fluid-work IRC Logs-2008-04-22

[04:43:17 EDT(-0400)] * phiggins (n=dante@c-68-34-199-67.hsd1.tn.comcast.net) has joined #fluid-work
[08:27:14 EDT(-0400)] * anastasiac (n=team@142.150.154.105) has joined #fluid-work
[08:58:23 EDT(-0400)] * colinclark (n=colin@bas1-toronto09-1279475430.dsl.bell.ca) has joined #fluid-work
[09:37:27 EDT(-0400)] * theclown (n=theclown@142.150.154.101) has joined #fluid-work
[09:58:16 EDT(-0400)] * EricDalquist (n=dalquist@bohemia.doit.wisc.edu) has joined #fluid-work
[10:09:31 EDT(-0400)] * colinclark (n=colin@142.150.154.101) has joined #fluid-work
[10:10:26 EDT(-0400)] * michelled (n=team@142.150.154.197) has joined #fluid-work
[12:02:54 EDT(-0400)] * slightlyoff (n=alex@204.14.154.209) has joined #fluid-work
[13:10:07 EDT(-0400)] * phiggins (n=dante@c-68-34-199-67.hsd1.tn.comcast.net) has joined #fluid-work
[13:45:01 EDT(-0400)] * Bosmon (n=Antranig@ginger.caret.cam.ac.uk) has joined #fluid-work
[13:45:14 EDT(-0400)] <Bosmon> WE ARE GOING TO HAVE TO IMPLEMENT OUR OWN SELECTOR ENGINE!!!!!!!!
[13:54:31 EDT(-0400)] <phiggins> ?
[13:56:24 EDT(-0400)] * davidb (n=davidb@142.150.154.101) has joined #fluid-work
[13:57:11 EDT(-0400)] <davidb> colinclark: i've performed a Window-Eyes experience test with the existing jq tabs from fluid.
[13:57:21 EDT(-0400)] <davidb> colinclark: and i've written up a summary of the dijit approach
[13:57:33 EDT(-0400)] <davidb> colinclark: with observations based on WE (window eyes) experience
[13:57:48 EDT(-0400)] * davidb proceeds to doctor jquery tabs...
[13:58:33 EDT(-0400)] <davidb> btw all, the dijit tabcontainer window-eyes experience rocks!
[13:58:43 EDT(-0400)] <davidb> (WE 6.1)
[14:11:19 EDT(-0400)] * davidb looks for cc
[14:37:51 EDT(-0400)] <davidb> correction, we dijit tabcontainer experience almost rocks
[14:38:01 EDT(-0400)] <davidb> we/Window Eyes
[14:41:52 EDT(-0400)] <colinclark> davidb: I'm here. Just distracted.
[14:41:54 EDT(-0400)] * phiggins (n=dante@c-68-34-199-67.hsd1.tn.comcast.net) has joined #fluid-work
[14:42:06 EDT(-0400)] <davidb> np
[14:42:09 EDT(-0400)] <colinclark> davidb: Nice work on the jQuery Tabs. I'm impressed!
[14:42:23 EDT(-0400)] <davidb> colinclark: right back at ya
[14:42:30 EDT(-0400)] <colinclark> Bosmon: We are certainly not going to implement yet another selector engine!
[14:42:43 EDT(-0400)] <davidb> lol
[14:43:01 EDT(-0400)] <davidb> colinclark: yes call it jQuiffy
[14:43:03 EDT(-0400)] <colinclark>
[14:43:35 EDT(-0400)] <colinclark> I call it jCrazy!
[14:48:11 EDT(-0400)] <davidb> colinclark: okay the WE experience is decent in your tabs i can go over browse mode for the panes with you later.
[14:48:38 EDT(-0400)] <colinclark> Sweet@
[14:48:40 EDT(-0400)] <colinclark> !
[14:48:46 EDT(-0400)] <colinclark> Our tabs.
[14:48:49 EDT(-0400)] <colinclark> Klaus' tabs.
[14:49:05 EDT(-0400)] <davidb> woot! theclown has the WE 6.1 install cd... no more "your demontsration time is about to expire"
[14:49:26 EDT(-0400)] <davidb> colinclark: yeah the greater "your"
[14:49:31 EDT(-0400)] <davidb> ok our
[14:49:41 EDT(-0400)] <colinclark>
[14:49:54 EDT(-0400)] <colinclark> Someone recently called me a communist, which is silly.
