W3C

UWA WG F2f, Dublin, 7-8 June 2007

See also: IRC log 1, IRC log 2, IRC log 3

Face to face photo, Dublin, June 2007
Dublin, June 2007. From left to right: José Manuel Cantera Fonseca, Kevin Smith, Rhys Lewis, Stéphane Boyera, Dave Raggett, Giles Payne, Rotan Hanrahan, Sailesh Sathish, and Kangchan Lee

Attendees

Agenda
http://www.w3.org/2007/Talks/0607-dsr-uwa/
Present
José Manuel Cantera Fonseca, Kevin Smith, Rhys Lewis, Stéphane Boyera, Dave Raggett, Giles Payne, Rotan Hanrahan, Sailesh Sathish, and Kangchan Lee
Regrets
to be added ...
Chair
Dave

Contents

Additional Materials


Charter review

dave presents the charter based on his talk

KC: clearly bug problem concerning bridging between devices. big problems 2 systems completly un-interoperable

Sailesh: problem is the interconnecting world beow the device layer. no consensus yet. semantic interoperability is the ideal goal, but we not here yet

Dave: business driver are not yet clear.

KC: proprietary solutions: people are afraid to loose business share if they open their system

Giles: same in japan

Jose: Stand. group in Korea ?

KC: yes, home network appliance std group in Korea but after 1 one year, all the members left

Scribe: Sailesh

Dave: continuing with presentation given at xtech

Jose: binding: putting xml tag means binding to service?

dave: element becomes proxy for service
... images being resource

Ontologies

Rhys presents.

Rhys: Groups in OMA, W3C interested in properties associated with devices.

... good idea to have vocab

... not achieved in cc/pp

... simple one in UAProf

... good idea to use tools.

... Protege is leading Open Source tool for ontologies.

... (Demonstrates on screen)

... started using UAProf (as it exists).

Rhys demonstrates samples in Protege: currently showing PixelCount

Considering relationships between properties, and how change in value might affect other values.

Discussing how ontology would represent relationships between things held therein.

This includes representation of unit conversions, and derivative relationships (e.g. aspect ratio is derived from width and height).

Simple mechanisms within ontology to handle such relationships.

APIs can either use these directly, or use convenience functions for the common usage.

For example, default units.

Rhys: possible we could be looking at multiple ontologies. One for device properties, one for application behaviours, etc.

Dave: identify different classes of users of ontologies, so that demands of different users don't impede each other's work.

(Moves on to instances. Actual ontologoes for actual devices.)

Note that there are "Common Names", names as used in ontology may be different from names as used in APIs/vocabs etc.

(Moves on to DisplayHeight example.)

You have to have a class for each units of measure.

E.g. millimetres.

Note: there's a relationship between units and values. If the units changes, then the value needs to change too.

Rhys: conversion factors encoded in ontology too.

RDF limitation - cannot include computations.

Conversion factors can introduce inaccuracies. Issue is whether or not it would be useful to actually record this.

Rhys: we can mention it. Probably won't be practical issues in implementations.

Categorisation of devices could be useful.

RH: Could look at "primitive" device concept.

Rhys: Yes, like "printer-scanner-fax multifunction device" is a composite.

RH: So we could identify the primitives and say devices are composites.

Rhys: yes.

Dave: describing capabilities, in terms of what is available to the application. E.g. access to a camera.

Rhys: Like if it has access to pictbridge.

Dave: could we phase into PUCC to see how this would be considered there.

Giles: OK

... Output of this will be an OWL file?

Rhys: that and a W3C spec.

... Delivery Context Ontology.

Rhys: need to cover DDC in ontology.

(Shows generated XHTML page showing ontology summary.)

Rhys: note also removal of "has".

... Guidance from SemWeb says verb should be involved, but new thinking disagrees.

... Can cause confusion. Looks like Java methods isXYZ() and thus gives the wrong impression.

... So we no longer prepend a "has". These are not methods, these are relationships.

(Moves on to focus on Device class.)

Dave: getting ontology out to support content adaptation is fine.

... requirements for DCCI second.

... That could reference this work.

Rhys: DCCI stuff is just more properties. Need to discuss interface requirements.

Dave: why is battery in there?

Rhys: Wanted to show how to model dynamic properties.

Dave: so you envision sending data to the server?

Rhys: Yes.

Sailesh: Can we use DCCI on the client, serialize data, send to server, then use that data there?

Rhys: Yes, though also OMA DCAP can help here.

Sailesh: example, speech recognition could be server side. When in high-bandwidth context, I have a better speech recognition (server side) because it can deal with better resolution data.

(Dave explains - the ontology Rhys presented is aimed at properties that are accessible by the server for adaptation purposes, but in future we will also be interested in supporting discovery/binding which will require new kinds of information)

Rhys: event representing change in bandwidth can be sent to flag shift in capability.

... Can register for events with the DPE server.

Rotan: So, Delivery Context was just for context affecting delivery of *content* but now we are considering effects of context on application itself.

... Should we expand defn of DC or do we introduce a new term.

Sailesh: propose we change name of DCCI.

(lol)

Kevin: how do we determine if a vocabulary is consistent with the ontology?

... e.g., how would WURFL people say that their implementation conforms?

Rhys: more likely that WURFL would say they comply with an API.

Dave: need to discuss what the exit criteria of this would be.

Rotan: could be sufficient for vocab groups to reference ontology in their specs.

Dave: normative references from spec groups, and commercial support.

Rhys: three industry groups normatively referencing. DDR Core Vocab, OMA-BT-DCAP and DISelect (UWA).

PUCC

Giles: PUCC local task force to look at service metadata.

...Potential start in September 07.

...How to work with W3C (formats, admin, communication)

...Precedent with W3C character set investigations (Asian character sets), was a good model

Dave: May be problem with different patent policies. If a spec wants to become part of a W3C rec, we need to follow our patent policy.

Giles: PUCC amending patent policy to facilitate interworking with other SDOs

...(Giles takes action to follow this up)

Rhys: NB only those parts submitted to W3C need comply.

