/irc-logs / freenode / #whatwg / 2007-11-26 / end

Options:

  1. # 00:00 < othermaciej> Does Mozilla's 3d canvas stuff use OpenGL?
  2. # 00:01 < Philip`> othermaciej: Yes
  3. # 00:01 < Philip`> (The API is OpenGL ES, so it's a subset of OpenGL, and it's currently implemented using OpenGL)
  4. # 00:01 < othermaciej> does that mean you can't alpha-composite things over it? Or does it just turn off hardware acceleration in that case?
  5. # 00:02 < othermaciej> actually, I guess it must be rendering to a backing store anyway just to draw over normal canvas stuff
  6. # 00:02 < othermaciej> so likely not fully accelerated all the way to the screen
  7. # 00:03 < alp> othermaciej: it used to use glitz to render the graphics context to gl as well, iirc. don't know if it still does
  8. # 00:03 < Lachy> when did Mozilla add 3d canvas?
  9. # 00:04 < Philip`> othermaciej: I can put a translucent <div> over the 3d canvas and it all looks properly blended
  10. # 00:04 < Lachy> is it compatible with Opera's 3D canvas?
  11. # 00:04 < Philip`> Lachy: It's currently just an unreleased extension
  12. # 00:04 < Lachy> ok
  13. # 00:04 < othermaciej> I don't think it's compatible
  14. # 00:04 < Lachy> so looks like we really need to get this standardised soon
  15. # 00:05 < othermaciej> and I think its existence as an extension predates Opera's API
  16. # 00:05 < Philip`> It's totally different to Opera's
  17. # 00:05 < Lachy> Philip`, which one do you think is better from an authoring perspective?
  18. # 00:05 < othermaciej> my understanding is that Mozilla's is very low-level, pretty close to the actual OpenGL API
  19. # 00:06 < Philip`> Someone said Opera "implemented 3D Canvas 2-3 years ago" though I don't know how much they've changed since then
  20. # 00:06 < aroben> http://my.opera.com/timjoh/blog/2007/11/13/taking-the-canvas-to-another-dimension has some of the motivation behind Opera's API
  21. # 00:07 < Philip`> Lachy: Opera's one is easier to use if you're programming straight to the API and not using any helper libraries, except it barely has any features (e.g. you can't do dynamic lighting) and it would need some redesigning in order to support those features
  22. # 00:07 < othermaciej> it seems like if you want things like collision detection you might also want an API model with a fully retained scene graph
  23. # 00:08 < Philip`> I'd prefer to implement high-level scene graph things in JS, and have the browser provide just a low-level rendering API
  24. # 00:08 < Philip`> (hence me trying to implement X3D in JS to see if that's feasible)
  25. # 00:09 < othermaciej> is X3D a sensible spec?
  26. # 00:09 < othermaciej> I don't really know how it works
  27. # 00:10 < othermaciej> does it natively provide for things like collision detection, or embedding foreign markup as a texture and still being able to hit test it?
  28. # 00:11 < othermaciej> I gotta reboot
  29. # 00:11 < Philip`> I've only looked at small bits of X3D, and it seems far from perfect but not totally useless
  30. # 00:11 < othermaciej> back soon
  31. # 00:11 < Philip`> It has Collision sensors which can send events to make stuff happen
  32. # 00:12 < Philip`> X3D is full of undefined behaviour so I don't think browsers could reasonably implement it
  33. # 11:16 < hsivonen> http://www.flickr.com/photos/psd/1805709102
  34. # 12:00 * gsnedders stays by the shores of the sea of interoperability
  35. # 12:00 < gsnedders> (having followed the link to TBL's similar thing)
  36. # 12:29 < gsnedders> annevk2: in #whatwg!? what's happened?
  37. # 12:40 < hsivonen> Hixie: what Keynote feature did you use to create a timeline?
  38. # 13:05 * hsivonen is totally puzzled with Keynote not having an obvious way to turn a CVS table of labels and numbers into a bar chart
  39. # 14:22 < hsivonen> where might I find a reference to the W3C decision to leave 4.01 as the last version of HTML and move to XHTML
  40. # 14:28 < Dashiva> Not a good reference, but /tr/html leads to XHTML 1.0
  41. # 14:30 < hsivonen> Dashiva: I meant the meeting and its date when the XML course was decided
  42. # 14:45 < zcorpan> hsivonen: last time i searched for it i didn't find anything
  43. # 14:47 < hsivonen> zcorpan: ok. good to know
  44. # 14:48 < zcorpan> which was... hmm, perhaps 2 years ago
  45. # 14:52 < Lachy> hsivonen, it would have been decided around 1998-1999 when work on XHTML began
  46. # 15:59 < zcorpan> <title> <textarea> <style> <script> <xmp> <noscript> <iframe> <noembed> <noframes>
  47. # 15:59 < zcorpan> are there any others that are [R]CDATA elements?
  48. # 15:59 < zcorpan> <xml> in ie, i think, but other than that?
  49. # 16:04 < zcorpan> i guess <applet> and <object> are pseudo-CDATA in ie too
  50. # 17:50 < zcorpan> hsivonen: feature request: it would be useful to be able to group errors by message (perhaps so by default even), because often there are lots of errors of the same type and you usually fix them with a search-and-replace
  51. # 17:50 < zcorpan> e.g. http://validator.nu/?doc=http%3A%2F%2Fwww.cafefernando.com%2Fturkce%2F
  52. # 17:51 < zcorpan> s/group/collapse/
