W3C

- DRAFT -

app backplane XG teleconference

24 Jun 2008

Agenda

See also: IRC log

Attendees

Present
+31.20.616.aaaa, +34.91.549.aabb, Charlie, Gregory_Rosmaita, Jack_Jansen, John_Boyer, Steven
Regrets
Kevin, Uli
Chair
Charlie Wiecha
Scribe
Gregory_Rosmaita

Contents


 

 

<scribe> Scribe: Gregory_Rosmaita

<scribe> ScribeNick: oedipus

<jsalvachua> Hello, good morning

CW: markup examples - embedding in dojo
... First example of using ubiquity modules in AJAX: xforms model and submission in Dojo

http://www.webdialogs.com

<Charlie> webdialogs.com

<Charlie> meeting code is 801862

CW: welcome to Jack - joining us today
... brief introductions

Jack: member of SYMM WG work on SMIL; SMIL state

JJ: integration into web apps for SMIL states

CW: welcome

<jsalvachua> i think the noise its me

<jackjansen> oedipus, I am +31.20

CW: background

<Charlie> http://www.w3.org/MarkUp/Forms/wiki/XForms_Future_Features

XForms Face2Face Update

CW: spent good chunk of times refactoring XForms spec into modules
... XForms 1.2 M12n
... different from original backplane note (2006 Amsterdam)

<jsalvachua> 34 spain

CW: modulized XForms with finer grade modules; trying to reflect XHTML m12n
... backplane - implementing modules into host languages
... started in XG investigating ubiquity xforms (google code)

<John_Boyer> http://code.google.com/p/ubiquity-xforms/

JJ: had quick look - needs serious study

CW: Forms POV and Backplane POV is digestable in small pieces

<John_Boyer> For developers (consumers) of ubiquity, here is a google interest group: http://groups.google.com/group/ubiquity-xforms

CW: what if author just interested in Data Model and want to support in Web 2.0 application; just need to use relevant modules

JJ: ubiquity forms stuff very very interesting; no support on Safari

<jackjansen> agrees with steven

CW: by virtue of AJAX implementation, get out of plug-ins
... hope is twofold: 1) cross-platform implementation; 2) beyond the desktop (mobile, etc.)
... ubiquity and XG symbiotic relationship to extend ubiquity model with non-traditional types of content (beyond forms and HTML)

JB: couple of google groups for developers - consult URIs pasted above -
... how to use, rather how to develop

oedipus@hicom.net

<John_Boyer> Here is a link to a group that will be focused on consumers of the new ubiquity library: http://groups.google.com/group/ubiquity-xforms

CW: goal - sense of direction to drive examples - using ubiquity library for other purposes
... dojo data provider
... discuss support of custom controls in dojo with XForms wrappers

<John_Boyer> This is different than contributing to the ubiquity library, which has this other group: http://groups.google.com/group/ubiquity-xforms-eng

CW: what would be something to implement - email client? who are we addressing
... 1) who are we trying to impact and 2) choosing an example to code

<Steven> I am

<Steven> I know

JB: CW has to relaunch before everyone can join

<John_Boyer> Once Charlie launches: www.webdialogs.com/join code 801862

<Steven> I was waiting to be invited to talk

<Charlie> meeting is up you can rejoin

Steven: question who are we trying to impact - ourselves are the customers - all W3C WGs doing things with similar use cases but weork independently so all end up slightly different and think use cases different, see W3C and its WGs prime candidates - makes W3C Submit standard; XML Events 2 library; put as much in core as possible

CW: rather than focusing on AJAX integration and dojo integration - mix and match with other XML namespaces?

SP: that should be our aim; dojo as an implementation of concept and no more than that; in end, ideally, use cases should become markup without need to program out - find way to declaratively markup so can be supported natively or through script, but in the end gives you what you want

CW: return to SP's point after demo

<Steven> "Your Web browser does not meet the requirements

<Steven> for joining a presentation."

<Steven> Doesn't work for me

SP: doesn't work with opera

CW: do you have FF?

TL: follows dojo api which allow any dojo widget (XForms data module) do things server side to process; allows one to export data model, dojo stores model and uses DOM;
... XForms data module - accept data, on javascript side create adjacent object with data model id, formation id, and custom ??? id - uses XPath to add to bucket
... dojo object registered through DOM
... demo does search on picasa - used modular extensions and mozilla data model extension
... if type "XForms" and do a search, returns XML document; use dojo data storage to put image into thumbnail anchor
... can use title as well as thumbnail - if click on it, you get the image
... all done by dojo widget - data model allows dojo widget to crunch data from XForm - any dojo widget that could use this as way to send data up the web
... also have another example - how easy to use - picasa example
... bunch of images - one library lists - only thing diff between 2 is data model and call for item and call for form

JB: traditional dojo author would write script that wouldn't capture differences between picasa and flickr - 2 different declarative statements

TL: XPath retrieves data (images) to show picture

CW: Use case - differentiate interfaces using dojo - is it the kind of thing trhat has "legs" - injects some XForms modules into AJAX framework

JB: piece of XForms markup not very big so not doing much - just doing submission - what would look like without XForms markup?