Giles: And there is an overlap of membership, also the PUCC universities are usually patent-free.

Giles: Looking at establishing high level goals (of the interworking)

...e.g. compatability between specifications.

...Need semantic richness, so that interfaces can be created dynamically.

Dave: Analog to academic work in this area.

Giles: This was captured in the workshop. Also state transitions are of interest to PUCC

Dave: eg. agent control, responding to different events/commands

Giles: Looking at events relevant to a given state, so a dynamic interface would be relevant to state

Dave: Levels of application support (from static to rich dynamic applications, and sophisticated state transition behaviours required for agents)

...Access control and security also relevant to PUCC

Accessibility?

Giles: No, that's a different level.

Dave: The richness of description of services and interfaces, just naming or more?

Giles: Also note Rhys' ontology work.

Dave: When will it be published?

Rhys: It's public now, and potentially publishable.

Rotan: The DD process has been launched this week, OMA liason statement sent.

...This will bring in suggested properties from OMA [Rotan describes the process]

http://www.w3.org/2002/09/wbs/1/ddwg-core-vocab-public-form

Rhys: So first working draft before next F2F (November), posssibly Q3 depending on response.

Dave: Giles, what do you expect from W3C?

Giles: As long as there is some communication/feedback loop that the work is going through normal W3C process (guidance, allowing comments etc.) then that is fine.

One more time... http://www.w3.org/2002/09/wbs/1/ddwg-core-vocab-public-form

Giles: Note about nested primitive devices, will report that back to PUCC as regards possibility of using it.

Dave: Do we see W3C developing ontologies for printers etc. based on PUCC work?

Rotan: In DDWG we don't build the vocab, it's a public process. The rec is a gathering of sources.

Dave: We would need to publish a roadmap to assist recruitment of input.

Rotan: (Notes there could be a problem with the current form submission mechanism, as per DDWG experience)

Dave: As per roadmap, we could wait until PUCC publishes specs.

Giles: September to end of year, possibly.

Dave: Want to say 'this is what we are doing, here's an example' to garner company interest.

...To what extent is PUCC looking at security?

Giles: Noted as question

Giles: Marie-Claire said she would be in Japan in July, good opportunity to meet with Kaz.

Dave: We can review PUCC specs when they are published.

ScribeNick: Giles

Rotan: If there is anything we can achieve in the short term - then we should do it
... Otherwise people will think we ignored the output of the Workshop

Dave: How about the Gaps and Feedback to the UI Groups

Rhys: ... Jose interested in the User Interface stuff
... the page flows stuff looks a good target for getting something quickly
... a lot of people have implementations

Rotan: Late binding and loose coupling related to the discussion about layers this morning

Dave: Had the question "What do we mean by late binding?"

Rotan: Analogy with hyphenation - guidance within the document needed for the late binding

Dave: "role" is a good example and easily doable

Rhys: How do you want to articulate back that we've taken on board these ideas

Dave: Put something in the Workshop report

Rotan: The Workshop Report should be what happened in the Workshop - not what is coming after it

Rhys: Can include a section on intended actions at the end to say what will do
... based on the output from the Workshop

Dave: Gap analysis -see whats missing from DIAL, XForm or other layout specs

Rotan: Need to identify the technologies that support adaptation (existing technologies)

Rhys: OK but we need to say that we confirmed our suspicions there

Dave: There are bunch of ideas CSS, XSLT FO etc. we need to consider

Kevin: Could have a primer 2.0

Rotan: Output will be a document to say these are the technologies that are out there
... that we think are good and then say why we think they're good

Rhys: Everybody agreed that a declarative semantic approach was good

Dave: Need to find ways of addressing composability

Rhys: Do we need a requirements document?

Dave: W3C usually gathers use cases and then make a requirements document based on that

Rotan: Use cases could be gathered in a Wiki

Kangchan: What about the eco-system?

Dave: Yes that is very important

Rhys: Might be the same as the DD eco-system - could cut and paste

Dave: Often requirements documents gather dust

Rotan: If you use a Wiki - its faster since you don't have to worry about it being a formal document

Dave: E-mails on the public mailing list are often ignored - nobody takes responsibility

Rotan: In DD we also blog - you can use it to set up debates

Dave: W3C don't want to support 2 different Wikis

... don't want to use Moin Moin

... Stephane can you help persuade them to let us use Media Wiki?

Stephane: Take an action to discuss this

RH http://meta.wikimedia.org/wiki/Semantic_MediaWiki

... What is the home page for it?

Rhys: Dan Conolly also supports Semantic Media Wiki

... lets you have RDF

Stephane: Could look at trying it out at Sofia.

... I'll see what I can do

TimBL mentions Wikipedia recently - http://www.w3.org/2007/Talks/0222-3gsm-tbl/text.html

He says ... "Web 2.0 community Web sites, eBay, and Flickr are possible because the Web standards, in turn, were widely implemented in an interoperable way, before those innovations. The same for the wikis, like Wikipedia, and blogs, and so on."

Giles_ ACTION: Stephane to report back on possibility of using Semantic Media Wiki by end of June [recorded in http://www.w3.org/2007/06/dublin-uwa-f2f-mins.html#action06]

trackbot-ng Created ACTION-7 - Report back on possibility of using Semantic Media Wiki by end of June [on Stéphane Boyera - due 2007-06-14].

Rhys: Need section of report that deals with overall outcome

... and fit in these items (list of output from workshop)

Rhys: Send to the public list a list of conclusions from the workshop

Giles_ ACTION: Dave to prepare an e-mail as an interim summary of the workshop [recorded in http://www.w3.org/2007/06/dublin-uwa-f2f-mins.html#action07]

trackbot-ng Created ACTION-8 - Prepare an e-mail as an interim summary of the workshop [on Dave Raggett - due 2007-06-14].