[14:49:56 EDT(-0400)] <Bosmon> HA!
[14:49:58 EDT(-0400)] <Bosmon> COLIN!
[14:50:01 EDT(-0400)] <Bosmon> YES YES!
[14:50:02 EDT(-0400)] <Bosmon> WE are!!!
[14:50:08 EDT(-0400)] <colinclark> But we could call them the Collective's Tabs.
[14:50:16 EDT(-0400)] <davidb> i'm going to check the aria specs... the experience is similar with these tab containers.. but the roles are used slightly differently.
[14:50:18 EDT(-0400)] <colinclark> Bosmon: No, we're not.
[14:50:18 EDT(-0400)] <davidb> colinclark: ^
[14:50:21 EDT(-0400)] <Bosmon> Yes yes yes yes
[14:50:43 EDT(-0400)] <davidb>
[14:50:49 EDT(-0400)] <Bosmon> Because it is not fast enough to get "innerHTML" out of crabbed things inside the DOM
[14:50:53 EDT(-0400)] <colinclark> davidb: Cool.
[14:50:57 EDT(-0400)] <Bosmon> We will have to do it FOR PERFORMANCE REASONS!!!
[14:51:18 EDT(-0400)] <colinclark> Bosmon: File a jQuery bug and we'll talk to Resig about it if there is a performance problem.
[14:51:33 EDT(-0400)] <Bosmon> I doubt Resig can fix the speed of innerHTML
[14:51:40 EDT(-0400)] <Bosmon> He is a genius, but not a magician
[14:51:43 EDT(-0400)] <colinclark>
[14:51:50 EDT(-0400)] <Bosmon> Anyway
[14:51:51 EDT(-0400)] <davidb> Bosmon: you can tell resig how to write better javascript
[14:51:57 EDT(-0400)] <colinclark> So what is it that you propose?
[14:51:59 EDT(-0400)] <Bosmon> Even more exciting speculations!!!!!!!
[14:52:09 EDT(-0400)] <Bosmon> (19:45:57) Antranig: AAAAAAH!
[14:52:09 EDT(-0400)] <Bosmon> (19:46:02) Antranig: We can have a DUCK TYPING IKAT!
[14:52:09 EDT(-0400)] <Bosmon> (19:46:09) Antranig: I have been thinking too JAVAER-y
[14:52:09 EDT(-0400)] <Bosmon> (19:46:19) Antranig: We don't need all these fucking component classes
[14:52:09 EDT(-0400)] <Bosmon> (19:46:39) Steve:
[14:52:10 EDT(-0400)] <Bosmon> (19:47:01) Antranig: Basically any String in itself represents a UIOutput
[14:52:12 EDT(-0400)] <Bosmon> (19:47:21) Antranig: Any structure with a field called "valuebinding" potentially represents a UIBound
[14:52:14 EDT(-0400)] <Bosmon> (19:47:37) Antranig: With the potential presence of "willinput" settling the the debate about whether it will input or not
[14:52:17 EDT(-0400)] <Bosmon> (19:48:04) Antranig: To create a "UIVerbatim" you just slap the string in a field called "makrup"
[14:52:19 EDT(-0400)] <Bosmon> (19:48:05) Antranig: markup
[14:52:21 EDT(-0400)] <Bosmon> (19:48:38) Antranig: An Array of anything naturally represents a bunch of BranchContainer peers
[14:52:23 EDT(-0400)] <Bosmon> (19:48:52) Antranig: Which if any of them contain "jointID" are joints
[14:52:30 EDT(-0400)] <anastasiac> davidb: I'm trying to catch up here: what Query tabs work are you referring to / doing?
[14:53:41 EDT(-0400)] <davidb> anastasiac: hi, the ones for the demo next week. the fluid-tweaked jquery tabs
[14:53:42 EDT(-0400)] <colinclark> Bosmon: But why does this mandate a new selector engine?
[14:53:56 EDT(-0400)] <davidb> anastasiac: have you tried with screen reader? we should talk.
[14:54:19 EDT(-0400)] <Bosmon> No, this is a separate track
[14:54:21 EDT(-0400)] <Bosmon> 2 different ideas!
