> Dashiva, one password. For everything. You can use the same public key *everywhere*. That's the beauty. Plus you could do biometrics or whatever instead, if you felt like it.
683. # [19:07] <Philip`> Dashiva: It should use biometric identification
684. # [19:08] <Dashiva> Biometrics is flawed
685. # [19:08] <AryehGregor> Dashiva, you'd keep a copy in your bank vault that you could reimage if you lost it.
686. # [19:08] <gsnedders> AryehGregor: What happens when someone takes that USB key?
687. # [19:08] <Dashiva> AryehGregor: And normal people are going to do this?
688. # [19:08] <TabAtkins> Biometrics just means that they'll cut off your finger when they mug you.
689. # [19:08] <Philip`> Dashiva: The USB key can contain a small pin and a receptacle for blood DNA analysis
690. # [19:08] <AryehGregor> Dashiva, and the protocol would be designed so the computer doesn't see the private key. It just passes the data to the device, which encrypts it.
691. # [19:08] <Dashiva> I don't even have a bank vault
692. # [19:08] <AryehGregor> Philip`, biometrics aren't a magic bullet. They're just as susceptible to forgery as other things.
693. # [19:08] <TabAtkins> I don't want criminals to have *any* reason to do invasive surgery on me.
694. # [19:09] <Philip`> AryehGregor: I have one password, for everything, I can use it everywhere, and I don't have to worry about losing anything
695. # [19:09] <Dashiva> AryehGregor: And what if you don't want everyone to know who you are everywhere?
696. # [19:09] <Philip`> s/I/and I/
697. # [19:09] <gsnedders> Dashiva: Yes Magus…
698. # [19:09] <Dashiva> You'd need several different USB keys?
699. # [19:09] <gsnedders> *Magnus
700. # [19:09] <AryehGregor> gsnedders, make it password-protected or such.
701. # [19:09] <Dashiva> gsnedders: This isn't my secret nick :P
702. # [19:09] <Philip`> AryehGregor: so it's much easier than this hypothetical new world yet has all the benefits :-)
703. # [19:09] <gsnedders> Dashiva: Nor is this mine :P
704. # [19:09] <AryehGregor> Dashiva, yes, people will do it, if it's what they use to authenticate to everything. They do it for their property deeds and so forth.
705. # [19:10] <AryehGregor> Philip`, if you sign up to my site and I remember your password, I now can log in as you to any site. Congrats. Or if I MITM you when you log in to any unsecured site.
706. # [19:10] <Dashiva> No, people do not sign their porn site subscriptions
707. # [19:10] <AryehGregor> Dashiva, a key is what, 8192 bits? You could store 10,000 pregenerated ones on one device for barely more cost than one. Have it use a different actual key for each site if you like, but the same device.
708. # [19:11] <AryehGregor> Dashiva, what?
709. # [19:11] <AryehGregor> ("what?" is responding to that last line)
710. # [19:11] <Dashiva> I.e. eople will want different keys for different kinds of activities
711. # [19:11] <Dashiva> *people
712. # [19:12] <gsnedders> AryehGregor: Dashiva makes his living through pr0n
713. # [19:12] <AryehGregor> So have the device store thousands of keys. Let the negotiation protocol ask it "what's your public key for <unique stable site id>"?
714. # [19:12] <TabAtkins> AryehGregor: one of the benefits of openid - you get the single-password login, but without the possibility of a malicious server-owner remembering your info.
715. # [19:12] <Philip`> AryehGregor: As far as I'm aware, nobody running a site has ever stolen my password, so I have no reason to be concerned about that
716. # [19:12] <Dashiva> TabAtkins: Yeah, you just get a massive phishing opportunity instead
717. # [19:12] <AryehGregor> TabAtkins, yes. That slightly improves one small aspect of the problem.
718. # [19:12] * Quits: dglazkov (n=dglazkov@c-67-188-0-62.hsd1.ca.comcast.net)
719. # [19:12] <AryehGregor> Philip`, I've seen it happen. It's pretty trivial.
720. # [19:12] <TabAtkins> Dashiva: Eh, phishing is always going to be a problem forever.
721. # [19:13] <Philip`> (Actually I don't really have only one password, I have half a dozen and I type them all into every login form until I guess the right one)
722. # [19:13] <Dashiva> TabAtkins: OpenID makes it easier, though
723. # [19:13] <AryehGregor> Philip`, a guy lured the moderator of a big forum to sign up to his site, logged in to the forum, and caused havoc.
724. # [19:13] <TabAtkins> Dashiva: Depends. It's easy to personalize the openid page so it's harder to duplicate and thus phish.
725. # [19:14] <TabAtkins> For example, my bank site shows me a picture and phrase that I chose so I know that someone else isn't impersonating them.
726. # [19:14] <Dashiva> TabAtkins: And it's easy to say "We are currently performing maintenance, your personalized data will be back soon. Please use thois temporary form instead."
727. # [19:14] <AryehGregor> The only way to stop automated theft of authentication is to make authentication impossible to steal automatically.
728. # [19:14] <AryehGregor> Some places do this in weak ways like having a widget that displays pseudorandom numbers that change every minute.
729. # [19:14] <Dashiva> Or just MITM the whole process
730. # [19:14] <TabAtkins> Dashiva: At that point, aren't we in the same situation with individual logins?
731. # [19:15] <Dashiva> TabAtkins: Sure, but with OpenID you steal all logins in one fell swoop
732. # [19:15] <TabAtkins> Considering the password security of the average person, I'm not sure that's much of a difference.
733. # [19:15] * Joins: erikvvold (n=erikvvol@96.49.192.204)
734. # [19:15] <AryehGregor> PKA really is the answer here. For all authentication.
735. # [19:15] <Dashiva> Not much, but it's a step in the wrong direction
736. # [19:15] <AryehGregor> It's more secure than physical keys and so on as well.
737. # [19:16] <AryehGregor> Dashiva, you don't get security by giving up on the issue and just avoiding single points of failure. You get security by making sure there's only one point of failure, and it's *strong*.