Rhys: Conclusions:

General agreement that declarative approaches are a good idea

Importance of semantics

Importance of high level abstractions

Set of technologies listed in the UWA charter is more or less appropriate

State machines can be used to represent page flows

Application adaptation seems to be an appropriate approach

Dave: Being able to reduce the amount of server side scripting (by capturing the end-to-end semantics)

Composability is important for mash-ups

Rhys: Didn't do much on ontology

Dave: Got a clearer idea of what is need for rich service metadata

Agree the need for gap analysis

Rhys: Importance of not re-inventing the wheel

Rhys by using existing standards and technologies and identifying the gaps

Dave: Full report on the workshop due end of July (?)

Dave: Tomorrow - where next for DIAL?

(end of minuting for day 1)

slides for agenda topics: http://www.w3.org/2007/Talks/0607-dsr-uwa/

agenda topics again: http://www.w3.org/2007/Talks/0607-dsr-uwa/

F2F Day 2

ScribeNick: Rotan

Krcsmith http://www.w3.org/2001/di/Group/di-profile/DIALspec.html

Kevin: There are issues in the doc to resolve, which can then be published.

Rhys: but that's just editorial. Might not need republishing.
... Might give people impression there's something really new to look for.

Kevin: OK.
... So let's just look at existing issues.
... In this session, look at future of DIAL.

Dave: Expectations of when we intend to public new version.

Kevin: OK, look at issues and matters re implementations.
... Re the primer...
... MIME Type for DIAL. Look at CDF.
... What's the proper way to register.
... application/xml+dial ?

Rotan: might be an issue for future evolution of DIAL. What would be the future MIME types?

Rhys: Version finding in TAG. 3 volumes. 200 pages. Not sure if it delivers a crisp answer :)

Kevin: How do they do in in HTML?

Rhys: In v5 might not reveal anything at all.

Dave: Under debate/

Rotan: so when do we have to resolve this issue?

RH ScribeNick: RH

ScribeNick: Rotan

Rhys: May have to leave this issue until we see what happens with XHTML2.

Kevin: we should add these issues to Trackbot

Rhys: Need to retest trackbot.

Kevin: I can take action to add remaining actions to trackbot.
... Issue with schemas.

Rhys: Could be challenges. Need to have extension points.

Kevin: Action from us to XHTML2 that we need extension point?

Rhys: Don't think we need to put anything in for this version of DIAL.

Kevin: substitution may only work for certain data types.

Rhys: need to structure in right way to support substitution.

Rotan: seems like we're not completely sure. Need to check.

Rhys: XForms is using schema.

Kevin: do we have to produce more than one schema?

Rhys: probably
... probably have some things to offer here.

Dave: take action to coordinate with SP.

scribe ACTION: Kevin to coordinate with Steven Pemberton on XHTML2, and schema related issues in two weeks. [recorded in http://www.w3.org/2007/06/dublin-uwa-f2f-mins.html#action08]

trackbot-ng Sorry, couldn't find user - Kevin

scribe ACTION: Krcsmith to coordinate with Steven Pemberton on XHTML2, and schema related issues in two weeks. [recorded in http://www.w3.org/2007/06/dublin-uwa-f2f-mins.html#action09]

trackbot-ng Sorry, couldn't find user - Krcsmith

scribe ACTION: Smith to coordinate with Steven Pemberton on XHTML2, and schema related issues in two weeks. [recorded in http://www.w3.org/2007/06/dublin-uwa-f2f-mins.html#action10]

trackbot-ng Sorry, couldn't find user - Smith

(Group struggles with action recording tools.)

Dave: will see if action can be recorded via alternative interface.

scribe ACTION: Kevin Smith to coordinate with Steven Pemberton on XHTML2, and schema related issues in two weeks. [recorded in http://www.w3.org/2007/06/dublin-uwa-f2f-mins.html#action11]

trackbot-ng Sorry, couldn't find user - Kevin

stef trackbot-ng, help ?

trackbot, status

trackbot-ng, status

Kevin: issue with validation

Rhys: DISelect designed to be validated in its source form.

Kevin: should we validate all possible results.

Rotan: That's "Theorem Proving"!
... Suggest we avoid Complexity problems. Test generated material at run-time perhaps?

stef tracbot, status ?

Kevin: DIAL doc must be well-formed XML.... (reads from conformance part of spec).

stef trackbot, status ?

stef trackbot-ng, status ?

trackbot-ng, status

stef trackbot-ng, help ?

stef trackbot-ng, add kevin

stef trackbot-ng, add kevin uwa

Rotan: suggest we be consistent in conformance spec about using "input doc" and "generated doc".

trackbot-ng, init

trackbot-ng Reloading Tracker config

trackbot-ng, status

stef http://www.w3.org/2005/06/tracker/irc

stef @@ describe how usernames work

trackbot-ng Tracking ISSUEs and ACTIONs from http://www.w3.org/2007/uwa/trackbot/

scribe ACTION: Kevin to coordinate with Steven Pemberton on XHTML2, and schema related issues in two weeks. [recorded in http://www.w3.org/2007/06/dublin-uwa-f2f-mins.html#action05]

trackbot-ng Created ACTION-9 - Coordinate with Steven Pemberton on XHTML2, and schema related issues in two weeks. [on Kevin Smith - due 2007-06-15].

Krcsmith ACTION: Kevin ref- DIAL spec issue-validation, add text 'A DIAL document must be valid before and after processing' to the spec. [recorded in http://www.w3.org/2007/06/dublin-uwa-f2f-mins.html#action12]

trackbot-ng Created ACTION-10 - Ref- DIAL spec issue-validation, add text \'A DIAL document must be valid before and after processing\' to the spec. [on Kevin Smith - due 2007-06-15].

Krcsmith ACTION: Kevin , ref- DIAL spec issue-validation, use consistent text throughout for source infoset and result infoset (as per DISelect) [recorded in http://www.w3.org/2007/06/dublin-uwa-f2f-mins.html#action13]

trackbot-ng Created ACTION-11 - , ref- DIAL spec issue-validation, use consistent text throughout for source infoset and result infoset (as per DISelect) [on Kevin Smith - due 2007-06-15].

Kevin: in XHTML2 they allow entities to be used within an XHTML2 documents.
... Schema doesn't allow them.

Dave: We could use a DTD, or just say they're not allowed.
... What do customers want?