[14:54:32 EDT(-0400)] <davidb> anastasiac: sandbox/tabindex/trunk/examples/jquery-ui-tabs
[14:54:33 EDT(-0400)] <colinclark> Aha. Multitasking.
[14:55:17 EDT(-0400)] <colinclark> I approve.
[14:55:52 EDT(-0400)] <colinclark> k
[14:56:06 EDT(-0400)] * colinclark is typing in the wrong window.
[14:56:26 EDT(-0400)] <anastasiac> ok, thanks davidb. haven't tried with screen reader - yes, we should talk
[14:58:14 EDT(-0400)] <Bosmon> No, we need the selector engine to be able to specify "cutpoints" within the document
[14:58:43 EDT(-0400)] <Bosmon> But, given a selector cutpoint, we need to be able to go from there to a relevant template node ("XMLLump") very quickly
[14:59:17 EDT(-0400)] <Bosmon> Really the "template" representing a document can only afford to be parsed just once
[14:59:39 EDT(-0400)] <Bosmon> But after this point, the natural association between "DOM elements" and "template elements" has broken down
[14:59:45 EDT(-0400)] <anastasiac> davidb: have you tried the screen reader on http://build.fluidproject.org/sakai-imagegallery-tool/sample-code/reorderer/jquery-tabs/jquery-tabs.html ?
[14:59:52 EDT(-0400)] <Bosmon> And reestablishing this association "node by node" will simply be TOO EXPENSIVE
[14:59:55 EDT(-0400)] <anastasiac> this is the vanilla jquery tabs, with the reorderer added
[14:59:56 EDT(-0400)] <davidb> anastasiac: NO
[15:00:06 EDT(-0400)] <davidb> sorry didn't mean to yell!
[15:00:18 EDT(-0400)] * anastasiac was wondering
[15:00:23 EDT(-0400)] <Bosmon> We cannot use the DOM structure as our own template respresentation, its performance characteristics are too appalling
[15:00:24 EDT(-0400)] <davidb> i can't focus on this atm talk later?
[15:00:27 EDT(-0400)] <davidb> anastasiac: ^
[15:00:31 EDT(-0400)] <Bosmon> Worse even than those of vanilla Javascript
[15:00:40 EDT(-0400)] <anastasiac> of course, davidb, no problem
[15:00:44 EDT(-0400)] <colinclark> Bosmon: Thank goodness you walked me through all this code. You're actually making some sense.
[15:00:45 EDT(-0400)] <davidb> k
[15:00:58 EDT(-0400)] <colinclark> But we're not writing a whole new selector engine!
[15:01:09 EDT(-0400)] <Bosmon> I can't see an easy alternative
[15:01:20 EDT(-0400)] <Bosmon> We have mere milliseconds to get this thing working
[15:01:24 EDT(-0400)] <colinclark> lol
[15:01:28 EDT(-0400)] <Bosmon> SAVE THE MILLISECONDS
[15:01:54 EDT(-0400)] <colinclark> Well, if you use the OCTACORE like anastasiac does, you have plenty milliseconds to spare.
[15:01:59 EDT(-0400)] <Bosmon> Yes
[15:02:09 EDT(-0400)] * anastasiac gloats
[15:02:11 EDT(-0400)] <Bosmon> I will not repeat my comments on the OCTOCORE
[15:02:38 EDT(-0400)] * colinclark realizes he has been spelling OCTOCORE wrong.
[15:02:41 EDT(-0400)] <colinclark> Eek.
[15:02:57 EDT(-0400)] <Bosmon> Anyway, it will have to be our own selector engine in any case.... since at least one form of selector it needs to support is "old-style" rsf:id
[15:03:28 EDT(-0400)] <colinclark> Bosmon: You expect it to support other forms? Aside from the potential of a plain id selector?
[15:03:34 EDT(-0400)] <Bosmon> Yes, why not
[15:03:41 EDT(-0400)] <colinclark> Why?
[15:03:43 EDT(-0400)] <Bosmon> We want this to be "idiomatic" to existing JQuery users
[15:03:55 EDT(-0400)] <Bosmon> They should not be forced to put weird things in their documents that they don't necessarily want to
[15:04:00 EDT(-0400)] * anastasiac looks up the propoer spelling of octo/octa, and discovers that octa is also correct
[15:04:12 EDT(-0400)] * colinclark is relieved.