738. # [19:16] <Dashiva> AryehGregor: Who said anything about giving up?
739. # [19:16] <AryehGregor> Well, so what's your proposed solution?
740. # [19:17] <Dashiva> Get people to use password managers
741. # [19:17] <TabAtkins> Even considering the phishing possibilities, I consider the strengths of having a single password that you can make strong and use everywhere to be significant enough to be worthwhile.
742. # [19:17] <AryehGregor> Dashiva, and if you're using a public computer?
743. # [19:17] <Dashiva> Strong passwords are overrated, even a medium strength password is more than enough
744. # [19:17] <TabAtkins> Dashiva: a password manager-based solution is effectively identical to PKA
745. # [19:18] <AryehGregor> TabAtkins, strength of the password doesn't help much. It can be pretty weak and still infeasible to crack.
746. # [19:18] <Dashiva> TabAtkins: No, it saves you from carrying around a thing and storing things in vaults
747. # [19:18] <AryehGregor> TabAtkins, no, it's markedly inferior.
748. # [19:18] <AryehGregor> Dashiva, a weaker PKA scheme would just keep the key on your computer.
749. # [19:18] <AryehGregor> Pretty much identical to a password manager, but secure against replay attacks and so on.
750. # [19:18] <Dashiva> Security against replay attacks is unrelated
751. # [19:19] <TabAtkins> Dashiva: I'm not sure how it's different. Either way you have an identifier that you don't remember, and instead store elsewhere.
752. # [19:19] <Dashiva> If you change the auth procedure to support PKA you can also change it to use passwords with nonces etc
753. # [19:19] <AryehGregor> Both of those schemes are useless if you're using an untrusted computer, or not using a computer at all, or your computer gets hacked.
754. # [19:19] <TabAtkins> All schemes are useless if you get hacked, of course.
755. # [19:19] <AryehGregor> Dashiva, yeah, okay, granted. They're fairly comparable in that respect, I guess.
756. # [19:19] <AryehGregor> TabAtkins, nope.
757. # [19:19] <AryehGregor> Not my proposal.
758. # [19:19] <AryehGregor> Since your own computer wouldn't know your private key.
759. # [19:19] <Dashiva> TabAtkins: It's easier to press "enter my password" in the browser than to fiddle with a thing
760. # [19:20] <AryehGregor> I don't think having private keys on your computer is much more useful than a password manager that autogenerates strong passwords.
761. # [19:20] <TabAtkins> Unless your fob stores the identifying keys from your programs as well, a trojan can mitm the program you're trying to log into.
762. # [19:20] <Dashiva> Imagine how much better off we would be if domains weren't backwards
763. # [19:20] <TabAtkins> That would be nice, yeah.
764. # [19:20] <AryehGregor> TabAtkins, and get what from it? Nothing. You don't hand out any reusable info when using PKA.
765. # [19:21] <AryehGregor> It's all one-shot.
766. # [19:21] <Lachy> The problem with password manager based solutions is that if you use lots of random passwords on the basis that you don't have to remember them all, you're reliant on your ability to always have the password manager available and never lose it
767. # [19:21] <TabAtkins> No more www.yourbank.com.shadyrussiansite.com.
768. # [19:21] <AryehGregor> Dashiva, MediaWiki's externallinks table has a whole extra column with the URL having domain name reversed, so as to allow efficient prefix-based lookups.
769. # [19:22] <Dashiva> Lachy: This is of course true, but then you can always have a USB pin with the password manager, as a minor tribute to PKA ;)
770. # [19:22] * Quits: cardona507 (n=cardona5@c-67-180-160-250.hsd1.ca.comcast.net)
771. # [19:22] <TabAtkins> PortableApps!
772. # [19:22] <AryehGregor> Lachy, yes. In that way, having a small physical object is actually more convenient. Keep it in your wallet.
773. # [19:22] <Lachy> also, password managers (especially Firefox's) store passwords unencrypted, so if the password file is compromised, everything is
774. # [19:22] <AryehGregor> It's also much more secure.
775. # [19:22] <AryehGregor> Lachy, Firefox has a master password option, which I assume encrypts the stored passwords, no?
776. # [19:22] <Lachy> 1Password and Safari use OSX's keychain, which is a little more secure
777. # [19:23] <Lachy> AryehGregor, yes, but at the expense of invonvenience of having to enter it all the time, negating the major benefit of a password manager
778. # [19:23] <Lachy> *inconvenience
779. # [19:23] <AryehGregor> Well, you only have to remember one password.
780. # [19:24] <AryehGregor> The keychain improves things how?
781. # [19:24] * Quits: myakura (n=myakura@p4115-ipbf3106marunouchi.tokyo.ocn.ne.jp) ("Leaving...")
782. # [19:24] <Lachy> the keychain stores passwords in encrypted form. Gaining access to the computer doesn't let anyone read the passwords without knowing the user's system login password
783. # [19:25] * Quits: erikvold (n=erikvvol@96.49.192.204) (Read error: 110 (Connection timed out))
784. # [19:25] <Lachy> although, once the password manager fills in the form on a page to login, it still remains possible to intercept the password at that point
785. # [19:26] <AryehGregor> Ah, so it decrypts them using the login info.
786. # [19:26] <Dashiva> The next solution would be minification so that the USB thing becomes an entire computer, you just attach it to a monitor, network and input devices
787. # [19:27] <AryehGregor> You could get the same benefit using home-directory encryption, which has lots of other nice benefits if an attacker gets hold of your computer. Or even full-disk encryption. Although those tend to have some performance overhead.
788. # [19:27] <AryehGregor> Dashiva, no. You want to keep your sensitive data in a device that's as simple to secure as possible. It should be impossible to run arbitrary code on it.
789. # [19:27] <TabAtkins> I don't notice too much problems with full-disk encryption, luckily. Large file transfers take a bit longer, but for normal use it's most unnoticeable.
790. # [19:27] <AryehGregor> Everything should be unchangeable down to the firmware.