Rotan: they want DTDs.

Rhys: They need DTDs to load namespaces.

Kangchan: also, power of client may be low, so validation on client could be expensive.

Rhys: in future may process DIAL on client, but no intent right now.

Rotan: do we specify where validation failure events go?

Rhys: no, we only specify the nature of the event, not the post-DIAL processing.

Kevin: in text, it says DIAL doc must be valid before and after processing.

Rhys: there's the input doc (must be valid), there's the resulting doc (can't say anything about it's validity), and may already have stated in DISelect that the result of applying DISelect is a valid doc.

(Discussing validity re material that DISelect might include from external sources.)

Dave: After processing DISelect part of DIAL you get XHTML2+XForms (+ whatever extensions you may have added?)

(Hi-speed three-way discussion on DIAL/DISelect processing ensues. Summary will follow...)

DIAL Schema must be capable of extension.

Kevin: will explain in text that DIAL doc must be valid against schema before processing.

Rhys: could say the source infoset.
... encourage you to look at Section 3 (processing model) of DISelect text. Possibly even lift some of that text.

scribe ACTION: Kevin to mention the intent to ensure that the DIAL schema is extensible in two weeks. [recorded in http://www.w3.org/2007/06/dublin-uwa-f2f-mins.html#action14]

trackbot-ng Created ACTION-12 - Mention the intent to ensure that the DIAL schema is extensible in two weeks. [on Kevin Smith - due 2007-06-15].

(Kevin will close off actions from tracker.)

Kevin: issue - XHTML2 Conformance
... Thinks in XHTML2 that are not in DIAL.
... If XHTML2 used image, ruby etc...

Rhys: Take position that DIAL is a superset of XHTML 2.

Rotan: so you suggest we just say it's a superset rather than the copy/paste we did before?

Rhys: HTML2 is a moving target.
... Suggest we state intention to have DIAL as a superset of XHTML 2.

Stephane: Can have a DIAL doc that doesn't support ruby etc.

Rhys: But would certainly like it to support these.

Rhys Proposed resolution: We should state that DIAL is a proper superset of XHTML 2

Krcsmith +1

RH +1

Rhys Hearing no objections...

Rhys RESOLUTION: DIAL specification should state that DIAL is a true superset of XHML 2

scribe ACTION: Kevin to include text stating DIAL specification should state that DIAL is a true superset of XHML 2, as per resolution, in two weeks. [recorded in http://www.w3.org/2007/06/dublin-uwa-f2f-mins.html#action15]

trackbot-ng Created ACTION-13 - Include text stating DIAL specification should state that DIAL is a true superset of XHML 2, as per resolution, in two weeks. [on Kevin Smith - due 2007-06-15].

(issue-redefinition can be removed, by Kevin)

Rhys: CSS validation is out of scope, so issue-CSSSupport is not something we can address.

Dave: perhaps policies could help here.

Krcsmith http://www.w3.org/2001/di/Group/di-profile/DIALspec.html#applying-styles-CSS-selection

Dave: may not be best practice to link to stylesheet from within DIAL doc, with or withour DISelect.

Rhys: inclusion of stylesheet, later, can take care of adaptation of stylesheet.

Rotan: assuming one communicates the delivery context.

(issue-mediaqueries?)

Kevin: media queries issue is addressed by above

RH Issue is that any attribute can be the result of an AVT (computation), as permitted in DISelect.

RH Does this affect validation?

scribe ACTION: Kevin to clarify wording to address AVT issue, in two weeks. [recorded in http://www.w3.org/2007/06/dublin-uwa-f2f-mins.html#action16]

trackbot-ng Created ACTION-14 - Clarify wording to address AVT issue, in two weeks. [on Kevin Smith - due 2007-06-15].

Rotan: note that DISelect Basic does not support AVTs (by design) as per http://www.w3.org/TR/cselection/#sec-avt

Rhys: in case where AVTs are supported, you are dealing with GIGO.
... So cannot state much about output.
... DIAL can require that output is valid DIAL.
... even though the result of processing would not contain DISelect markup.

Rotan: we previously avoided the issue of delivering the whole original DIAL doc to the device (to allow reprocessing) because of the complex use cases that would enuse.

Kevin: So AVT 2 issue is also resolved.

Issue DISelect Full/Basic?

Rhys: Full.

RESOLUTION: DIAL will reference DISelect Full.

Issue XHTML 2 legacy

Dave: If XHTML 2 contains legacy material, and we think it a bad idea to use that material, then it's merely an issue of Best Practices.
... We can document that separately.

Issue selidname

(Rhys explains background to selidname)

Krcsmith http://www.w3.org/2001/di/Group/di-profile/DIALspec.html#content-inclusion

Kevin: we have comments around inclusion.
... actual inclusion mechanism for XHTML 2 is still under discussion.

Rhys: src attribute is an inclusion mechanism.
... issue was about XInclude etc.
... we've said we'd use the XHTML 2 inclusion mechanism.
... issue about including fragments, which don't have roots.
... suggestion to remove second paragraph of 2.10 and the ed note.

scribe ACTION: Kevin to remove para 2 and ed note from section 2.10 "Content Inclusion" of DIAL in two weeks. [recorded in http://www.w3.org/2007/06/dublin-uwa-f2f-mins.html#action17]

trackbot-ng Created ACTION-15 - Remove para 2 and ed note from section 2.10 \"Content Inclusion\" of DIAL in two weeks. [on Kevin Smith - due 2007-06-15].

Krcsmith http://www.w3.org/2001/di/Group/di-profile/DIALspec.html#sec-applying-layouts

Leave 3.2

Kevin: should identify in DIAL doc what features of XHTML (etc.) that we are explicitly using in support of adaptation (and why).

Dave: OK, let's look at new material.
... XBL and SVG and WICD.
... Trying to take advantage of best state of the art of compound docs./

Rhys: was disagreement about modularisation.
... CDF and XHTML.
... Need to track developments in this area.

(Compound Docs by Ref framework could be considered.)

(Discussion on how CDF works, eg WICD incorporating SVG.)