TL: if don't have XForms markup, have to use data storage which has to fetch from web server - implementation in dojo that does that

CW: doing comparission might be a good thing - declarative layer boiled out of code

JJ: very excited - touches on demo i'm working on - can point to demo if keep confidential

<jackjansen> smil state demo: http://homepages.cwi.nl/~jack/NoBudget/

JJ: only works in safari on mac
... integrating SMIL and googlemaps and XForms and web content - big problem: tying all the stuff together
... dojo would have made life easier

CW: backplane POV would prefer XML Event notification between model and ticker; complete backplane compliance should have XForms value change ond notification events
... data model piece simplifies even if databinding is custom

JB: what does this look like without the XForms markup? if have XForms markup for submission piece alone, what if using YUI or other library - benefit of standardization - AJAX programmers don't all use dojo; want to standardize states and submission which makes them available to mujltiple application languages - actual markup languages as well

CW: get certain amount of simplification through declarative markup, but still framework issues if differences show through programming model

JB: more backplane - submission working in demo, even though not using Events; chunk of markup that could be used by dojo programmer or YUI programmer, or in SMIL app

CW: can capture results and use own notification
... would a dojo programmer want to use this?
... how to invest effort?
... build assets and compelling examples to motivate AJAX community to pay attention to our work
... helpful exercise for that? should that be our agenda?
... who are our audiences

JB: think is right tack
... excited by demo and want a copy to demonstrate
... even with small part of big picture - what is CW's comment about "smothering" related - markup could be reused if someone crteated the right "glue" could be used by dojo user, glue determines library

CW: misunderstood - want output useable by dojo and YUI user

JB: even this little demo shows benefit of encoding most important pieces - don't have to change dojo widget (strength of dojo) - gives dojo neew capaibility to provide data sources

CW: don't have to change script to go from flickr to picasa
... do YUI implementation to show binding to diff script objects

JB: right

CW: Events not burried in script - can register own listeners if use submit
... mix-and-match - element semantics in parallell with dojo widgets and listeners

TL: already added some but used dojo callback - in model argument, specify what kind of code you want

CW: listening directly with XML listener along with DOM listener

JB: if have SOAP based web server, could do some clever bits when information came back
... TL's code uses events (submitDone, submitError) in glue codes to tell dojo when and what info available
... backplane - events are key to how components talk to each other

CW: return to JJ's use case?

JJ: suggest people have a look at it if can
... question: always looked at XBL for glue language - particularly gluing changes in data model to javascript - does XBL have part to play here?

SP: XBL has good chance of enhancing work - seen excellent examples linking XHTML, XForms and SVG together

JJ: sample URI?

SP: XForms+SVG version of googlemaps
... not currently working; will ping author again

CW: ubiquity library in FF uses XBL to add behaviors to each XForms namespaced elements in page - calls conventional javascript object

SP: that's what XBL does

CW: yes, but use object attached with XBL - javascript could be used to manipulate DOM

SP: does ubiquity use XBL?

CW: no
... not too much XForms in XBL directly due to javascript limitations
... postpone custom controls discussion until next week
... what examples should we build?

<Steven> I vote for Google maps in XForms

CW: something more complete?
... agree? ideas?

SP: i vote for google maps in XForms - hundreds of thousands of lines in javascript - shrunk dramatically when put into declarative form (XForms)
... combine more than 1 W3C tech into example
... splitting google maps into functional part (looking at latitue and longitude at particluar zoom level) - rest presentational on top of essential function - just chaning latitude, longitude and zoom

CW: overlays, pinpointe, etc.
... other mashups?
... like example - clear separation of backplane from UI
... Silverlight in XForms in backplane - navigator for olympics as good as Silverlight?

GJR: Silverlight port to linux underway by mono project at novell

JB: no depository for demos

TL: picasa allows to create album

CW: google docs - calanders and such

GJR: would like to "see" one non-visually dependent example or an example of using declarative markup to reuse data stream

CW: feed reader example might be good - google example would be good too - entry point into data
... discuss on list components and stictch together next call

JB: with JJ on call and expertise with SMIL and SYMM - binding media presentations with underlying data - short presenation possible?

JJ: be gone for 4 weeks begining next week - submitted paper and if i can share, will post URI on list, otherwise will have to wait until presentatoin formally given

CW: pick up next week with custom controls

Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.133 (CVS log)
$Date: 2008/06/24 16:02:05 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.133  of Date: 2008/01/18 18:48:51  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Succeeded: s/webexamples.com/webdialogs.com/
Succeeded: s/: follows/TL: follows/
Succeeded: s/??:/TL:/G
Succeeded: s/TL:/TL:/
Succeeded: s/picassa/picasa/G
Succeeded: s/picasa/picasa/G
Found Scribe: Gregory_Rosmaita
Found ScribeNick: oedipus
Default Present: Charlie, +31.20.616.aaaa, John_Boyer, Gregory_Rosmaita, +34.91.549.aabb, Steven, jackjansen?
Present: +31.20.616.aaaa +34.91.549.aabb Charlie Gregory_Rosmaita Jack_Jansen John_Boyer Steven
Regrets: Kevin Uli
Agenda: http://lists.w3.org/Archives/Public/public-xg-app-backplane/2008Jun/0016.html
Got date from IRC log name: 24 Jun 2008
Guessing minutes URL: http://www.w3.org/2008/06/24-backplane-minutes.html
People with action items: 

WARNING: Input appears to use implicit continuation lines.
You may need the "-implicitContinuations" option.


[End of scribe.perl diagnostic output]