Thursday, 24th July 2007, CNR, Pisa, Italy.
A one day meeting hosted by the HIIS Laboratory of the
Istituto di Scienze e Tecnologie dell'Informazione
Area della Ricerca del CNR
Istituto di Scienze e Tecnologie dell'Informazione
Via G. Moruzzi 1, 56124, PISA, Italy
Aula "A.Faedo"
Local Participants:
Remote Participants:
Regrets:
We will try to enable remote participation via IRC and skype. Contact us on IRC and we will then try to sort out how to use skype effectively. We will have to use IRC to manage things due to the limitations of a close talk microphone. Please send any presentation materials and other documents to Dave Raggett who will make them available to the group.
The meeting will start at 9:30 promptly on Thursday 24th and Friday 25th July, so aim to get to ISTI around 9am. Lunch will be provided by our hosts at 13:00, together with breaks at 11:00 and 16:00. We will be on our own for dinner, but we should be able to get help with reservations for group meals for those interested. I will be staying at the hotel Bolgna in the center of town (Via Mazzini 57, tel: +39-050-502120) and arriving on Tuesday evening for the model-based design meeting at ISTI on Wednesday.
There will be wireless network access, and we will take minutes on IRC on channel #uwawg, see http://www.w3.org/Project/IRC/ and http://www.w3.org/2007/uwa/telecons.html
14:00 Access mechanisms - Jose, Rotan and Kevin
15:00 DIAL and XHTML1 - Jose, Rotan
16:30 Looking forward
ACTION-94: jcantera to lead the creation of a page splitting and layout format for DIAL. Will Contact Volantis, MobileAware and Vodafone to try to produce something very quickly
ACTION-95: Dave to ask W3C Team for clarification of purpose of test suites and whether the one for DCCI needs to fully cover DOM Events.
ACTION-96: KRosenbl to prepare draft PR for DCCI
ACTION-97: KangChan create revision of Device Coordination Note
ACTION-98: jcantera to contact Ivan to arrange a joint teleconference on units and see about possibility of a joint meeting during the Tech Plenary.
ACTION-99: Kevin to assess reuthoring of XPATH access functions to consider new generic syntax.
ACTION-100: jcantera to investigate adding tabindex into DIAL for use with any element and also adding values for zero and minus one.
ACTION-101: dsr to contact XHTML2 and XForms WGs on updated DIAL proposal and ideas for a joint profile.
<Cloud> Good morning from Galway
We go around the table introducing ourselves.
Good morning John!
<johnbreslin> i will join in when i get to deri, still at home
<johnbreslin> If there is a Skype channel set up later I can join
<scribe> agenda: http://lists.w3.org/Archives/Member/member-uwa/2008Jul/0046.html
John, we are trying to set up skype now. I don't think it will be practical for you to hear us, but we should be able to hear you.
<johnbreslin> Okay Dave, no problem
<johnbreslin> okay can you hear me?
it seems like we are connected, but you can't hear us, we can hear you
This is something that we inherited from the DIWG
It has lain dormant and the question is whether we want to resume work on it or not.
The driver was a means to control how pages are split into smaller ones for delivery tp mobile devices.
Jose thinks that the requirements are well understood and we should invite indiustry submissions on potential solutions.
Jose is interested in submitting a proposal, but wants to first talk with Rotan and Paul to see what MobileAware and Volantis would like.
The idea is to re-use an existing proven solution rather than to develop a new one.
Fabio: the concrete UI can
include compositional operators which may be of value.
... doesn't this assume that you are starting from one page for
all devices?
Jose: yes, that is what DIAL provides, and this is about splitting this page into pieces suited to particular device constraints.
Dave asks for clarifications. It seems that this work item is about a short term need for DIAL.
The XG would work on a longer term that looks at interaction more generally than web pages.
Rich: annotations on web pages should be of value for the rules used for the page splitting process and navigation amongst the subpages.