Rhys: next phase probably too early. Need to get DIAL 1 and implementations.

Dave: proposal to include XBL (later).
... support contextually dependant binding of controls.

Rhys: XBL useful for binding XForms widgets, esp those in SVG.
... XBL about helping binding to rich controls.
... surfaces into authoring space, instead of being hidden.

Dave: current approaches messy, and difficult to maintain.
... CSS3 UI module discusses relationship between input fields and how they appear, for example.

Kevin: would like to see what we want, the drivers, requirements, not how they could be achieved.

Dave: could use wiki to start discussions on roadmap for DIALv2

Kevin: and that gives traceability for our decisions.
... Input from José and Morfeo useful here.

Dave: need discussion around what it means to support rich interaction.

Rhys: workshop report feeds into this.

Dave: also relationship to SCXML.
... application flow.

Rhys: pageflow was the term we used, but it's more than that.
... potentially within the page as well.
... essentially application flow.

Dave: wiki will help to explain to others why these are important.

Kangchan: interested in using delivery context to reduce the interaction between server and client.
... using flow on the client.
... intent-based services, page flow based on intent.

Dave: if you have idea of state-machine driven by events, then identify intents with those events.

Rhys: may be abstract events. Like navigation.
... and concrete events (click).

Kangchan: no meaning to associate with event if you just go directly to the page.

Dave: can used state-machine to control flow.

Rhys: page generated in response to request.

Rotan: seeing trend away from pages. URL of entry page, and then Web2.0 DOM morphing events modify fragments. No more "pages", just morphing presentations. Where are the URLs in here?

(Discussion of relationship between pages, URLs, state-machines, representation of state, and reconstruction of state...)

Sailesh: decomposition of page is an issue.

Dave: we're too focussed on concept of a page. Need to move to state.
... New issue, where DIAL fits in the family of XML langauges.

Rhys Scribe: Rhys

New Work Items

After lunch, Stephane reported on Semantic MediaWiki. Was discussed in the sys team last week. Seeing some people requesting MediaWiki.

Will look at migration. Will take time. Global sys team says its ok for us to go ahead and install an instance. Won't be supported by sys team.

Group thanks Stephane

scribe ACTION: Stephane to investigate SemanticMediaWiki by end of June [recorded in http://www.w3.org/2007/06/dublin-uwa-f2f-mins.html#action18]

trackbot-ng Sorry, couldn't find user - Stephane

scribe ACTION: Stef to investigate SemanticMediaWiki by end of June [recorded in http://www.w3.org/2007/06/dublin-uwa-f2f-mins.html#action19]

trackbot-ng Sorry, couldn't find user - Stef

scribe ACTION: stef to investigate SemanticMediaWiki by end of June [recorded in http://www.w3.org/2007/06/dublin-uwa-f2f-mins.html#action20]

trackbot-ng Sorry, couldn't find user - stef

scribe ACTION: sboyera to investigate SemanticMediaWiki by end of June [recorded in http://www.w3.org/2007/06/dublin-uwa-f2f-mins.html#action21]

trackbot-ng Created ACTION-16 - Investigate SemanticMediaWiki by end of June [on Stéphane Boyera - due 2007-06-15].

Dave shows the chart of deliverables and timescales. Indicates what we said we would do

DR: We have new proposals from CSS for layout. Also Grid layout from Microsoft.
... Need requirements, editor,

http://www.w3.org/2007/Talks/0607-dsr-uwa/#(7)

DR: that is the link to the latest CSS 3 layout module
... Need a leader for the work leading to the layout
... Would like one person to lead the work on this.

RH, RL and JCF have all indicated interest in working on this

DR: Need to look at existing materials, at what is going on in CSS, and at the requirements for richer layout and support for composition
... Seems like the kind of thing that could proceed on the Wiki

RH: Important to distinguish our individual opinions from group opinions

DR: Layout is distinct from just applying CSS. Often, implementations have used a combination of scripting, markup and styling

SS: Does the layout policy include the interaction policy?

DR: Trying to keep the various kinds of ontology separate

SS: Policies can apply to things at all levels. The layout policy seems to refer to just a single page. Need policies to refer to the page flow too

KS: Layout constraints are policies

SS: Need to address behaviour too.

DR: Need to understand the boundaries of this work too.

RL: Need to have explicit layout representations too. Continuum from explicit authored layout to layout constraints

DR: Like TeX springs for constraints
... Could define new XML layout rules. Need to be able to explain why CSS is insufficient.
... Have to be able to support implementations, of course

RESOLUTION: Form a task force on policy-based layout. Initial members, RH, RL and possibly JCF.

Keith hey!

DR: Use Wiki to collect reviews of existing materials, use cases and requirements

Hi Keith, we are working on policy based layout right now and will switch to DCCI in a while

DR: Question about the amount of time to use on calls

Keith Hello all !

Kangchan KS, the minutes is forbidden.

stef done

Krcsmith (Rhys and Dave discuss layout policies vs behaviour policies)

Summary is that layout needs to be concerned with dynamics and time explicitly.

DR: Points out that CSS approach causes issues for editors

RH: Lots of libraries give useful function and perhaps they could be categorized

DR: Need to keep separation of the concrete and abstract UI

RH: Suggests contact with ecmascript committee

DR: Think that they have a different focus

RH: Could send a note to their list.
... Libraries drive layout using script behaviour.

DR: Should try and capture the kinds of layout that people want to achieve. Both through markup and through script to change layout dynamically

RH: Could be that because this is a new use for Ajax and script, that we could talk to the people developing the libraries to see what is happening

Behaviour in the discussion on layout includes movement of things within containers as a change of layout

DR: When do we expect to need time on the weekly teleconference

scribe ACTION: Dave to invite JCF to the layout taskforce [recorded in http://www.w3.org/2007/06/dublin-uwa-f2f-mins.html#action16a]

trackbot-ng Created ACTION-17 - Invite JCF to the layout taskforce [on Dave Raggett - due 2007-06-15].

scribe ACTION: Rotan to report back on the kinds of scripted components for layout that are being used in the Ajax world [recorded in http://www.w3.org/2007/06/dublin-uwa-f2f-mins.html#action22]

trackbot-ng Created ACTION-18 - Report back on the kinds of scripted components for layout that are being used in the Ajax world [on Rotan Hanrahan - due 2007-06-15].

scribe ACTION: Rhys to review the latest draft of the CSS layout module [recorded in http://www.w3.org/2007/06/dublin-uwa-f2f-mins.html#action23]

trackbot-ng Created ACTION-19 - Review the latest draft of the CSS layout module [on Rhys Lewis - due 2007-06-15].