791. # [19:28] <Dashiva> AryehGregor: You could have that device as part of it
792. # [19:28] <AryehGregor> TabAtkins, depends on your CPU, I guess. Netbooks will likely notice a difference.
793. # [19:28] <TabAtkins> Yeah, probably.
794. # [19:28] <AryehGregor> Dashiva, if you can compress your whole computer enough to fit in your wallet, sure . . .
795. # [19:28] <Lachy> OpenID can be secure if the user is careful to never reveal their openID password by using it on any other site, and if they trust their open id provider and if they are resistant to phishing attacks
796. # [19:28] <TabAtkins> Lachy: Yeah, phishing is the major weakness.
797. # [19:29] <Dashiva> AryehGregor: It's about as likely as users becoming resistant to phishing attacks ;)
798. # [19:29] <Philip`> Can't the OpenID provider just use something better than passwords?
799. # [19:29] <Philip`> (like PKA or whatever)
800. # [19:29] <TabAtkins> Philip` sure.
801. # [19:29] <Philip`> so the main benefit is you can choose a more secure login method without requiring a zillion sites to independently support it
802. # [19:29] <Lachy> I want to set up and run my own personal openid server, which would mean that I can control the login and make it as secure as possible, while being virtually invulnerable to phishing attacks
803. # [19:29] <TabAtkins> You can authenticate however you feel like. Which is one of the nice things about it - you don't need to rely on the website to be secure.
804. # [19:29] <AryehGregor> Dashiva, PKA is immune to phishing in the sense of being able to steal someone's login info.
805. # [19:30] <TabAtkins> Lachy: if you have PHP, there's a really simple two-file server that takes all of 5 minutes to get running.
806. # [19:30] <Dashiva> "We are upgrading your device. Please send it to us, together with the password"
807. # [19:30] <Lachy> TabAtkins, yeah, I found some on openidenabled.com that I will look into
808. # [19:30] * Joins: webben (n=benh@dip5-fw.corp.ukl.yahoo.com)
809. # [19:30] <Lachy> is that the one you're referring to?
810. # [19:31] <TabAtkins> Hmm, dunno. I use http://siege.org/projects/phpMyID/
811. # [19:32] <AryehGregor> Dashiva, "We are upgrading your bank account. Please send us your debit card, together with the PIN."
812. # [19:33] <AryehGregor> How often do you think that works?
813. # [19:33] <AryehGregor> There's practically no reason to have *any* authentication other than PKA.
814. # [19:33] <Dashiva> AryehGregor: A card is just a card. We're talking about a computer in your pocket
815. # [19:33] <AryehGregor> People's house keys should be replaced with the same PKA.
816. # [19:33] <AryehGregor> What difference does that make to phishers?
817. # [19:34] * Quits: gsnedders (n=gsnedder@host86-167-19-36.range86-167.btcentralplus.com)
818. # [19:34] <AryehGregor> "We are upgrading your computer. Please send your laptop to us."
819. # [19:34] <AryehGregor> Not much more plausible, is it.
820. # [19:34] <AryehGregor> The reason people don't protect their passwords is because they're so used to giving them out to anyone who asks.
821. # [19:35] <Dashiva> AryehGregor: Well, wordpress has a system where you send them your ftp auth and they login and upgrade your installation
822. # [19:35] * Quits: erlehmann (n=erlehman@tmo-104-56.customers.d1-online.com) (Read error: 54 (Connection reset by peer))
823. # [19:35] <Lachy> Dashiva, what the?
824. # [19:36] <Lachy> Why is that necessary. Recent releases of WordPress support automatic upgrading from within the Dashboard
825. # [19:36] <jgraham> Speaking of which, did we upgrade to the leatest wordpress?
826. # [19:36] <AryehGregor> Dashiva, yeah, but . . . you can't read your private keys off the device either. You *can't* give them to someone over the Internet.
827. # [19:36] <jgraham> *latest
828. # [19:36] <Lachy> jgraham, yes
829. # [19:36] <AryehGregor> A complete scheme would require some delegation infrastructure.
830. # [19:36] <AryehGregor> But the authenticator could always refuse to accept the delegation.
831. # [19:37] <jgraham> Lachy: Oh good. Now aall we need is a plan to migrate to a less insecure system
832. # [19:37] <Lachy> jgraham, but one thing I should upgrade is out mediawiki installation. I haven't done that in a while
833. # [19:37] <AryehGregor> Comparable attacks would always be possible if people are dedicated enough, obviously.
834. # [19:37] <Dashiva> AryehGregor: No, but it's a complicated device. People are liable to believe all kinds of things when they don't understand the subject.
835. # [19:37] <AryehGregor> Dashiva, the reason people give out their passwords all the time is because that's what you're *supposed* to do with passwords.
836. # [19:38] * Quits: Unixmonkey (n=Unixmonk@ppp-69-208-150-144.dsl.ipltin.ameritech.net)
837. # [19:38] <AryehGregor> The way password authentication works is that to authenticate, you give someone else all the info they'd need to impersonate you.
838. # [19:38] <AryehGregor> That's the problem here.
839. # [19:38] <Lachy> btw, did someone go through all the registerred users in the whatwg blog and delete all the spammer accounts? When I looked today, there were only about 30 accounts of relatively trusted people
840. # [19:38] <AryehGregor> People aren't inclined to give out sensitive info per se. They're only inclined to give out info that they reasonably expect someone else should have reason to want.
841. # [19:39] <Dashiva> ...
842. # [19:39] <AryehGregor> Phishers don't get people to give them their car keys.
843. # [19:39] <AryehGregor> In fact, phishing doesn't really work at all in IRL. Only digitally.
844. # [19:39] <AryehGregor> And you can't give out the device in question digitally. It's a physical device.
845. # [19:39] <Dashiva> Oh, so social engineering doesn't exist?
846. # [19:39] <AryehGregor> Lachy, whoa, you're at 1.11.
