Rich Web Applications Backplane Weekly Call -- 09 Sep 2008

<Charlie> Meeting: Rich Web Backplane Weekly telecon

<Charlie> ah...sorry

<Charlie> go for it

<scribe> Scribe: Gregory_Rosmaita

<scribe> ScribeNick: oedipus

Previous: http://lists.w3.org/Archives/Public/public-xg-app-backplane/2008Aug/0021.html
... http://www.w3.org/2008/08/19-backplane-minutes.html

<Charlie> k

Announcements & News

Kevin: can only stay for part of the meeting

CW: topics for today? sent out strawman for TPAC

http://lists.w3.org/Archives/Public/public-xg-app-backplane/2008Sep/att-0007/loan_v.xhtml

CW: multi-modal, voice and xml events
... start with TPAC planning?

<Zakim> Steven, you wanted to ask about CDF

SP: CDF and @role

CW: particular question?

SP: kevin was chair of CDF -- CDF at CR stage - since that is about combining techs together, is there a cross-over and compatibility check with CDF

CW: kevin, where is CDF

Kevin: ensuring aligned with backplane, why in group; stuck in a rechartering mode working with AC rep to be allowed to rejoin; charter change, and change in IP policy; CDF WG membership dropped from 25 to 5 because all had to rejoin; no longer chair, looking for new chair now
... everyone has to rejoin individually; AC review of charter, etc.
... CDF probably out-of-action until end of year
... hopefully restart in february 2009

CW: understand general CDF framework - where are the touchpoints with RWAB

Kevin: both frameworks - doesn't impinge on anything planning for backplane; like to see backplane XG get to state where create profile based on CDF base profile and get content created on that profile with ideas working in RWAB

CW: XForms & GUI stuff; voiceXML profile, etc.

Kevin: yes, like to reuse CDF work on identification and slide profile underneath CDF infrastructure; hope RWAB proves CDF framework is viable; designed for the type of work RWAB doing

CW: deliverables?
... expressing in generalized way in profile will be good

Kevin: assumed at end of XG, will have profile for CDF

<Zakim> jackjansen, you wanted to ask for intro reading on cdf

CW: examine example profiles and how to approach

JJ: intro reading/background on CDF you would recommend?

Kevin: will put pointers into IRC

CW: best practices profile
... modularization discussions in Forms WG; what spec or schema looks like

Kevin: only have 1 profile currently
... mobile flavor (SVG tiny + XHTML) and full SVG + XHTML profile

CW: look at those in a couple of weeks-- have kevin walk us through

Kevin: not available next week; should be on 9/21/2008 call; away end of month

<Kevin> CDF Framework (intro reading) http://www.w3.org/2004/CDF/Group/specs/CDR/wp1/cdf.xml

CW: shoot for 21 September 2008 for profile intro and questions
... rechartering, and then...

Kevin: several candidates for other profiles; should meet with CDF if WG members there; SMIL timing profile been on radar

Jack: hard to contact over last few months

CW: good direction - do have meeting scheduled with MultiModal group

<Kevin> WICD profile Core http://www.w3.org/2004/CDF/Group/specs/CDR/wp1/wicd.xml

CW: trying to tiptoe around HTML versus XHTML WG politics; deal strictly with XML Events and VoiceXML

<Kevin> WICD mobile profile (based on WICD Core and CDF Framework) http://www.w3.org/2004/CDF/Group/specs/CDR/wp1/wicd-mobile.xml

CW: anything else?

<Kevin> WICD Full http://www.w3.org/2004/CDF/Group/specs/CDR/wp1/wicd-full.xml

TPAC Planning

CW: formal call for presentations on wednesday track not out yet

SP: ball in my court now, so need to follow up
... vague theme about different techs working together - interoperability proposals - can send in proposal pre-call

CW: how long?