[15:04:19 EDT(-0400)] * Bosmon blargs
[15:04:28 EDT(-0400)] <Bosmon> blarg
[15:04:34 EDT(-0400)] <colinclark> Bosmon: But the cost of such a thing is inordinate.
[15:04:41 EDT(-0400)] <Bosmon> Nonsense
[15:04:47 EDT(-0400)] <colinclark> And seemingly insane.
[15:04:57 EDT(-0400)] <Bosmon> Nothing compared to the cost of imprisoning humanity within a set of Concrete Eggcups
[15:06:35 EDT(-0400)] <Bosmon> So, here is another issue for the group - "nonexistence of outerHTML"
[15:06:38 EDT(-0400)] <Bosmon> <thead><tr><th>Count</th><th>Name</th><th rsf:id="header:">1</th><th>Median ave</th></tr></thead>
[15:06:44 EDT(-0400)] <Bosmon> Imagine the following "template segment"
[15:06:59 EDT(-0400)] <Bosmon> In performing a "self-rendering" we would naturally want to expand the "header:" node to any number of replicates
[15:07:11 EDT(-0400)] <Bosmon> The problem with this is that there is no easy "all-in-one" DOM operation which can effect this
[15:07:33 EDT(-0400)] <Bosmon> If there is any node we can assign to the "innerHTML" of, it will have to be the <thead> which is incredibly wasteful
[15:07:46 EDT(-0400)] <Bosmon> And "outerHTML", whilst it exists on IE, exists nowhere else
[15:08:57 EDT(-0400)] <davidb> colinclark: rich has added tablist to the list of containers that manage their children (internal draft grrrr)
[15:09:09 EDT(-0400)] <davidb> colinclark: so you are ahead of the game
[15:10:50 EDT(-0400)] * colinclark is catching up...
[15:11:11 EDT(-0400)] <colinclark> davidb: What containers don't manage their children?
[15:11:20 EDT(-0400)] <davidb> colinclark: dunno
[15:13:14 EDT(-0400)] <colinclark> Bosmon: I'm reading...
[15:14:24 EDT(-0400)] <colinclark> Bosmon: You're correct. I don't know of any easy way to do this.
[15:15:56 EDT(-0400)] <Bosmon> Yet it is essential
[15:16:28 EDT(-0400)] <Bosmon> How can we achieve it?
[15:16:59 EDT(-0400)] <Bosmon> I believe this is also "comes together" with the issue of corresponding "DOM spines" to template nodes
[15:17:09 EDT(-0400)] <colinclark> Bosmon: I'm in the midst of something else. Can you double check that none of jQuery's DOM manipulation functions will work?
[15:17:11 EDT(-0400)] <Bosmon> DOM elements are such crappy things
[15:17:22 EDT(-0400)] <Bosmon> OK, never mind, we can talk next week
[15:17:39 EDT(-0400)] <colinclark> They all seem to be intended to handle HTML strings, not just DOM elements.
[15:18:14 EDT(-0400)] <Bosmon> It should be "instant", given any block of innerHTML, to be able to splice out some relevant bit of
[15:18:23 EDT(-0400)] <Bosmon> I guess writing our "lunatic fast parser" is the only way out of this
[15:21:03 EDT(-0400)] <michelled> colinclark, Bosmon: jQuery.before or jQuery.after should do what you want
[15:24:23 EDT(-0400)] <colinclark> Bosmon: Is that what you were looking for?
[15:24:27 EDT(-0400)] <Bosmon> I am reading
[15:24:38 EDT(-0400)] <Bosmon> But instead I seem to have bumped into this page
[15:24:39 EDT(-0400)] <Bosmon> http://ejohn.org/blog/i-learned-some-things-about-jquery-today/
[15:26:54 EDT(-0400)] <Bosmon> http://jquery.com/api/
[15:26:58 EDT(-0400)] <Bosmon> I'm rather disappointed by this page
[15:27:01 EDT(-0400)] <Bosmon> Firstly it is URL-less
[15:27:12 EDT(-0400)] <Bosmon> Seconly there are lots of things which are underlined which don't really seem to do anything
[15:27:16 EDT(-0400)] <Bosmon> I tried about 4 browsers on them...
[15:28:01 EDT(-0400)] <colinclark> Bosmon: That page is ancient!
[15:28:07 EDT(-0400)] <Bosmon> ....