847. # [19:40] * Joins: cardona507 (n=cardona5@c-67-180-160-250.hsd1.ca.comcast.net)
848. # [19:40] <Lachy> AryehGregor, set up a fake valet service in a car park, in an area where valets are common, and see how many people hand over their keys!
849. # [19:40] <Dashiva> Night drop scams should be a great example of how people are trusting even with very important stuff
850. # [19:40] <Lachy> AryehGregor, yeah, I know. I must upgrade that
851. # [19:40] <AryehGregor> Lachy, possible, I'm sure. Widespread problem? No.
852. # [19:40] * TabAtkins belatedly realizes that he should have been testing all these functions as he wrote them. Damn this asynchronous write-save-minimize-transfer-browser-refresh cycle.
853. # [19:40] <AryehGregor> Doing anything IRL is a lot riskier, and harder to scale.
854. # [19:41] <AryehGregor> Set up a website and you can easily get thousands of hits for little effort, and be close to untraceable.
855. # [19:41] <AryehGregor> In real life? Very risky. What if one guy the whole day gets suspicious and tells someone?
856. # [19:41] <Lachy> AryehGregor, that's because stealing a car and getting away with it isn't as easy as stealing a login by phishing over the internet, and remaining completely anonymous
857. # [19:41] <AryehGregor> Precisely.
858. # [19:41] <AryehGregor> Stealing physical objects is much harder.
859. # [19:42] <AryehGregor> Stealing a physical object plus the password needed to unlock it, still harder.
860. # [19:42] <AryehGregor> Impossible? Hardly. But much harder.
861. # [19:42] <AryehGregor> PKA schemes normally have provision for revocation too, remember.
862. # [19:42] <AryehGregor> Unlike password schemes.
863. # [19:42] <TabAtkins> AryehGregor: So write up an openid server that uses PKA. ^_^
864. # [19:42] <TabAtkins> Best of all worlds.
865. # [19:42] <AryehGregor> TabAtkins, you don't get the advantage of my scheme without a real infrastructure.
866. # [19:43] <AryehGregor> With support by all OSes, browsers, and so forth.
867. # [19:43] * Quits: cardona507 (n=cardona5@c-67-180-160-250.hsd1.ca.comcast.net) (Client Quit)
868. # [19:43] <AryehGregor> It's hard to see how to do it incrementally while retaining the benefits.
869. # [19:43] <AryehGregor> I guess you could replace existing PKA systems, for SSH and so on, pretty easily.
870. # [19:43] <AryehGregor> But that's really niche.
871. # [19:43] <jgraham> TabAtkins: Yeah the development loop for browser stuff still really sucks doesn't it...
872. # [19:43] <TabAtkins> I don't see why not. You don't get as *comprehensive* of a scheme, but using PKA as a secure way of establishing identity without actually compromising your identity should be fine.
873. # [19:44] <AryehGregor> TabAtkins, it would hardly be an improvement over password managers if the private keys remained on your computer.
874. # [19:45] <TabAtkins> I'm much less concerned about your computer being compromised, precisely because it *does* require a very complete idea of security and will be defeated anyway by the first idiot that sticks an untrusted usb in his computer, and also because it's simply less common.
875. # [19:45] <TabAtkins> But you could certainly also use a usb-mounted program which communicates with the openid page.
876. # [19:46] <AryehGregor> Sure. But in that case, you may as well just use a password manager.
877. # [19:46] <AryehGregor> Over HTTPS, of course.
878. # [19:46] <AryehGregor> The advantages of PKA in that case would be marginal.
879. # [19:47] <TabAtkins> jgraham: it would be a lot better if I were developing on a linux box with apache on it. I don't want to even *try* putting php on windows again.
880. # [19:47] <AryehGregor> I had a perfectly good WAMP setup.
881. # [19:48] <AryehGregor> You just need to get one of the prepackaged things.
882. # [19:48] <AryehGregor> You know, something that installs the AMP part together.
883. # [19:48] <TabAtkins> Any idea where I can find a good one?
884. # [19:49] <AryehGregor> Just Google it. It's been a while since I've had to use Windows for much of anything, thankfully.
885. # [19:49] <AryehGregor> XAMPP is widely used, I think.
886. # [19:50] <TabAtkins> Yeah, one of my friends set up XAMPP. Maybe I'll try it again today.
887. # [19:51] <TabAtkins> When I build another desktop box for myself in a few months, I'll run Windows on that for gaming purposes and then wipe this laptop and load it with ubuntu again.
888. # [19:51] <TabAtkins> Well, I'll probably dualboot the desktop.
889. # [19:51] * Joins: erlehmann (n=erlehman@tmo-104-56.customers.d1-online.com)
890. # [19:51] * Joins: dglazkov (n=dglazkov@c-67-188-0-62.hsd1.ca.comcast.net)
891. # [19:52] <AryehGregor> (Actually, I used Windows for gaming as of fairly recently. Now I've finished Bioshock and started playing Baldur's Gate II, so Linux works fine.)
892. # [19:52] <AryehGregor> (Without having to load any wretched binary video drivers.)
893. # [19:52] * TabAtkins wonders if it would be possible to make his monitors wireless.
894. # [19:53] <TabAtkins> Ah, excellent. Looks like tools to do so have started popping up in the relative mainstream since last year.
895. # [19:54] * Joins: Unixmonkey (n=Unixmonk@ppp-69-208-150-144.dsl.ipltin.ameritech.net)
896. # [20:00] <AryehGregor> You'd still need power, surely.
897. # [20:03] <TabAtkins> Yes, of course. That's not a problem.
898. # [20:04] <TabAtkins> But my vision for the desktop box is to go with an open design, and actually build it onto a shelf.
899. # [20:04] <TabAtkins> And I'd prefer not to have wires running from my monitors to the shelf.
900. # [20:05] <Lachy> AryehGregor, this mediawiki upgrade procedure isn't very clear
901. # [20:05] <Lachy> I'm reading http://www.mediawiki.org/wiki/Manual:Upgrading
902. # [20:06] <Lachy> I ran the tar command with the --strip-path=1 parameter, and it seemed to do the extraction. Though I'm not sure which files it actually replaced and which files it kept unchanged.
