/irc-logs / w3c / #webapps / 2009-05-16 / end

Options:

  1. # Session Start: Sat May 16 00:00:01 2009
  2. # Session Ident: #webapps
  3. # [00:20] * Quits: arve (arve@84.202.133.45) (Quit: Ex-Chat)
  4. # [00:20] * Joins: arve (arve@84.202.133.45)
  5. # [01:02] * Quits: aroben (aroben@71.58.77.15) (Connection reset by peer)
  6. # [01:22] * Joins: Hixie (ianh@129.241.93.37)
  7. # [01:42] * Quits: Hixie (ianh@129.241.93.37) (Quit: bbiab)
  8. # [02:28] * Joins: Hixie (ianh@129.241.93.37)
  9. # [05:17] * Quits: heycam (cam@124.168.80.239) (Quit: bye)
  10. # [05:34] * Quits: gsnedders (gsnedders@86.136.52.180) (Client exited)
  11. # [09:24] * Joins: heycam (cam@124.168.80.239)
  12. # [12:38] * Joins: ArtB (c0647cdb@128.30.52.43)
  13. # [13:13] * Joins: gsnedders (gsnedders@86.136.52.180)
  14. # [13:48] * Joins: Marcos (Marcos@84.215.160.79)
  15. # [14:14] * Quits: Marcos (Marcos@84.215.160.79) (Quit: Marcos)
  16. # [15:17] * Joins: tlr (tlr@128.30.52.30)
  17. # [15:18] <tlr> art?
  18. # [15:18] <tlr> artb?
  19. # [15:18] <tlr> what's up?
  20. # [15:54] * Joins: Lachy (Lachlan@124.171.255.145)
  21. # [15:55] * Quits: tlr (tlr@128.30.52.30) (Quit: tlr)
  22. # [15:55] * Quits: arve (arve@84.202.133.45) (Ping timeout)
  23. # [16:07] * Joins: arve (arve@84.202.133.45)
  24. # [18:49] * Joins: Marcos (Marcos@84.215.160.79)
  25. # [20:31] <arve> Marcos: you're wrong on the Internet :D
  26. # [20:32] <Marcos> Ha! on the internet no one knows I'm a dog!
  27. # [20:32] <arve> no, but we know you're wrong
  28. # [20:32] <Marcos> hehe
  29. # [20:32] <arve> no, but srsly, I'll fight it tooth and claw if you try to treat URI's in feature as anything more than an opaque identifier
  30. # [20:33] <Marcos> ah fuck, I just rewrote that whole section
  31. # [20:34] <Marcos> :(
  32. # [20:35] <arve> I don't even think RDF pretends that http://foo/bar#baz and http://foo/bar#xyzzy are one and the same concept
  33. # [20:35] <Marcos> No, that is different because # refers to a fragment
  34. # [20:36] <Marcos> But it still refers to the same resource, but something within the resource
  35. # [20:37] <Marcos> so, for example http://marcos#arm and http://marcos#legs are different fragments of me
  36. # [20:37] <Marcos> but they still refer to me (resource)
  37. # [20:37] <Marcos> and http://marcos?mood=agry is still me, but angry
  38. # [20:38] <arve> Marcos: and I'm not using the same API to manipulate your mind as I'm using to manipulate your arm
  39. # [20:38] <arve> no
  40. # [20:38] <arve> ?mood=agry is a random, unidentifiable string that alters the resource marcos
  41. # [20:38] <arve> or even the representation of such
  42. # [20:38] <Marcos> and http://marcos#arm?mood=angry is angry Marcos' arm
  43. # [20:39] <Marcos> right
  44. # [20:41] <arve> my entire point here is that you'll run into problems as soon as you encounter a uri scheme for which the query string doesn't exist
  45. # [20:41] <arve> data: tag: uuid: opera: chrome:
  46. # [20:41] <arve> file:
  47. # [20:41] <arve> the only thing you win by applying magic to the query string is to complicate processors
  48. # [20:42] <arve> I'd perhaps even go as far as to say that it's an opaque string, but that an author SHOULD use a URI that resolves, pointing to a description of the API
  49. # [20:43] <Marcos> I don't think any of the mentioned URI schemes care about queries, they are just ignored
  50. # [20:44] <Marcos> I think you have misunderstood my proposal a little bit
  51. # [20:45] <Marcos> param just takes strings as name value pairs
  52. # [20:45] <Marcos> those strings are then converted to URI queries at runtime and tagged onto the end of feature name
  53. # [20:45] <Marcos> yes, feature name is URI however
  54. # [20:47] <Marcos> For example, for the following feature element, the resulting feature name would be "http://example.org/api.geolocation?Gr%C3%BCnerl%C3%B8kka,0554,%20Oslo".
  55. # [20:47] <Marcos> <widget xmlns="http://www.w3.org/ns/widgets">
  56. # [20:47] <Marcos> <feature name="http://example.org/api.geolocation">
  57. # [20:47] <Marcos> <param name="home" value="Grünerløkka, 0554, Oslo"/>
  58. # [20:47] <Marcos> </feature>
  59. # [20:47] <Marcos> </widget>
  60. # [20:47] <Marcos> woops, booboo, make that "http://example.org/api.geolocation?home=Gr%C3%BCnerl%C3%B8kka,0554,%20Oslo"
  61. # [20:56] * Quits: ArtB (c0647cdb@128.30.52.43) (Quit: CGI:IRC (Ping timeout))
  62. # [21:33] <arve> Marcos: I got what you meant, but the feature in this case would, either way, be "geolocation"
  63. # [21:33] <arve> that I input Grünerløkka, Oslo into the API as some default doesn't change the fact that the api is still "geolocation"
  64. # [21:33] <Marcos> right
  65. # [21:34] <arve> nor do I think that pretending that configuration data is resolvable is even remotely useful
  66. # [21:34] <arve> which is my main beef with the proposal
  67. # [21:34] <arve> you're making processing harder
  68. # [21:34] <Marcos> true
  69. # [21:34] <arve> for no apparent reason
  70. # [21:34] <Marcos> well, the reason was that we chose to go with URIs
  71. # [21:34] <Marcos> we can't say they are URIs and then say they are not
  72. # [21:35] <Marcos> either they are URIs or they are not
  73. # [21:35] <Marcos> If they are not URIs, then we need to change the spec
  74. # [21:35] <arve> I'll have to think about it
  75. # [21:35] <Marcos> We get a lot of free stuff with URIs... but yes, we get complexity too
  76. # [21:37] <arve> What do other specs that use the I in URI's think?
  77. # [21:40] * Quits: Marcos (Marcos@84.215.160.79) (Quit: Marcos)
  78. # [21:43] <Marcos> I guess we could see what HTML5 does with origin.
  79. # [22:03] * Quits: shepazu (schepers@128.30.52.30) (Quit: shepazu)
  80. # [23:36] * Quits: Lachy (Lachlan@124.171.255.145) (Quit: This computer has gone to sleep)
  81. # Session Close: Sun May 17 00:00:01 2009

The end :)