W3C

Model-Based User Interfaces XG F2F

Mandelieu, 24 Oct 2008

See also: IRC log

Attendees

Present
Dave Raggett, Fabio Paternò, Sebastian Käbisch, Joachim Laier, José M. Cantera Fonseca, Juan Manuel Gonzalez Calleros, Jaroslav Pullmann
Observers
Charlie Wiecha, Roland Merrick, Daniel Appelquist, François Daoust, Richard Schwerdtfeger, Andy Heath, Wonsuk Lee, Seungyun, joseph
Chair
Dave Raggett
Scribes
Richard Schwerdtfeger, Daniel Appelquist, Dave Raggett, José M. Cantera Fonseca

Contents


Presentations/Papers


Preamble

Charlie: Patent Policy?

dsr: As an incubator group we are not creating standards

dsr: we have a more fluid approach

dsr: there is no commitment to make patent disclosures unless we move work into a working group

dsr: let's look at the charter

Charlie: ... reading patent policy

Charlie: Looks like option 3 - the RF policy

dsr: it was weakened to encourage broader participation

Charle: This is the W3C so could there be a charter statement about web architecture principles

Charlie: It should say we are interested in Web Centric approach - a layered UI

dsr: I think that was in the beginning

Charlie: It is more a point of discussion

dsr: My company is looking more at the web architecture

Charlie: we should expect things to run in a loosely coupled, http like, environment

dsr: the idea to have loosely connected teams plays well into this

Charlie: I was looking at reemphasis of this

dsr: Can we assume this is a working assumption of the group

RESOLUTION: It is agreed that we would have loosely coupled layered systems

Charlie: Transformations can get us to a run time point of view

dsr: we should get a better look as we go around the room

Fabio Presentation on Universal declaritive User Interface Description Languages

fabio: Lot of work in this area - late 80s

Charlie: 3rd generation of work

fabio: how to exploit multiple layers

fabio: Different abstraction layers (task and object, abstract interface, concrete interface)

fabio: this is not only a top down approach

fabio: older proposal for notation representing task models

fabio: concurrent task trees

fabio: set of temporal relations

fabio task allocation

fabio: Concurrent Task Trees (CTT) have become a defacto standard for task modeling

fabio: CTT should be a standard model

Charlie: Is this the human model of tasks

Charlie: no

Charlie: tie in with BPMN

Charlie: human processing might be a nice complement

fabio: XForms does not completely solve adaptive description

fabio: XForms have abstract and concrete concepts

Charlie: we wanted this to be a web application spec. initially

Charlie: there are separate applications for RIAs

Charlie: Nothing in xforms relates to the task model

Charlie: in Flex and Silverlight will find similarities

fabio: you can model drag and rop interaction in XForms

dsr: We can collect use cases and do gap analysis on the spec.

fabio: Teresa XML

fabio: we created an authoring tool abstract and concrete

fabio: support various platforms: DXHTML, SVG.HTML/javascript, Digital TV, ...

Charie: where is Teresa

fabio: abstract and concrete

fabio: 2 European projects. ServFace model driven service engineering methodology

open: http://www.ict-open.edu

fabio: illustrating open project model-based incubator framework

fabio: Maria XML

fabio: Maria XML based on Teresa supports abstract data types

fabio: able to generate UI AJAX/JavaScript

fabio: Maria authoring tool

jose: you could do much more than complex transformations

fabio: the source and target are xml documents then you are ok , otherwise ...

fabio: you can generate more powerful approaches but then there is performance

dsr: how easy is important too

dsr: there is a lot of work on transformation languages

jose: transformations between layers of abstraction is a key part

dsr: We would like to open authoring of applications to non-programmers. So, we need easy to use authoring tools

fabio: Maria is at the abstract layer

fabio: Maria would replace Teresa

dsr: lets get back to the various layers later

UsiXML a UI Model-based authoring approach and language engineering approach

juan: our group has 20 years of experience

juan: we have experience with multi-modal, haptic, 3d, 2d interfaces

juan: we are in Belgium

juan: we are working with UsiXML consortium

juan: we modeling the UI of the HMI

juan: in a plane cockpit

juan: http://www.usixml.org

juan: all information can be downloaded to an image CD

juan: we divide our work for different access

juan: usixml defines a common language

juan: I want to give you the UI transformation model where each layer is described

juan: we are not looking at just the task model but also the business process model

juan: defining roles of each actor, aspects of collaboration

juan: demonstrates gesture-based authoring tool of widgets

juan: the method is 4 layered.

juan: 4 layers for defining the task (abstract, concrete, etc.)