903. # [20:06] <Lachy> though AdminSettings.php was not overwritten
904. # [20:06] <AryehGregor> Lachy, AdminSettings.php and LocalSettings.php shouldn't be overwritten.
905. # [20:06] <AryehGregor> You just need to unpack it and run maintenance/update.php.
906. # [20:07] <AryehGregor> It would be a good idea to upgrade extensions at the same time.
907. # [20:07] <Lachy> oh, I didn't see where that page said update.php was in the maintenance directory
908. # [20:08] <AryehGregor> Most of the docs on mediawiki.org are half-nonsense. Stupid wikis.
909. # [20:08] * Quits: dglazkov (n=dglazkov@c-67-188-0-62.hsd1.ca.comcast.net)
910. # [20:08] <Lachy> oh damn. I need Hixie to switch wiki.whatwg.org to use PHP5
911. # [20:08] <AryehGregor> The file UPGRADE in the tarball is probably more useful and accurate.
912. # [20:08] <AryehGregor> Er, 1.11 doesn't run on PHP 5 either.
913. # [20:08] <Lachy> hmm, the php command used php4. Maybe there's a separate command to run php5
914. # [20:08] <AryehGregor> If php maintenance/upgrade.php fails, try php5 maintenance/upgrade.php.
915. # [20:09] <Lachy> there is no php5 command
916. # [20:09] <AryehGregor> Well, it's got to be installed somewhere, since the wiki currently works.
917. # [20:10] <AryehGregor> What distro is this? I assume it's Linux.
918. # [20:10] <AryehGregor> And what version of the distro.
919. # [20:11] <Lachy> I don't know. It's Dreamhost.
920. # [20:11] <AryehGregor> Try creating a file named phpinfo.php with the contents "<?php phpinfo();" and viewing it to get some info on where the executable is coming from for the website.
921. # [20:11] <AryehGregor> Oh.
922. # [20:11] <AryehGregor> Hmm.
923. # [20:12] <AryehGregor> Apparently you need to do: /usr/local/php5/bin/php maintenance/upgrade.php
924. # [20:13] <Lachy> AryehGregor, http://wiki.whatwg.org/phpinfo.php
925. # [20:13] <AryehGregor> Well, that doesn't give me the info I wanted, but I found it on Google anyway, so try what I just said.
926. # [20:14] <Lachy> upgrading with that command
927. # [20:15] <Lachy> http://wiki.whatwg.org/wiki/Special:Version now says 1.15.1
928. # [20:15] <AryehGregor> You're probably fine, then.
929. # [20:16] <AryehGregor> Doubtless there are some useless files hanging around, but that doesn't really matter.
930. # [20:17] <Lachy> what useless files?
931. # [20:17] <Lachy> like files from the old version that are no longer needed in the new version?
932. # [20:17] <AryehGregor> Things that were moved around.
933. # [20:17] <AryehGregor> Or removed.
934. # [20:17] <AryehGregor> Yeah.
935. # [20:18] <Lachy> ok
936. # [20:18] <AryehGregor> Like we made a bunch of subdirectories in includes/ and moved tons of stuff there.
937. # [20:18] <AryehGregor> It doesn't really matter, they don't hurt anything.
938. # [20:18] <Lachy> if I wanted to to a fresh installation next time, does it just require extracting to a new directory, copying AdminSettings.php, LocalSettings.php and .htaccess across and running the update script the same way?
939. # [20:18] * Joins: dglazkov (n=dglazkov@c-67-188-0-62.hsd1.ca.comcast.net)
940. # [20:19] <AryehGregor> A fresh installation with the same database?
941. # [20:19] <Lachy> yeah
942. # [20:19] <AryehGregor> That should work in theory, yes, but LocalSettings.php does contain some paths that might need to be changed.
943. # [20:19] <TabAtkins> Aw man, rock on. The big problem was just that I was passing invalid json to json_decode. Now my parser works great.
944. # [20:19] <AryehGregor> MediaWiki developers tend to just use SVN for all our wikis, so it does the deletion for us. :P
945. # [20:20] <TabAtkins> <3 svn
946. # [20:20] <Lachy> AryehGregor, how about I just give you shell access to wiki.whatwg.org and you can handle the wiki upgrades for us?
947. # [20:20] <AryehGregor> MediaWiki is weird because we don't mostly develop it for the sake of releasing it. It's mainly for Wikipedia.
948. # [20:20] <AryehGregor> So niceties like a usable, well-documented upgrade procedure tend to go by the wayside.
949. # [20:21] <AryehGregor> Wikipedia doesn't even use the update script, it would kill the site for hours. Schema updates get applied manually.
950. # [20:21] <AryehGregor> Lachy, could do. I could upgrade it to a 1.16alpha version that uses HTML 5, in fact. :) But the problem is, I don't know right now if I'll have the time as of a few days from now. So I don't want to commit to anything.
951. # [20:21] <Lachy> does the update script make changes to the database structure?
952. # [20:22] <AryehGregor> Feel free to ask again in a week or two.
953. # [20:22] <AryehGregor> Yes, that's basically all it does.
954. # [20:22] <AryehGregor> Well, it also populates things with defaults sometimes, and various other stuff.
955. # [20:22] <AryehGregor> But most of it is schema changes.
956. # [20:23] * Quits: kristallpirat (n=kristall@c-base/crew/kristall) (Remote closed the connection)
957. # [20:23] <Lachy> AryehGregor, no need to commit to anything immediately. But if you send me an SSH public key, I can add it and you can login using: ssh whatwikiuser@whatwg.org
958. # [20:24] <AryehGregor> ssh-rsa AAAAB3NzaC1yc2EAAAABJQAAAIEAgP8oDfL75zl8aiMA7nmSjhHIqM2MZQ5MmGbRxfRiaaWw9nYjz+sQJ6qntJ+BimNm+lq+ggOK5pxQQ+KD4HwvDJTkGuMIvk4+VjwGHkZdgLiNfl/dTQQD9XNEfq6ilwy/tSn/ka6uNKZaMOjk8z+AVNE/Z71naTUe7E4JRAaz6WM=
959. # [20:28] <Lachy> AryehGregor, should work now
960. # [20:29] <AryehGregor> Works.