<scribe> Scribe: dsr
Jose: splitting and layout are separate. There needs to be some control over splitting.
For example, certain things must be kept together.
Rich: do you (Jose) use Ajax as part of the splitting process?
Jose: no, it is done server side based upon the delivery context.
Fabio: do you support client side scripting?
Jose: it depends on the device.
<danh> hi all
<danh> I'm Danh Le Phuoc from DERI
Jose: Ajax may be used within page components where appropriate.
But we can't rely on that as not all devices support Ajax.
<johnbreslin> Hi Danh - I'm on a skype connection but it comes and goes
<johnbreslin> So can monitor on IRC anyway
Dave waves to Danh, and asks if you want to introduce yourself via John's skype connection?
<johnbreslin> Danh - you can come to my room if you want to say hello
<danh> I'm not at DERI building now
<johnbreslin> Or maybe you can join the Skype connection directly
<danh> yes,how can I join by skype?
<johnbreslin> is your ID lpdanh??
<johnbreslin> let me see if i can pull you in, or else dave might be able to
<jcantera> ACTION: jcantera to lead the creation of a layout declarative format for DIAL. Will Contact Volantis, MobileAware and Vodafone to try to produce something very quickly [recorded in http://www.w3.org/2008/07/24-uwawg-minutes.html#action01]
<trackbot> Created ACTION-94 - Lead the creation of a layout declarative format for DIAL. Will Contact Volantis, MobileAware and Vodafone to try to produce something very quickly [on Jose Manuel Cantera Fonseca - due 2008-07-31].
danh, we are trying to add you into the call ...
Fabio presents on work in his lab at ISTI.
(Slides to be provided later)
The essential idea is to suspend an application and to resume it after changing to a new device.
Device discovery, when to migrate, where to migrate, what to migrate, how and what means to support state persistence.
<richardschwerdtfe> *
Fabio's approach utilizes a model-based approach where the concrete UI is reverse engineered.
from XHTML
Semantic rules are used to control what components need to be kept together during page splitting for delivery to mobile devices.
Fabio notes that few sites use XHTML cleanly, and it is interesting how JavaScript is being used.
Jose: JavaScript usage is a complete mess.
Fabio: we are proposing guidelines for how to use scripting.
Rich: have you thought about encouraging authors to include metadata to guide adaptation?
Fabio: our challenge was to infer such metadata.
The page splitting process uses an iterative search for the most appropriate way to recasting a page to match the constraints for a given target device.
Rich explains how ARIA would help with the semantic labelling, increasing the reliability over automatic labelling.
The challenge is how to convince authors to provide labels.
Rich: people are creating mashups by combining components, and the concepts should be transferable to content adaptation.
Fabio: we have another project for annotation WSDL descriptions of web services.
We use Ajax to push the state of the web page DOM to the migration server.
In addition to migrating from desktop to mobile, we have worked on mobile to digital TV.
Fabio shows a video of a demonstration.
The user can discover devices to migrate to and save the state for the migration.
Andy: this is something that could be applied to migrating other kinds of application than HTML.
Fabio: reverse engineering from Java is tough. XHTML is much easier to work with.
Application state covers a wide variety of things, e.g. user input, client-side variables, dynamic content (Ajax scripts), cookies, bookmarks, history, and query strings.
Fabio talks about factors relating to the usability of migratory interfaces.
<andy> rather - other kinds of media (report of my earlier comment)
Fabio asks if W3C has anything relevant to managing the state as needed to support migration.
Dave: one thing is to separate the state from the UI, e.g. as in XForms, so that the state can be decoupled from its location.
For bookmarks, there are existing services for storing your bookmarks online (e.g. Delicious), so there is user experience with that aspect.
KangChan asks about the relevance of W3C work on the delivery context and DIAL?
Fabio: yes, these could be used to assist with constructing the models.