Krcsmith Scribe: Kevin

Kangchan Thanks Kevin

Krcsmith Dave: (shows list of requirements for document advancement)

Krcsmith ...don't *have* to do tests before CR (but advisable)

DISelect and XAF transistion to CR

Krcsmith ACTION: stephane to report back on the steps needed for a transition to CR for DIselect and XAF [recorded in http://www.w3.org/2007/06/dublin-uwa-f2f-mins.html#action24]

trackbot-ng Sorry, couldn't find user - stephane

Krcsmith ACTION: sboyera to report back on the steps needed for a transition to CR for DISelect and XAF [recorded in http://www.w3.org/2007/06/dublin-uwa-f2f-mins.html#action25]

trackbot-ng Created ACTION-20 - Report back on the steps needed for a transition to CR for DISelect and XAF [on Stéphane Boyera - due 2007-06-15].

DISelect documents and issues lists are at http://www.w3.org/2001/di/Group/di-selection/

DCCI (Keith Waters)

Krcsmith Keith: Delivery Context Remote Interfaces

Krcsmith ...idea is to put down ideas about delivery context client interfaces

[Showing title page of Keith's slides]

[Next slide]

Krcsmith ...how they can work in 'lighter' devices

Krcsmith ...DCCI, property access and manipulation

Krcsmith ...Spec provides *declarative* access

Krcsmith ...What if you mix it with dynamic (changing over time) properties, e.g. location services?

Krcsmith ...DCCI does not provide implementations of properties (it's a framwework)

Krcsmith ...Status; we are moving toward full recommendation.

Krcsmith ...In UWA charter DCCI is proposed for Nov 07.

Krcsmith ...Are implementation reports to be complete at that time?

Krcsmith Dave: we need those at exit CR

Kangchan Implementation Report : http://www.w3.org/2001/di/dci/ir/

Krcsmith Keith: We have published a set of test suites (atomic tests) through IDL

Krcsmith ...more comprehensive tests (combinations involving DOM traversal) being made.

Krcsmith ...We now have an implementation that passes most of the tests.

Krcsmith ...We need further usage examples. Specifically to show the usefulness.

Krcsmith ...Important for test harnesses to allow cross-validation between two implementations.

Krcsmith ...[shows location hierarchy slide]

Krcsmith ...Have Firefox implementation of this example.

Krcsmith Sailesh: updateDistance and position nodes in example tree. If you provide a value to say when an event is fired by that node, does that go against the spec?

Krcsmith ...Because by providing a value you are controlling the sibling position node.

Krcsmith Keith: No, this is an updateDistance unconnected to the position node.

Krcsmith Sailesh: So what is purpose of setting updateDistance?

Krcsmith Keith: see subsequent example slides.

Krcsmith Keith: We are calculating distance of contacts to the yellow marker. We centre on contact 1, and can retrieve info from property hierarchy about calculated distance.

Krcsmith Sailesh: So is ontology tied to application

Krcsmith Keith R: Yes, that's correct. Don't know if that violates spec though.

Krcsmith Sailesh: Just wondering, not a problem for now.

Krcsmith Keith: Loosely coupled but does not have to affect that property layout

Krcsmith ...(DCCI implications). Looking at relationship between UWA and DCCI.

Krcsmith ...DCCI fine for 'fat browsers'. But UWA considers 'lighter' devices too.

Krcsmith ...This makes DCCI support hard.

Krcsmith ...However we believe the DCCI framework for manipulation/adaptation is valuable, so can we do it without a browser implementation?

Krcsmith ...(slide 8, describes DCCI processing)

Krcsmith ...(slide 9) Delivery Context Remote Interface. Considers the lightweight devices.

Krcsmith ...Looks at having a server interface which provides tailored content. So transactions between server and remote DCRI

Krcsmith ...How to get client updates back to remote interface.

Krcsmith ...Should not be a showstopper.

Krcsmith ...(Conclusion) Potential for DCCI for mobile handsets. Another context for 'lighter' clients hence the need for DCRI.

Rhys mentions BT DCAP.

Krcsmith Rhys: This is analog to DCAP, the OMA browser WG. OMA is a mobile organisation, so would not cover full scope of UWA.

Krcsmith ...Not sure whether DPE/DCAP would replace DCCI or DCRI.

Krcsmith Keith: Agree to take a look at the OMA work. Interesting overlap.

Krcsmith Sailesh: Nokia drives heavyweight (capable) devices with powerful browsers.

Krcsmith ...Most of the context data (location, contacts) are local to the handset. Device itself can generate this info. There can be other data sent from the server.

Krcsmith ...Client apps will access most data through the client (even if the client has to contact the server) rather than a direct call to a remote interface.

Krcsmith ...i.e. there will always be a local model for access.

Krcsmith Keith R: This is not intended to replace the local model but to help light devices without e.g. Javascript.

Krcsmith Sailesh: But is that lightweigtt, to have a remote interface? e.g. Bandwidth problems?

Krcsmith ...If you have implemented DCCI from scratch it is not a heavyweight problem.

Krcsmith Dave: Practically we want to move out of WD. Rhys has applied edits. Rhys will provide URI of edits, please review pre-publication.

Latest DCCI spec version is at http://www.w3.org/2007/uwa/editors-drafts/DeliveryContextClientInterface/2007-06-30/DCCI.html

Krcsmith ...XTech presentation slides, one topic was describing different kinds of devices. e.g. Sensors etc. which can generate of respond to events.

Krcsmith ...Then more powerful systems that request and respond to events, also state driven (SCXML).