961. # [20:29] <Philip`> TabAtkins: Low-res monitor might be 1280x800, 24-bit, 60Hz, which is two gigabits per second
962. # [20:29] <Philip`> I'd quite like a two gigabit wireless network, but it doesn't seem like near-future technology :-(
963. # [20:30] <TabAtkins> That's assuming no compression, Philip`.
964. # [20:31] <Philip`> TabAtkins: I expect it'd be hard to compress without introducing unacceptable latency
965. # [20:31] * Joins: SavageX_ (n=maikmert@Z82d2.z.pppool.de)
966. # [20:31] <TabAtkins> Apparently it's okay. They produced tech demos last year that seemed to be okay with very low latency.
967. # [20:31] <AryehGregor> For a 60 Hz screen?
968. # [20:32] <TabAtkins> It appeared to be a pretty standard LCD, so I would assume so.
969. # [20:32] <AryehGregor> You can compress a screenful in less than 17 ms, surely.
970. # [20:32] <TabAtkins> The engadget report quoted them as saying that the only weakness is trying to use them while doing high-end gaming, as the game chews up too much cpu to let the compression run smoothly.
971. # [20:32] <AryehGregor> But you can't rely on lossless compression. If the output is too hard to compress, you're going to introduce stuttering.
972. # [20:32] <AryehGregor> Unless you want lossy compression for your monitor.
973. # [20:33] <AryehGregor> Games won't have good lossless compression at all.
974. # [20:33] <AryehGregor> Unless they're, like, Tetris.
975. # [20:33] <Philip`> Lossless compression of video doesn't really exist, unless you're happy with something like a 2:1 compression ratio
976. # [20:33] <TabAtkins> Depending on the level of lossiness and the application, I'd be okay with that.
977. # [20:33] <TabAtkins> Best is if it could identify when I start running into compression issues and switch between lossless and lossy.
978. # [20:34] * Quits: erlehmann (n=erlehman@tmo-104-56.customers.d1-online.com) (Read error: 110 (Connection timed out))
979. # [20:34] <Philip`> If you have the bandwidth for lossless, why wouldn't you always use it?
980. # [20:34] * Quits: dglazkov (n=dglazkov@c-67-188-0-62.hsd1.ca.comcast.net)
981. # [20:34] <Philip`> (and if you don't have the bandwidth, you can't use the option anyway)
982. # [20:34] <TabAtkins> If I had the bandwidth, of course I would.
983. # [20:35] <AryehGregor> Lossy compression tends to seriously mess up normal text. At least JPEG does.
984. # [20:35] <AryehGregor> It becomes blurry.
985. # [20:36] <TabAtkins> Yeah, but if I'm viewing text I'm almost certainly doing something that will compress well losslessly.
986. # [20:37] <TabAtkins> JPEG isn't a good example of lossy compression on text, though. It's specifically designed to work well on natural photographs.
987. # [20:37] <TabAtkins> PNG compresses text relatively well.
988. # [20:37] <Philip`> PNG isn't a good example of lossy compression on text, or on anything else
989. # [20:38] <TabAtkins> Oh, but hum. Is PNG lossless?
990. # [20:38] <Philip`> Yes
991. # [20:38] <TabAtkins> Never mind, then. ^_^ But I am vindicated anyway! Text can be losslessly compressed pretty well.
992. # [20:38] <AryehGregor> Very well, yes.
993. # [20:39] <AryehGregor> Especially if you don't antialias or such, I guess.
994. # [20:39] <TabAtkins> Yeah, that would make it even better, but still. a screen of text is still *mostly* just the background color.
995. # [20:40] <TabAtkins> Also: Phase 1 of my radial-gradient() parser is complete. Now I begin the transformer that will take the parsed values and figure out the parameters I need for generating an image.
996. # [20:40] <Philip`> What do you do when the user views a PDF brochure with text on some swishy fading photo background?
997. # [20:41] <Philip`> (It's no good having a monitor that only works for easy cases, and turns into unreadable mush once a month when you do something unusual)
998. # [20:42] <TabAtkins> I dunno. I'm not a wireless-monitor manufacturer. But at least one company is apparently confident they can do it.
999. # [20:42] <Philip`> You mean, one company's PR department is confident they can do it?
1000. # [20:43] <TabAtkins> Though it looks like I'll have to use normal monitors for now anyway, as I don't think there will be anything in my price range by November.
1001. # [20:43] <Philip`> Why do you want a wireless monitor anyway?
1002. # [20:43] <Philip`> The easier solution is to wire a computer to it, then use a wireless mouse and keyboard, and you get exactly the same effect
1003. # [20:44] <TabAtkins> Because I want a computer mounted on a shelf.
1004. # [20:44] <Philip`> You could mount a computer on a shelf, and then have another one hidden away somewhere wired to a monitor and use that one instead
1005. # [20:44] <TabAtkins> Though I am open to possible other suggestions that still express a steampunk motif.
1006. # [20:45] <TabAtkins> That's just silly, Philip`.
1007. # [20:45] <Philip`> You'll get the same external effect
1008. # [20:45] <TabAtkins> Doesn't matter. It's not cool to say "Yeah, that thing on the wall behind me is just for show. I *really* use this perfectly ordinary box under here."
1009. # [20:45] <Philip`> So don't say that
1010. # [20:46] <Philip`> Just lie
1011. # [20:46] <TabAtkins> I prefer not to lie? Especially for no reason?
1012. # [20:46] <Philip`> Or you could run a monitor cable through the walls to the monitor
1013. # [20:46] <TabAtkins> Instead I'll just run monitor cables to the shelf for now. Wireless later.