<Kevin> why CDF article http://www.axistive.com/how-does-compound-document-framework-benefit-us.html

SP: standard TP length - lightning talks and half-hour slots
... RWAB would be a 30 minute slot

http://lists.w3.org/Archives/Public/public-xg-app-backplane/2008Sep/att-0007/loan_v.xhtml

CW: been 2 years since had backplane panel (linked-to from wiki)
... generally got interest going

<Kevin> some CDF precursor articles http://www-128.ibm.com/developerworks/xml/library/x-cxdp1/ & http://www-128.ibm.com/developerworks/xml/library/x-cxdp2/

CW: want to get people excited; remind people of ideas that came out of 2006 panel - sharing submission and data; patterns of interactive design; when something ocurs in XForms page, may be message in data model; instead of recalculating, allow other parties to weigh in on default action
... more collaborative and transparent approach -- what RWAB needs
... illustrate different things from same data sources
... experimental approach, not spec writing approach;
... Jack
... Jack's stuff is ready; new example up - VoiceXML support in FF -- have what needed to make work
... Mark Birbeck suggested a second talk -- coordinated discussions; makes sense, but may be too long and off topic for TPAC this year
... work on interaction alternatives?

SP: looks good to me; probably won't be talking in session; who would like to talk and what to demo

Kevin: concern is that we show something new -- VoiceXML and extensions been demoed

CW: concern of mine, too -- need an angle

Kevin: demo needs to show something people haven't seen before

<jackjansen> offtopic: what's x+v?

<Steven> XHTML+VOICE

<Steven> Implemented in Opera

<jackjansen> steven, thanks

CW: targetted at streamlined syntax of XForms - explaning HTML4 and HTML5 in terms of XForms - sync element hooking up to HTML input field, MVC arcitecture not available in HTML4 - depper architecture happening on GUI -- hooking up Voice with streamlined Xforms; issues a lot clearer with streamlining

<Steven> http://www.w3.org/TR/xhtml+voice/

CW: thoughts?

Kevin: what would be too subtle?

CW: explain the evolution of streamlined XForms syntax - interpret HTML4/5 element which adds a lot of power, including validation; put in Voice dialog via data elements - voice form could be economical - don't need sync elements - don't have to invent new markup to implement; coordination part of story; GUI part of story
... let Forms group take care of background?
... devil in the details -- refresh interest

JJ: not only Voice, but having a real data model behind form -- help auto-filling of forms

CW: that's my point about streamlined syntax and why using that approach is better for Voice hookup; may not see in markup or care about at runtime, but if do care, it is there; register event handler on form element as off XForm control and sync to Voice - show can do better job with Multimodal interaction
... if pair with MarkB and let him give overview of streamlining; explain backplane using xforms to give fresh look at voice interaction
... model is there for SMIL case

JJ: would be great - you can do it?
... will be there monday and tuesday only

CW: will be presenting on wednesday

Discussion of Straw Man Streamlined Example

http://lists.w3.org/Archives/Public/public-xg-app-backplane/2008Sep/att-0007/loan_v.xhtml

CW: ARIA implementation? just begining to streamline in ubiquity process - transforming cannonical XForms syntax; overloading HTML/XHTML with XForms or vice versa
... not use streamline version of loan form

GJR: work off version of form demoed at last meeting

CW: yes

GJR: ok
... calendar widget biggest piece of work

<Charlie> http://code.google.com/p/ubiquity-xforms/wiki/SimplifiedLoanForm

<scribe> ACTION: Gregory - provide roadmap for ARIA enabling loan form (simplified loan form) - due 2008-09-15 [recorded in http://www.w3.org/2008/09/09-backplane-minutes.html#action01]

GJR: using streamlined attachment get an error message from FF3.01

XML Parsing Error: prefix not bound to a namespace

CW: semantics vis a vis HTML5 and XHTML; Voice interaction implications; formal presence of model and instance and behavior of instances in terms of events that fire, can then reuse events more robustly than before
... what does it mean to write this from a "markup" Point-of-View
... input markup that tranforms/transcodes into underlying xforms markup (replace INPUT with XForms namespace element); example: LABEL element - transformation on input doc before executing or is this the runtime execution as well are we overloading xforms semantics?
... explore some of the markup
... main goal: approach author with mindset of writing web page (visual elements on screen then associating interactivity) - support that use case, but explain what can be done behind scene by finding right architecture
... a lot of attributes put on INPUT element; lot of elements would expect in bind or data element type
... attributes on control; define in streamline notation is how attributes used to generate formal elements, bindings, and constraints and used to define elements to which attached
... avoid XPath subtleties - evaluation context; want to avoid need for tree navigation in XPath expressions;
... cumbersome notation if not coming from heirarchal data model; HTML author wants flat model
... implicitly define in terms of siblings; when author uses @name in redefined context world

SP: streamlined syntax is like putting FONT element back into HTML so authors don't have to be exposed to "complication" of CSS selectors; whole effort tailored to HTML5 -- appreciate that a lot of demand, but think win of using it far outweighs anything else
... for same reason CSS is good idea, XForms is good idea

JJ: agree with Steven; hiding what appears to be complex behind simple syntax is not good idea in long run; but from practical POV, looks very much like HTML forms - will encourage use of XForms; at some point streamlined model will break down; cannonical XForms still needs to be end-path for migration; relearn new language or will there be help for that?