juan: we need some transformation of the UI

juan: we express how to transform form one level to the other ( a graph transformation on each node)

juan: shows example of the methodology

juan: task descriptions, abstract descriptions, concrete UI, and final UI

juan: We have some authoring tools that might help you but not one tool to support the whole process.

juan: GrafiXML allows the user to create multi-language GUI

juan: you cna edit yourself some parts of the XML

You can create a conextModel for Drag/Drop

juan: you can show how can split between different UIs including the mobile phone

jose: Basically all your tools generate code but no runtime

juan: that is correct

juan: we use the CTT notation

jose: Your teresa is only at design time?

fabio: we have tools to generate runtime using the same language

Charlie: the XG should be clear on what W3C is doing for a more flexible runtime across a set of distributed organizaitons

Charlie: business partner may be adding their own UI components

Charlie: to get the W3C language we need pictures of those runtime scenarios

joseph: does one single document define the process

juan: we split the pieces into different repositiories

joseph: does UsiXMl describe all the parts?

joseph: yes

joseph: so one UsiXML includes all the design decisions made

joseph: so UsiXML could be an executable language

yes

Model Based Service Composition presentation from Joachim [Jo] Laier

Jo: [contents of presentation]

Jo: I would like to suggest some work items for the group [as part of my presentation]

Jo: ["service mashups" slide]

Jo: Energy management - available of electricity variable - establishing technology to enhance usability and efficiency

Jo: [Also healthcare, business process, industrial automation]

Jo: Common ground to these and common issues.

Jo: From the lessons learned from the mobile web, we could enhance usability in enterprise-oriented systems.

Jo: [slide "our expectations towards xg results"]

Jo: First want to learn.

Jo: would like to understand better how to combine relevant standards

Jo: [slide: "current work @ siemens"]

Jo: Working on service composition.

Jo: So far using x-forms - but no powerful complete x-forms implementations exist.

Jo: X-Forms also limited for distributed systems.

Jo: We have just started to investigate the semantics behind all of this.

Jo: If all of these services [being composed] bring their own information than this information needs to be aggregated into a single UI.

Jo: Network should be a dynamic topology -- if you bring a new device home, it should find other devices and re-configure itself, so you don't need to be an expert for installation.

Jo: [slide on "proposed work items"]

Jo: We should agree on a common set of use cases. If we don't focus we won't achieve anything.

Jo: From the use cases, it may be sensible to [find] requirements.

Jo: from requirements, we hope to develop to methods and strategies to get some long-term result

Jo: For the report, we should have a compliation of best-practice rules, design patterns, and recommendations for new working group.

Jo: [slide: discussion - mind map]

Dave: Thanks!

[Talk from José Cantera]

Jose: We started a long time ago trying to develop a declarative language in order to be adapted to mobile devices

Jose: MyMobileWeb project

Jose: Can support different scenarios. Displaying data as a map or paginating data (for example).

Jose: This adaptation happens at run time.

Jose: You precompile which generates the runtime support that allows at runtime for you to render the UI for different devices.

Jose: This precompliation creates an object model and some code.

Jose: You can't generate code for every device so this happens at runtime.

Jose: Our language was built for the typical mobile web use cases.

Jose: We used CSS properties.

Jose: now (after 3 years) taking into account this experience, we have created a new version of the IDEAL language.

Jose: Necessary to be modular and extensible.

Jose: We have tried to reuse existing things where possible but x-forms didn't give us everything we needed.

Jose: We built it for mobile applications but it could evolve to any modality.

Jose: We have tried to align with DIAL.

Jose: [IDEAL history]

Jose: When you have a UI description language how do you bind to [e.g. Dojo, Yahoo UI, etc...] We should come up with something to bind from the UI language to the final toolkit you use.

Jose: We thing things like javascript are horrible for adaptation - you need something more declarative.

Jose: The idea is to allow not only XML data models but also JSON models or whatever model you would like.

Jose: The idea is to have a more advanced language to be able to create more compelling applications.

Jose: Looking at using RDFa to attach more semantics.

Jose: [next slide]

Jose: IDEAL specification suite could be defined around IDEAL. IDEAL is licensed under a creative commons licence.

Jose: We have a working draft published on the web: http://mymobileweb.morfeo-project.org/specs/ideal/

Jose: We are developing an engine behind this which will be open source.

Jose: Adaptation Policies - you need to policies on how layout can change depending on the context, etc...

Jose: context-aware reordering policies - e.g. reorder a table depending on location based on a geolocation API.

Jose: We need simple guidelines on how to translate from the abstract to the concrete.