<andy> There is a group in ISO/IEC JTC1 SC36 that is producing a technical report on how to migrate state (and other things) across mobile technologies - I think the world needs one solution here not two different ones - there is scope to harmonise. I have a toe in participating in that work so I will raise the issue again at appropriate points
Dave wonders if it would be practical to use metadata for describing the pertinent data for migrating application state.
This could perhaps be handled in a similar way to the delivery context, but for applications.
This would involve metadata that is held separately from the application rather then as Rich suggests embedded into web pages.
Jose asks about how links between pages are handled?
These are handled by the migration server.
Johannes asks about the cost functions used in searching for the best way to split pages.
The process operates at both the abstract and concrete UI levels as appropriate.
Rich: how does this work for secure website?
Fabio: we don't deal with that.
Rich: it is a really hard problem and I was involved in an IBM project on this some years back.
KangChan: there are many problems with secure/access controlled pages.
Dave wonders is the content transformation taskforce at W3C may have some insights.
Jose gives a brief description - the idea is to allow users and content owners to declare their preferences in regards to content adaptation.
Rich: this also relates to digital rights management issues, e.g. legal implications for transforming medical images.
Jose: transcoding proxies have a lot of issues, e.g. when they change the user agent header.
He cites a ring tone website where the ringtones weren't available because the content owner mistook the client as a desktop not a mobile.
Rich: online banking is another example where the risk of screwing up a transaction could lead to lots of lawsuits.
Dave wonders whether the default is to allow content adaptation unless the website explicitly forbids it.
Rich and Jose describe potential problems when trancoding pages with Ajax components.
<johnbreslin> Good question - I assume that's what Opera Mini does? Is there a spec to allow sites to forbid?
Jose would recommend dropping the work on reverse-engineering the page models.
John - take a look at the content transformation task force in the Mobile Web Best Practices WG.
<johnbreslin> Thanks Dave
Dave asks if Fabio has applied statistical machine learning techniques for the content labeling problem?
Fabio, no, not yet.
Dave notes that it requires a corpus of pages that have been labeled by people to drive the learning algoriths and the evaluation of their effectiveness.
Thinks this was in a paper in the Banff WWW conference a couple of years back.
Discussion about possible loss of information as a mobile version of an app may not include everything you would like on a big screen version.
Andy: we need to coordinate work on metadata to avoid problems.
where metadata is being developed by different organizations.
Jose describes a number of device description repositories.
scribe: we break for lunch ... back in an hour
Andy and Rich give us their presentation.
[time passes as we sort out projection problems]
Jose lends his PC and Rich gets started with the talk.
(slides to follow later)
ARIA is great for Web 2.0 apps with full interoperability wuth assistive technologies. keyboard navigation etc.
However, one size doesn't fit all.
Rich describes "mashups" as web apps combining data from more than one source into an integrated experience.
Rich talks about the problems with a one size fits all approach to accessibility. Essentially this results in a reduced quality user experience.
AccessForAll is an initiative focusing on reducing the mismatch to individual users.
This started in education, but has extended beyond that.
Andy cites the example of a library terminal where you want a fast and painless way for the application to dynamically adapt to the user.
This needs a description of the user's preferences and needs as well as a standardized description of the delivery context and resource capabilities.
We can then deliver an optimized experience to that user.
We are seeking a road map for harmonization.
This should cover metadata and transport.
There are a number of standards development organizations involved, e.g. IMS Access for All, ISO JTC1 SC36, EU4ALL. W3C delivery context ontology, OMA DPE.
The ISO work covers individualized adapability, and metadata for learning resources.
The ISO work has led to an approach for user modeling
We note that the delivery context doesn't currently cover modalities as such.
The display model has a variety of properties. The core is a property for enhancing visibility.
There are also modality preferences.
Hazaards are things like repetitive flashes that may trigger epilepsy
color coding can present problems for people with various forms of color blindness.
What kinds of adaptation are needed for desktop, mobile and digital TV clients?