Krcsmith ...Support for richer descriptions for service interfaces.

Krcsmith ...As per PUCC. They will let us know their work on service interfaces.

see http://www.w3.org/2007/Talks/0607-dsr-uwa/#(10)

Krcsmith ...Then binding work on device coordination (see link above)

Krcsmith ...So, moving DCCI spec forward, and new items, such as binding mechanisms in DCCI, geolocation

Krcsmith ...as a potential property to standardise.

Krcsmith ...As per KISS principle, have a simple interface to access geolocation.

Krcsmith ...Independent of location technology.

Krcsmith Sailesh: e.g. we can agree on the metadata and produce a concrete example, and can work with Ryan Sarver.

Krcsmith Dave: Access control mechanism is less simple. Need to avoid binding to a particular technology.

Krcsmith Keith: Yes, important.

Krcsmith Dave: Also provides a good way to get into security. In Banff talks around considering security throughout. How can that work with a simple interface.

Krcsmith ...Starting a study covering use case requirements, business drivers.

Krcsmith Keith: Sprint just announced interest in location based shopping services, many ways to monetize this.

Krcsmith Dave: The different technlogies have different models, e.g. phone with all GPS inside, or A-GPS, or cellID triangulation.

Krcsmith ...proposal for this as a work item.

Krcsmith ...Other example, access to camera in device. Specific properties that will provide most value in short term.

Krcsmith Keith: Geolocation a hot topic, should be lots of examples for monetization.

Krcsmith ...Are other groups looking at this?

Krcsmith Dave: OMA work?

Krcsmith Rhys: Not sure.

Krcsmith Keith: Good idea to invite Ryan Sarver.

Krcsmith Dave: MikeTM Smith and Ryan presented at XTech Paris. At level of Web applications, not so much at operator level.

http://www.sarver.org/

Krcsmith ...But would like more people. Also valuable for people from other companies, technologies.

Krcsmith Dave: We also discussed about public work in a Wiki.

Krcsmith ...So we should be able to publish WD of DCCI soon and move to CR.

Krcsmith ...In parallel work on complimentery items (location as property and binding).

Krcsmith Keith: +1. Binding is important, location has high visibility.

Krcsmith Rhys: I assert we are ready to publish DCCI as new WD

Krcsmith Keith: Agree that all action items taken care of, just will mail regarding wording of DOM3 implementation. (ref: previous URI)

Krcsmith Dave: Propose work item for geolocation as a property.

Krcsmith Sailesh: DCCI item, so I can help. How to proceed?

Krcsmith Dave: new work item separate to the DCCI.

Krcsmith ...service binding is separate.

Krcsmith PROPOSED RESOLUTION: Create work item on location as a case study for DCCI properties.

Krcsmith Kangchan: Nothing analog in geospatial incubator group

Krcsmith RESOLUTION: Create work item on location as a case study for DCCI properties.

Krcsmith ACTION: Dave to proceed with inviting expert Ryan Sarver. [recorded in http://www.w3.org/2007/06/dublin-uwa-f2f-mins.html#action26]

trackbot-ng Created ACTION-21 - Proceed with inviting expert Ryan Sarver. [on Dave Raggett - due 2007-06-15].

Krcsmith Dave: Binding mechanism. Suggestion for a general framework.

Krcsmith Sailesh: Agree that binding mechanisms are valuable.

Krcsmith Kangchan: What is definition of a device (in device coordination context)

Krcsmith Dave: Device coordination is an overlay mechanism for communicating between devices. There does not to be a browser in this model.

Krcsmith ...e.g. a state machine coordinating other devices and provisioning a remote interface.

Krcsmith ...Idea to extend Web to other (currently non-Web) devices.

Krcsmith ...Describing UIs, hiding network specifics.

Krcsmith Sailseh: Via a unified model.

Krcsmith Kangchan: We had a testbed based on Web services; a coffeemaker created an event to a Web service, then an STB acted as the Web service server, coffeemaker could send message to Web service that it had run out of water.

Krcsmith Dave: So an application could bind to the coffeemaker

Krcsmith Dave: Device coordination is about describing the resource to bind to, how to bind (as a local DOM), events to raise, exceptions; and how a service exposes these.

Krcsmith ...A binding mechanism could use an implicit mechanism (UpnP) or state the Web server playing the role.

Krcsmith Kangchan: How to bind a UWA and the proprietary protocol

Krcsmith Dave: With a layer on top that hides the proprietary protocol.

Krcsmith ...So a framework.

Krcsmith ...Aligned to existing mechanisms and existing markup languages (e.g. DCCI)

Krcsmith ...Need work on security, sessions etc. so examples will help.

Krcsmith ...Wiki will help gather information

Krcsmith PROPOSED RESOLUTION: Work item on device coordination binding

Krcsmith Rotan: any target dates?

Krcsmith Dave: Should have more detailed report for next F2F.

Krcsmith Sailesh: Start by putting ideas on the Wiki, will kick start the work item.

Krcsmith RESOLUTION: Work item on device coordination binding in time for November TP

Krcsmith Dave: REX update.

Krcsmith ...proposal for v2 of REX to cover broader range of events (shows REX v2 slide)

Krcsmith ...Keyboard events not well defined now, driver for DOM3 serialization.

Krcsmith Sailesh: Work on intent-based events

Krcsmith Dave: Separate issue, with Al Gilman

Krcsmith ...We have an item to cover it, but it's cross group (backplane)

Krcsmith ...Dave will discuss with Al and Stephen Pemberton. We don't need to cover it explicitly.

Krcsmith ...This group may get involved in PAG for REX

Krcsmith Sailesh: SCXML?

Krcsmith Dave: i.e. how to drive REX from SCXML and hence control a visual DOM tree.

Krcsmith ...Robin Berjon proposed two documents (ref: REX slide). Should not be a huge piece of work, quite straightforward. E.g. set of IDL to XML mapping rules.

Krcsmith Dave: CCPP 2

Krcsmith stef_dub: Have been discussing with Dan Connelly. Resolving mailout issues to chairs for review period.