[15:28:14 EDT(-0400)] <Bosmon> Well, it is hit #1 for "JQuery APi"
[15:28:24 EDT(-0400)] <Bosmon> OK, i have found more things now...
[15:28:40 EDT(-0400)] <colinclark> jQuery is up to 1.2.3 now.
[15:28:41 EDT(-0400)] <Bosmon> Yeah, JQuery.after seems very cool
[15:28:46 EDT(-0400)] <colinclark> http://docs.jquery.com/Main_Page
[15:28:57 EDT(-0400)] <colinclark> Bosmon: See, I told you!
[15:28:58 EDT(-0400)] <Bosmon> Just need to combine it with an operation that destroys the original thing....
[15:29:12 EDT(-0400)] <Bosmon> Is there a particularly readable version of the JQuery source?
[15:29:17 EDT(-0400)] <Bosmon> I would like to see how it is implemented
[15:29:22 EDT(-0400)] <colinclark> I find the trunk version quite readable.
[15:29:27 EDT(-0400)] <Bosmon> ok
[15:29:40 EDT(-0400)]

[15:29:58 EDT(-0400)] <colinclark> More than four...
[15:29:59 EDT(-0400)] <colinclark> http://jqueryjs.googlecode.com/svn/trunk/jquery/src/
[15:30:01 EDT(-0400)] <Bosmon> Ah
[15:30:03 EDT(-0400)] <Bosmon> This is interesting
[15:30:16 EDT(-0400)] <colinclark> Bosmon: It's quite beautifully written.
[15:30:36 EDT(-0400)] <Bosmon> Well, it is interesting, in that I think it will not work for our case
[15:31:53 EDT(-0400)] <Bosmon> The ultimate manipulation is simply a DOM.insertAfter
[15:32:07 EDT(-0400)] <Bosmon> Which of course could not work unless the inserted material contained a single root node
[15:33:53 EDT(-0400)] <Bosmon> Well, this one line is worth its weight in gold in any case
[15:33:55 EDT(-0400)] <Bosmon> return tag.match(/^(abbr|br|col|img|input|link|meta|param|hr|area|embed)$/i) ? all : front + "></" + tag + ">";
[15:34:05 EDT(-0400)] <Bosmon> We can just fold that into our mega-parser as is
[15:36:40 EDT(-0400)] <Bosmon> OK, this is a cool strategy
[15:36:44 EDT(-0400)] <Bosmon> But I worry it will be too slow
[15:36:57 EDT(-0400)] <Bosmon> Everyone should read JQuery.clean() when they get a moment
[15:37:54 EDT(-0400)] <Bosmon> It applies an innerHTML to a self-created Div, after various "de-plastering" operations, and then reads back all the top-level element children from the div
[15:38:04 EDT(-0400)] <Bosmon> Then, one-by-one, applies them to DOM.insertAfter
[15:38:33 EDT(-0400)] <colinclark> Bosmon: Yes.
[15:38:40 EDT(-0400)] <Bosmon> http://jqueryjs.googlecode.com/svn/trunk/jquery/src/core.js
[15:46:00 EDT(-0400)] * EricDalquist (n=EricDalq@ppp-70-226-161-218.dsl.mdsnwi.ameritech.net) has joined #fluid-work
[15:54:01 EDT(-0400)] * michelled (n=team@142.150.154.197) has left #fluid-work
[15:54:50 EDT(-0400)] <davidb> dang i gotta go! i have to buy a bday present for gwyneth's friend.
[15:54:58 EDT(-0400)] <davidb> catch you later.
[16:55:26 EDT(-0400)] * Bosmo1 (n=Bosmon@87-194-196-18.bethere.co.uk) has joined #fluid-work
[16:55:30 EDT(-0400)] * Bosmo1 (n=Bosmon@87-194-196-18.bethere.co.uk) has left #fluid-work
[17:18:26 EDT(-0400)] * theclown (n=theclown@142.150.154.101) has left #fluid-work
[19:22:07 EDT(-0400)] * phiggins (n=dante@c-68-34-199-67.hsd1.tn.comcast.net) has joined #fluid-work
[21:06:49 EDT(-0400)] * EricDalquist (n=EricDalq@ppp-70-226-161-218.dsl.mdsnwi.ameritech.net) has joined #fluid-work