Andy notes that sometimes you need to know if the device can perform an adaptation or whether this needs to be done server-side.
Jose: we need a new module in the ontology to address personalization.
Somethings will be shared, e.g. font properties.
Jose wonders whether we may want to consider a broader name than delivery context ontology.
Dave points out that the delivery context has included user preferences within its scope since the start, but content properties would be something new.
Rich: how can we indicate that some properties are only for the client and aren't required by the server?
Dave asks for clarification as to the scope of user modeling. Is it restricted to user preferences and needs or does it cover other information such where a user resides (user profile).
Andy: the former.
Dave: what is the road map for this work? what documents/specifications/liaison actions?
Dave thinks that there is value in publishing a Note that explains the context. as that is unlikely to be covered in details in the ontology specification.
Andy: we anticipate part of this becoming an ISO standard and mandated by governments etc.
<johnbreslin> Must be :)
Hi keith, we are resuming after a demo, and will try and call you now.
Keith, what is your skype id?
<KRosenbl> keithrosenblatt
------------
- Merged Implementation report and draft PR for DCCI
- Discussion with Sailesh before he left on an extended
vacation: http://www.w3.org/2008/06/19-uwawg-irck.
The DCCI property filter interface is at risk as neither
Keith nor Sailesh have implemented it.
Access control is mentioned in the spec but not
normatively. This is something to cover in the
security workshop at the end of this year.
Sailesh thinks that we should allow for implentations
involving access to static properties only.
------------
Jose displays the DCCI assertions (a large PDF file).
Keith walks us through them along with the implementation results.
Neither Nokia nor Orange have implemented the property filter interface, but these were identified as being at risk in the CR.
This is okay as we can drop this feature as we move to PR.
Keith notes that both implementations don't completely comply with the DOM specification when it comes to duplicate event handlers.
Keith isn't sure as how important that is. Perhaps he has gone too far into the DOM event specification in the tests.
Dave thinks we only need to cover assertions that are specific to the DCCI spec and not those that are already covered by existing W3C Recommendations.
This is something we can check with the W3C team.
<scribe> ACTION: Dave to ask W3C Team for clarification of purpose of test suites and whether the one for DCCI needs to fully cover DOM Events. [recorded in http://www.w3.org/2008/07/24-uwawg-minutes.html#action02]
<trackbot> Created ACTION-95 - Ask W3C Team for clarification of purpose of test suites and whether the one for DCCI needs to fully cover DOM Events. [on Dave Raggett - due 2008-07-31].
<andy> Collaboration on Metadata: We agreed that the Ubiquitous Web Working Group will work on integration of personalisation into the device context ontology. The DCO is synchronized with how preferences are passed to devices. The definition of the terms/vocabulary shall be owned by IMS/ISO and Rich and Andy are going to create a W3C note describing this working relationship with the W3C. The work may result in new versions or profiles of the IMS/ISO work. with device
Keith asks what we need to do next?
Dave: we need to prepare a draft PR.
<scribe> ACTION: KRosenbl to prepare draft PR for DCCI [recorded in http://www.w3.org/2008/07/24-uwawg-minutes.html#action03]
<trackbot> Created ACTION-96 - Prepare draft PR for DCCI [on Keith Rosenblatt - due 2008-07-31].
It might take a while to set up a transition meeting especially as we are in the holiday period.
We thank Keith for his hard work and look forward to reaching PR and after that REC status.
At this point we close the meeting for the day and turn to the pressing issue of a choice of restaurants for dinner.
<richardschwerdtfe> scribe: richardschwerdtfeger
<richardschwerdtfe> dsr: Kaz was going to make those contributions to the W3C. They wanted to describe the devices
<richardschwerdtfe> dsr: They wanted layering over the protocol mechanisms to describe the devices which is inline with our group
<richardschwerdtfe> dsr: It is not at the level we need.