1014. # [20:47] * Quits: maikmerten_ (n=maikmert@BAE1a14.bae.pppool.de) (Read error: 110 (Connection timed out))
1015. # [20:47] <Philip`> Lying for no reason is much more honest than lying for a dishonest reason
1016. # [20:47] <TabAtkins> But much *less* honest than not lying at all.
1017. # [20:47] <TabAtkins> Also: it sorta makes you a sociopath.
1018. # [20:47] <Philip`> Is that a bad thing?
1019. # [20:48] <Philip`> I saw a TV series about a sociopath and he seemed a nice enough guy
1020. # [20:48] <AryehGregor> Hmm, I guess PNG would be very suboptimal here. You'd want to do delta compression of some kind.
1021. # [20:48] <TabAtkins> Except for the sociopathy, sure.
1022. # [20:48] <AryehGregor> That would handle swirly PDFs just fine, losslessly.
1023. # [20:48] <Philip`> Well, yes, except for the murders and so on
1024. # [20:49] <TabAtkins> To be fair, very few sociopaths are murderers.
1025. # [20:49] <TabAtkins> Most of them are executives.
1026. # [20:49] <TabAtkins> But I'm not really interested in being an executive either.
1027. # [20:49] <AryehGregor> Sociopaths aren't nice. By definition, they don't really care about other people.
1028. # [20:49] <AryehGregor> But you *are* interested in being a murderer?
1029. # [20:49] <TabAtkins> AryehGregor: They can learn to fake it, and be relatively normal.
1030. # [20:50] <Philip`> AryehGregor: It's fine until you alt-tab in your fancy windowing system and it fades out the screen in a way that the delta compressor can't recognise and suddenly everything slows to a crawl
1031. # [20:50] <AryehGregor> Philip`, for one or two frames, yeah.
1032. # [20:50] <TabAtkins> No, the "either" at the end implies that neither are acceptable.
1033. # [20:50] <Philip`> AryehGregor: It might be a really slow fade
1034. # [20:50] <AryehGregor> I dunno, are there 2 Gbps wireless cards over sufficiently short range?
1035. # [20:50] <AryehGregor> TabAtkins, oh, okay. Just checking.
1036. # [20:50] <Philip`> AryehGregor: and the monitor compression would destroy it and the user would be sad and might murder somebody
1037. # [20:51] <AryehGregor> Or become an executive.
1038. # [20:51] <TabAtkins> I'll take that risk, Philip`.
1039. # [20:51] <TabAtkins> The real blame here is the windowing system, anyway.
1040. # [20:51] <Philip`> AryehGregor: BlueTooth can do 3 Mbps, which probably isn't quite enough
1041. # [20:51] <Philip`> *Bluetooth
1042. # [20:51] <TabAtkins> Command-line should be enough for anybody.
1043. # [20:51] <AryehGregor> But that's not so short-range, is it?
1044. # [20:52] <Philip`> AryehGregor: 802.11n can apparently do~150Mbps
1045. # [20:52] <TabAtkins> 2Gbps is only needed if you do no compression at all anyway.
1046. # [20:52] <AryehGregor> TabAtkins, yes, but the monitor still has to work with pixels. So when you go to the next level in NetHack, you might notice a delay as it has to redraw the whole screen.
1047. # [20:52] <AryehGregor> TabAtkins, there will always be some cases where you can get zero compression, for any compression algorithm.
1048. # [20:52] <Philip`> AryehGregor: What do you consider "short"? Bluetooth seems to be measured in meters, so you can't get much shorter if you want to cross a room
1049. # [20:53] <TabAtkins> AryehGregor, well of course. That's a pretty natural consequence.
1050. # [20:53] <TabAtkins> (In fact, *most* cases produce zero compression for any given algorithm.)
1051. # [20:53] <jgraham> FWIW I have seen people demo monitors-over-ethernet
1052. # [20:54] <TabAtkins> The company I was looking at does monitors over usb2
1053. # [20:54] <jgraham> I don't recall what they shows but they claimed that it worked well enough for computer games
1054. # [20:54] <jgraham> TabAtkins: url?
1055. # [20:54] <Philip`> AryehGregor: Do you mean "for any *lossless* compression algorithm"? (which isn't what you ever want for video)
1056. # [20:54] <AryehGregor> Yes, obviously.
1057. # [20:54] <TabAtkins> http://www.engadget.com/2007/01/15/displaylink-cuts-your-monitors-cord-utilizes-wireless-usb/
1058. # [20:55] <AryehGregor> And you do want lossless compression for some video.
1059. # [20:55] <TabAtkins> DisplayLink is the company. They don't have a wireless product for sale yet, though.
1060. # [20:55] <AryehGregor> If it's a video of someone clicking through text-dominated web pages.
1061. # [20:55] * Joins: gsnedders (n=gsnedder@host86-167-19-36.range86-167.btcentralplus.com)
1062. # [20:56] <TabAtkins> A different article about wireless monitors was talking about them deploying using some new wireless technologies like Ultra Wide Band.
1063. # [20:56] <Philip`> AryehGregor: You could do it lossily with 200Mbps H.264 and I expect nobody would notice the difference
1064. # [20:56] <jgraham> TabAtkins: those were likely the people I saw
1065. # [20:56] <AryehGregor> Probably, yeah.
1066. # [20:57] <Philip`> I don't think human eyeballs and brains have a bandwidth of 2Gbps, so it'll always be possible to lossily compress without anybody noticing
1067. # [20:58] <Philip`> Anyway, it's trivial to lossily compress any content by just deleting one pixel and the user will assume it's a dead pixel in their monitor and won't mind much
1068. # [20:59] <AryehGregor> That's not very useful, though.
1069. # [21:00] <Philip`> You expect me to suggest something useful?
1070. # [21:01] <TabAtkins> Lossy compression algorithm: just count the number of 1s in the file, and send that as a binary number. This can be iterated if necessary.
1071. # [21:04] <AryehGregor> Better lossy compression algorithm: send no data, let client make it up.