Krcsmith ...This issue should not impact us up to CR, but maybe afterwards.

Krcsmith ...Not critical timewise.

Krcsmith Rhys: We provided reviews as part of DIWG.

Krcsmith stef_dub: maybe CR beginning of July.

Krcsmith Dave: Need to get LC comments.

Krcsmith Rhys: Likely no more comments due to expert involvement in editing process.

stef_dub [17:04] IanSysTeam in the msg, indicate that the doc started LC on date but htat

stef_dub [17:05] IanSysTeam no announcement went to chairs, so extending the LC review period.

stef_dub [17:05] IanSysTeam There's an issue of the document saying "15 June" but the period being extended, but we'll have to live with that unless you want to republish.

stef_dub [17:06] stef_dub ok

stef_dub [17:06] stef_dub no republish thanks

stef_dub [17:06] IanSysTeam ;)

stef_dub about the patent policy

stef_dub [17:02] IanSysTeam regarding the exclusion period, I have not sent out the CFE since there is an open pat pol issue.

stef_dub [17:02] IanSysTeam The psig has not yet resolved whether licensing commitments from _di wg_ are still en vigeur or not

stef_dub [17:03] stef_dub ok

stef_dub [17:03] stef_dub does this prevent us to move to CR ?

stef_dub [17:03] IanSysTeam Probably not

Krcsmith Dave: patent policy, when the exclusion opportunities are, whether companies signed up in DI group, is there a form to persist their intent in UWA

Krcsmith Rotan: Could it turn out that we need cooperation of all members

Krcsmith Dave: Hopefully just this group.

stef_dub ACTION: dave to send a mail to chairs,public-uwa,www-mobile to announce the LC of cc/pp 2.0 and the extension of the review period till july 6 2007 [recorded in http://www.w3.org/2007/06/dublin-uwa-f2f-mins.html#action27]

trackbot-ng Created ACTION-22 - Send a mail to chairs,public-uwa,www-mobile to announce the LC of cc/pp 2.0 and the extension of the review period till july 6 2007 [on Dave Raggett - due 2007-06-15].

Krcsmith ACTION: Dave to contact Ian Jacobs for a status report on the patent policy details for transferred work items [recorded in http://www.w3.org/2007/06/dublin-uwa-f2f-mins.html#action28]

trackbot-ng Created ACTION-23 - Contact Iain Jacobs for a status report on the patent policy details for transferred work items [on Dave Raggett - due 2007-06-15].

external groups: http://www.w3.org/2007/uwa/external-groups.html

Discussion on above page...

These are groups that might be of interest at various stages of our work.

Could move to informative space on wiki.

Give group members responsibility of adding to informative pages.

For example, identifying which have official liaison with W3C, which have well-known contact points etc.

Krcsmith ACTION: sboyera to look at possibilty of Keio hosting F2F in March [recorded in http://www.w3.org/2007/06/dublin-uwa-f2f-mins.html#action29]

trackbot-ng Created ACTION-24 - Look at possibilty of Keio hosting F2F in March [on Stéphane Boyera - due 2007-06-15].

Krcsmith Kangchan: could look at co-hosting with Mobile Web 2.0 forum in March

Krcsmith Dave: Thanks Rotan for hosting (applause)

Krcsmith ACTION: Kangchan to look at hosting in Korea [recorded in http://www.w3.org/2007/06/dublin-uwa-f2f-mins.html#action30]

trackbot-ng Created ACTION-25 - Look at hosting in Korea [on Kangchan Lee - due 2007-06-15].

Krcsmith Meeting adjourned for day 2.

Summary of Action Items

Pending

[PENDING] ACTION: Dave to review waf widget spec.
[PENDING] ACTION: Dave to organise transition call for DISelect.
[DONE] ACTION: Keith to discuss DOM 3 implementation as a requirement for DCCI with Rhys when he returns.

New

[NEW] ACTION: Stephane to report back on possibility of using Semantic Media Wiki
[NEW] ACTION: Dave to prepare an e-mail as an interim summary of the workshop
[NEW] ACTION: Dave to contact Ian Jacobs for a status report on the patent policy details for transferred work items
[NEW] ACTION: Dave to invite JCF to the layout taskforce
[NEW] ACTION: Dave to proceed with inviting expert Ryan Sarver.
[NEW] ACTION: dave to send a mail to chairs,public-uwa,www-mobile to announce the LC of cc/pp 2.0 and the extension of the review period till july 6 2007
[NEW] ACTION: Kangchan to look at hosting in Korea
[NEW] ACTION: Kevin , ref- DIAL spec issue-validation, use consistent text throughout for source infoset and result infoset (as per DISelect)
[NEW] ACTION: Kevin ref- DIAL spec issue-validation, add text 'A DIAL document must be valid before and after processing' to the spec.
[NEW] ACTION: Kevin Smith to coordinate with Steven Pemberton on XHTML2, and schema related issues in two weeks.
[NEW] ACTION: Kevin to clarify wording to address AVT issue, in two weeks.
[NEW] ACTION: Kevin to coordinate with Steven Pemberton on XHTML2, and schema related issues in two weeks.
[NEW] ACTION: Kevin to include text stating DIAL specification should state that DIAL is a true superset of XHML 2, as per resolution, in two weeks.
[NEW] ACTION: Kevin to mention the intent to ensure that the DIAL schema is extensible in two weeks.
[NEW] ACTION: Kevin to remove para 2 and ed note from section 2.10 "Content Inclusion" of DIAL in two weeks.
[NEW] ACTION: Rhys to review the latest draft of the CSS layout module
[NEW] ACTION: Rotan to report back on the kinds of scripted components for layout that are being used in the Ajax world
[NEW] ACTION: sboyera to look at possibilty of Keio hosting F2F in March
[NEW] ACTION: stephane to report back on the steps needed for a transition to CR for DIselect and XAF

Chair notes that rrsagent tracking of actions goes to pieces when gluing together minutes from multiple days as it starts the numbering from 1 each day, sigh! I therefore deleted the recorded in text due to the broken links after a failed attempt to fix them.

[End of minutes]