  53. # 17:56 < zcorpan> (in case that page changes: it contains lots of "<a .../><img/></a>" which results in lots of "Stray / in tag." and "An a start tag seen with already an active a element." messages)
  54. # 18:16 < hsivonen> zcorpan: feature request recorded. However, I probably won't be implementing the requested feature any time soon, because Validator.nu doesn't have the concept of "same type" and the whole architecture is designed to assume that such grouping is unwanted
  55. # 18:17 < hsivonen> so introducing the feature inside Validator.nu wouldn't be much easier than writing a separate front end that called Validator.nu using the Web service API and did the sorting
  56. # 18:17 < zcorpan> hsivonen: it could be javascript
  57. # 18:18 < hsivonen> zcorpan: true. I hadn't though of that
  58. # 18:18 < zcorpan> btw, do you think it would be useful for you if the PFWG produced schemas for XHTML+ARIA?
  59. # 18:19 < hsivonen> zcorpan: which XHTML?
  60. # 18:20 < hsivonen> zcorpan: which schema language?
  61. # 18:20 < zcorpan> not sure
  62. # 18:20 < hsivonen> I'm probably going to write XHTML5+ARIA in due course
  63. # 19:22 < aroben> Lachy: yt?
  64. # 19:23 < Lachy> y
  65. # 19:24 < aroben> Lachy: do you think http://esw.w3.org/topic/HTML/Guide/ is an appropriate place for an overview of <canvas>?
  66. # 19:24 < aroben> Lachy: should the offline apps overview be merged in as well?
  67. # 19:26 < Lachy> aroben, if it's something that would be suitable for the Web Dev's Guide to HTML 5, then sure
  68. # 19:27 < aroben> Lachy: the offline apps overview is certainly something that developers will want to read, though I don't think it's developer-specific
  69. # 19:28 < Lachy> are you talking about an overview that already exists somewhere?
  70. # 19:28 < aroben> Lachy: for offline apps, yes
  71. # 19:29 < Lachy> link?
  72. # 19:29 < aroben> Lachy: Hixie and anne wrote it
  73. # 19:29 * aroben finds it
  74. # 19:29 < Lachy> this thing http://dev.w3.org/html5/offline-webapps/ ?
  75. # 19:29 < aroben> Lachy: yes
  76. # 19:33 < Lachy> it might be suitable, I'll have to think about it. I wasn't intending to look at that stuff for quite a while and will focus on the more stable sections
  77. # 19:34 < aroben> Lachy: I was thinking about starting to write a <canvas> overview, and so I wanted to figure out if I should just do it within the developer's guide
  78. # 19:39 < Philip`> aroben: Do you have ideas of what the overview would include? (I think duplicating the tutorial content that's already on the Mozilla wiki wouldn't be useful, but there are other things that aren't clearly written anywhere yet (as far as I'm aware))
  79. # 19:40 < hsivonen> is there a vector PDF whatwg logo available?
  80. # 19:40 < aroben> Philip`: I've only just started thinking about it
  81. # 19:41 < aroben> Philip`: existing documents I'm aware of are
  82. # 19:41 < aroben> http://developer.mozilla.org/en/docs/Drawing_Graphics_with_Canvas
  83. # 19:41 < aroben> http://developer.mozilla.org/en/docs/Canvas_tutorial
  84. # 19:41 < aroben> http://developer.apple.com/documentation/AppleApplications/Conceptual/SafariJSProgTopics/Tasks/Canvas.html
  85. # 19:42 < aroben> http://developer.apple.com/documentation/AppleApplications/Reference/SafariJSRef/Classes/Canvas.html
  86. # 19:45 < Philip`> aroben: Okay - I can't think of any other relevant documents except for scattered pieces in blogs and mailing lists
  87. # 19:45 < Lachy> aroben, you're welcome to contribute stuff about canvas if you like
  88. # 19:46 < Philip`> Oh, there's also some book stuff like http://safari.oreilly.com/0596102437/html_canvas which I can't read
  89. # 19:51 < aroben> Lachy: thanks, I guess we can always separate it out if we think that's more appropriate
  90. # 19:51 < aroben> Lachy: but I'd like to err on the side of fewer documents
  91. # 20:37 < Hixie> hsivonen: no special feature, just some shapes
  92. # 20:40 < hsivonen> Hixie: ok seems tedious. I implemented each year as a slide
  93. # 20:41 * gsnedders draws a circle around Hixie, and a square around hsivonen
  94. # 20:42 < hsivonen> (aside: I'm in awe over the level of suckiness of Keynote's chart tools and even more in awe over the suckiness of Numbers in general when trying to make something to paste to Keynote)
  95. # 20:48 < Hixie> hsivonen: oh it was _very_ tedious
  96. # 20:48 < Hixie> hsivonen: but it worked much better (imho) to have it all on one screen
  97. # 20:48 < Hixie> hsivonen: what version?
  98. # 20:51 < hsivonen> Hixie: '08
  99. # 20:54 < Hixie> hsivonen: ah. well. compared to v1, believe me, '08 is a dream.
  100. # 20:54 < Hixie> hsivonen: (also a dream compared to ppt)
  101. # 21:10 < hsivonen> hrm. Safari does not print SVG
  102. # 21:11 < hsivonen> Minefield print SVG but as bitmaps
  103. # 21:12 < hsivonen> and Opera crashes printing SVG
  104. # 21:13 < hsivonen> Prince to rescue
  105. # 21:18 < othermaciej> Safari doesn't print SVG?
  106. # 21:18 * othermaciej is mildly surprised
  107. # 21:19 < hsivonen> othermaciej: well, at least does not preview or save as PDF
  108. # 21:21 < othermaciej> does it refuse, or does the PDF come out blank?
  109. # 21:21 < hsivonen> othermaciej: comes out blank

The end :)