Jose: [slide on examples]

DKA: What license for the open source implementation?

Jose: GPL

Jose: Another policy we are trying to explore - have a dual mode - license the technology separately to someone doing something commercial.

Jose: but in general it's GPL.

Scribe: Dave

ScribeNick: dsr

Charlie Wiecha presents on XML technologies on the client

This could lead to collaboration either with the XForms WG or the Back Plane XG.

Mixing XML and JSON data is really interesting.

Three points: a) ubiquitous XForms through zero-install JavaScript libraries for web browsers

b) supporting XForms via additional attributes in HTML

where the goal is to preserve the HTML syntax whilst converging with the XForms architecture.

c) beyond forms with smaller bite sized specs

plus a means to support composition using forms modules.

Charlie: interested in plugging in behavior server-side.

The xforms bind elements can be used within html as empty elements.

We have the full XForms spec running on the client without any need for server-side transclusion.

This is thanks to a script that interprets the DOM tree and supplies the behavior.

This was developed by John Boyer and Mark Birkbeck under an Apache open source element.

This can be thought of as a poor man's XBL.

We support XPath expressions and some schema data types but not yet full schema validation.

We have XML Events and core XForms.

This allows you to choose between compiling down to HTML or to supporting XML technologies via JS run time libraries.

Charlie shows an example

This allows you to interact via voice via launching a VoiceXML component or via launching an ODF-based spreadsheet.

Fabio: the point is that the UI components share the same data-model, right?

Charlie: yes, they share data model via XML events.

Next Steps

We will have conference calls on Monday at 3 PM European Time

ACTION: dsr to create a wiki [recorded in http://www.w3.org/2008/10/24-mbui-minutes.html#action01]

ACTION: dsr to create a wiki [recorded in http://www.w3.org/2008/10/24-mbui-minutes.html#action02]

ACTION: Juan to provide a comparison of task models so far studied by the Academic Community [recorded in http://www.w3.org/2008/10/24-mbui-minutes.html#action03]

ACTION: Fabio to provide CTT materials to be studied by the group [recorded in http://www.w3.org/2008/10/24-mbui-minutes.html#action04]

Roland: What is the relationship of the final user experience?

charlie: looking for a cross-organizational scenario

charlie: the abstract UI comes from one organization, the concrete UI to other scenario, etc

charlie: interoperability is also central

Roland: this things can happen at different times

dsr: there is a lack of standards on this

dsr: we need to explore the interoperability needs

Jo: We would like to go into the use cases

dsr: we need to create the framework for the use cases

Jose: We need a category of use cases, and its comparison to other approchaes and the advantages

Jo: The first 6 months is the analysis phase

Jose: We have not talked about F2F meetings

charlie: A virtual F2F meeting

Fabio: It can be at some event

Fabio: suggests to go to an specific conference in Boston

By default a meeting every six months

If we find a conference or event in Europe that could be potentially attended by members we will consider to meet there

next call: November 10th

ACTION: Sebastian to coordinate collection of use cases [recorded in http://www.w3.org/2008/10/24-mbui-minutes.html#action05]

ACTION: Fabio to provide some material about XForms from the point of view of the abstract UI models coming from the -Academic Commiunity [recorded in http://www.w3.org/2008/10/24-mbui-minutes.html#action06]

dsr: free to invite other people to contribute

dsr: we need to follow the W3C rules, and make expert invitations

dsr: we look it at it further, about the rules for expert invitation

Jose: anyone can contribute via the public mailing list

Jose will help Dave with the chairing

[adjourned]

RSSAGent make minutes

RSSAGent, make minutes

Summary of Action Items

[NEW] ACTION: dsr to create a wiki [recorded in http://www.w3.org/2008/10/24-mbui-minutes.html#action01]
[NEW] ACTION: dsr to create a wiki [recorded in http://www.w3.org/2008/10/24-mbui-minutes.html#action02]
[NEW] ACTION: Fabio to provide CTT materials to be studied by the group [recorded in http://www.w3.org/2008/10/24-mbui-minutes.html#action04]
[NEW] ACTION: Fabio to provide some material about XForms from the point of view of the abstract UI models coming from the -Academic Commiunity [recorded in http://www.w3.org/2008/10/24-mbui-minutes.html#action06]
[NEW] ACTION: Juan to provide a comparison of task models so far studied by the Academic Community [recorded in http://www.w3.org/2008/10/24-mbui-minutes.html#action03]
[NEW] ACTION: Sebastian to coordinate collection of use cases [recorded in http://www.w3.org/2008/10/24-mbui-minutes.html#action05]
 
[End of minutes]