1072. # [21:04] <AryehGregor> It could prove to be very entertaining.
1073. # [21:04] <AryehGregor> You could just store an old session and replay it whenever the computer is turned on, automatically stitching it together somehow into a seamless loop.
1074. # [21:04] <TabAtkins> I like it.
1075. # [21:05] <AryehGregor> Unsophisticated users would think their computer was taken over by a ghost.
1076. # [21:05] <AryehGregor> Bonus points if the stored session is doing something embarrassing or incriminating.
1077. # [21:05] <TabAtkins> "BUG: Installation of software induces supernatural possession of user's computer. SUGGESTED FIX: Ghostbusters?"
1078. # [21:16] * Quits: Unixmonkey (n=Unixmonk@ppp-69-208-150-144.dsl.ipltin.ameritech.net)
1079. # [21:18] * Joins: Unixmonkey (n=Unixmonk@ppp-69-208-150-144.dsl.ipltin.ameritech.net)
1080. # [21:19] * Quits: Kalms (n=rasmuska@81.161.185.108)
1081. # [21:23] * Joins: tantek (n=tantek@adsl-63-195-114-133.dsl.snfc21.pacbell.net)
1082. # [21:35] * Quits: Lachy (n=Lachlan@85.196.122.246) ("Leaving")
1083. # [21:36] * Joins: vvv (n=vvv@mediawiki/VasilievVV)
1084. # [21:41] * Joins: nessy (n=nessy@203-214-73-15.dyn.iinet.net.au)
1085. # [21:46] * Joins: ttepasse (n=ttepas--@p5B014290.dip.t-dialin.net)
1086. # [21:52] * Quits: gsnedders (n=gsnedder@host86-167-19-36.range86-167.btcentralplus.com) (Read error: 104 (Connection reset by peer))
1087. # [21:52] * Joins: gsnedders (n=gsnedder@host86-167-19-36.range86-167.btcentralplus.com)
1088. # [22:02] * Joins: dglazkov (n=dglazkov@c-67-188-0-62.hsd1.ca.comcast.net)
1089. # [22:05] * Quits: dglazkov (n=dglazkov@c-67-188-0-62.hsd1.ca.comcast.net) (Client Quit)
1090. # [22:10] * Joins: zcorpan_ (n=zcorpan@c83-252-193-59.bredband.comhem.se)
1091. # [22:11] * Joins: Lachy (n=Lachlan@85.196.122.246)
1092. # [22:14] * Joins: roc (n=roc@121-72-174-84.dsl.telstraclear.net)
1093. # [22:20] * Joins: cardona507 (n=cardona5@c-67-180-160-250.hsd1.ca.comcast.net)
1094. # [22:20] * Joins: Super-Dot (n=Super-Do@adsl-76-231-44-168.dsl.pltn13.sbcglobal.net)
1095. # [22:21] * Quits: zdobersek (n=zan@cpe-92-37-68-231.dynamic.amis.net) ("Leaving.")
1096. # [22:26] * Quits: zcorpan_ (n=zcorpan@c83-252-193-59.bredband.comhem.se) (Read error: 110 (Connection timed out))
1097. # [22:31] * Joins: Kalms (n=rasmuska@81.161.185.108)
1098. # [22:31] * Quits: roc (n=roc@121-72-174-84.dsl.telstraclear.net)
1099. # [22:33] * Joins: Rik`_ (n=Rik`@pha75-2-81-57-187-57.fbx.proxad.net)
1100. # [22:33] * Quits: tantek (n=tantek@adsl-63-195-114-133.dsl.snfc21.pacbell.net) (Read error: 104 (Connection reset by peer))
1101. # [22:33] * Joins: tantek (n=tantek@adsl-63-195-114-133.dsl.snfc21.pacbell.net)
1102. # [22:34] * Quits: mlpug (n=mlpug@a88-115-164-40.elisa-laajakaista.fi) (Remote closed the connection)
1103. # [22:39] * Quits: gsnedders (n=gsnedder@host86-167-19-36.range86-167.btcentralplus.com) (Read error: 110 (Connection timed out))
1104. # [22:43] * Quits: ROBOd (n=robod@89.122.216.38) ("http://www.robodesign.ro")
1105. # [22:44] * Quits: Rik` (n=Rik`@pha75-2-81-57-187-57.fbx.proxad.net) (Read error: 110 (Connection timed out))
1106. # [22:44] * Rik`_ is now known as Rik`
1107. # [23:18] * Joins: roc (n=roc@203-97-204-82.dsl.clear.net.nz)
1108. # [23:18] * Quits: svl_ (n=me@g227035173.adsl.alicedsl.de) ("And back he spurred like a madman, shrieking a curse to the sky.")
1109. # [23:18] * Quits: da3d (n=opera@h11n1fls34o986.telia.com)
1110. # [23:20] * Quits: shepazu (n=schepers@31-33-83.wireless.csail.mit.edu)
1111. # [23:27] * Quits: tantek (n=tantek@adsl-63-195-114-133.dsl.snfc21.pacbell.net)
1112. # [23:29] * Joins: murr4y` (n=murray@85.84-49-67.nextgentel.com)
1113. # [23:30] * Quits: murr4y (n=murray@85.84-49-67.nextgentel.com) (Read error: 104 (Connection reset by peer))
1114. # [23:31] * Quits: Maurice (i=copyman@5ED548D4.cable.ziggo.nl)
1115. # [23:48] * Joins: murr4y (n=murray@85.84-49-67.nextgentel.com)
1116. # [23:53] * Quits: Rik` (n=Rik`@pha75-2-81-57-187-57.fbx.proxad.net) (Read error: 113 (No route to host))
1117. # [23:53] * Joins: Rik` (n=Rik`@pha75-2-81-57-187-57.fbx.proxad.net)
1118. # [23:59] * Quits: murr4y` (n=murray@85.84-49-67.nextgentel.com) (Read error: 110 (Connection timed out))
1119. # Session Close: Mon Sep 07 00:00:00 2009

The end :)