CW: XForms namespaced elements with HTML; add explicit bind to provide additional constraints, can chuck in XForms notation when needed in midst of streamlined notation
... technically feasible, if not philosophically popular

JJ: see problems

CW: events - allow people to register listener expecting to see events bubbling up out of live elements from refresh or recalculate
... that is what MarkB describing - behavior important, not markup
... same set of events as if using cannonical XForms

JJ: will come from very strange place to someone only familiar with streamlilned syntax

CW: for someone who wants more out of standard form - revalidation, etc. --

JJ: if push streamline, going to attach to pieces of javascript to do what can't do with constraints -- that is breaking point

CW: attach xforms behaviors via CSS selectors in ubiquity project; when change to rule, that will fire; why MarkB substituting XForms input for HTML input - remove ability to script, would break model
... has to be something for runtime processing
... collaboration with HTML WG will be key - bridging XForms and HTML4/5
... wanted to get unified architecture for HTML5 forms and XForms
... trying to do implementation for this form to get running
... opportunities for discussion

SP: so many more things that can be done rather than pull over a bunch of people who object not to XForms, but to namespaces

CW: report on AJAX claimed that obstacle in AJAX to do rich internet apps, is lack of higher declarative markup
... bit of backlash against pure javascript - at least in analyst community

SP: backlash inevitable -- HTML approach of just sticking stuff in, come up against barriers (no exension points - all power in hands of UA devs) - how to get implemented when UA devs won't play; proof that XML extensible and showing why it is the right thing to do, rather than throw a lifeline to those who don't get it anyway

CW: JohnB mentioned that if view as analog of what dojo doing (expansive decoration of web page) -- look at as HTML input - expands into DOM node as XForms behavior
... not just a line thrown to non-namespacers, but of utility itself; alternate to XForms syntax and not degenerate notation for ascii html world
... not do in HTML namespace, but extension of XForms namespace
... duration input element treat as input with name support
... shorthand notation for what already exists
... clearer what we need to be careful about - both points of view may be broached at TPAC
... implications for SMIL Jack doing or Voice integration i proposed -- help? muddy waters?

JJ: orthoganal - personally would use full XForms stuff, but this wouldn't hurt

CW: presentation - go through sequence XForms to second iteration with streamlined syntax (same event handlers)

JJ: path don't want to go down; in web apps stuff, emphasize base model as API - want fairly strict control over data model; how will streamlined syntax provide a data model and how? if want XForms to integrate with SMIL, better to use actual XForms

CW: fair amount of magic

JJ: if someone writes web app with streamlined XForms and SMIL will have to make mental translation, so why not do explicitly?

CW: explicit nature of model is what is wanted
... streamlined syntax for guy just doing form - otherwise needs too much magic

JJ: yes; big advantage of streamlined format is looks like HTML forms; if had web page with simple form and submit, would use HTML form; XForms pretty big for simple forms

SP: simple forms in XForms are also simple

JJ: 2 input fields and a button, fine; but when more complicated, can't do a lot of things with HTML forms

SP: if try to do in HTML4 would be more complicated

JJ: once start to have repeat, need xforms

CW: need to think a bit about presentation - clear story for streamline? if not, what about Voice demo?
... show not SMIL not Voice - simple focused example of pattern
... top of hour - adjourn

ADJOURNED

<Steven> Gregory?

yes, steven

<Steven> How do you like to be listed in an acknowledgments section?

<Steven> Do you have an affiliation?

Gregory J. Rosmaita, invited expert

<Steven> Ack

<Steven> thanks

thank you

<Charlie> Uli

<Charlie> yes

<Charlie> Ulrich Nicolas Lisse

<Charlie> yup

- DRAFT -

Rich Web Applications Backplane Weekly Call

09 Sep 2008

Attendees

Contents

Announcements & News

TPAC Planning

Discussion of Straw Man Streamlined Example

Summary of Action Items

Scribe.perl diagnostic output