<richardschwerdtfe> dsr: focusing on streaming media in the living room - specific class of devices
<richardschwerdtfe> dsr: The idea of having rich description of devices is a framework for coordination
<richardschwerdtfe> KangChan: I think the device meta data framework and metadata work can allow us to align
<richardschwerdtfe> KangChan: we can create a framework including a meta data definition
<richardschwerdtfe> dsr: Can we define an ontology, what is unique for each device, and can we coordinate
<richardschwerdtfe> KangChan: For device coordination we need to maintain the state of the device
<richardschwerdtfe> KangChan: I think the current deliverycontext ontology can support the state information
<richardschwerdtfe> dsr: we need to define how we support this in a modular way
<richardschwerdtfe> dsr: this is not really research because there has been a lot of working in describing devices
<richardschwerdtfe> KangChan: we need a framework for meta data and device coordination
<richardschwerdtfe> dsr: we are looking for input based on existing work
<richardschwerdtfe> dsr: we need to target companies working in this area
<richardschwerdtfe> KangChan: we can merge the use cases with the same intention and viewpoint for a device
<richardschwerdtfe> dsr: I am looking for Kaz to submit his materials and we need to decide whether to create as one document or broaden the context
<richardschwerdtfe> dsr: we will work on device coordination later
<richardschwerdtfe> dsr: We need to at least get some clarity on the roadmap for personalization
<scribe> Continued from Thursday evening.
One issue is how to describe fonts when it comes to readability?
Andy: we want to express this logically e.g. emphasis and contrast rather the font size in pixels, etc.
This is proving to be a challenge.
Jose: the DC Ontology has a property for proportional fonts, but that's all. We should plan on adding a new module for the required vocabulary.
We should define a class for fonts.
Andy: we would like to know if the client supports a means for users to make the fonts bigger or smaller, overall.
Jose: the key is to be able to model all the things that are relevant to adaptation.
The ontology isn't intended to cover every aspect of a device, but rather only those relevant to adaptation.
Rich: coming back to specifics, should we use bold or emphasis or contrast?
Dave notes that point size isn't a reliable indication of legibility as fonts vary in the x-height.
Perhaps we need to define qualitative measures, e.g. most legible font, and size relative to the default for that device.
Dave adds that contrast is perhaps best considered as a separate dimension as high contrast themes cover much more than fonts.
Rich and Andy agree on doing a study of the needs and expressing those for fonts.
Rich would like to cater for users to provide different profiles on different devices.
Andy: the user preferences may also change according to the context.
Dave suggests that it would be valuable to clarify the use cases for this.
<richardschwerdtfe> scribe: richardschwerdtfeger
<richardschwerdtfe> wonsuk: the goal of this document is device coordination within the w3c
<richardschwerdtfe> wonsuk: need to make a W3C standard
<richardschwerdtfe> wonsuk: introduction introduces concepts and scopes the coordination activities
<richardschwerdtfe> wonsuk: 1.1 conversion device coordination concepts
<richardschwerdtfe> wonsuk: There are companies have already developed services in this area but they don't provice for adequate interoperability and extensibility
<richardschwerdtfe> wonsuk: to meet this goal we have to consider challenges to support multiple devices
<richardschwerdtfe> wonsuk: We need to define the relationship between device and services and how to bind them
<richardschwerdtfe> wonsuk: we need to consider the user preferences.
<richardschwerdtfe> wonsuk: for UWA, services, devices, resources, nad user preferences we need to harmonize
<richardschwerdtfe> wonsuk: The device coordination is responsible for handling the interdependencies between all of these
<richardschwerdtfe> wonsuk: we have defined the mathmatical coordinaton
<richardschwerdtfe> wonsuk: device coordination = set of service request devices with capabilities times the set of service providing devices with a set of resources
<richardschwerdtfe> dsr: is there previous work using this mathmatical approach?
<richardschwerdtfe> KangChan: no
<richardschwerdtfe> wonsuk: cc/pp is mentioned and we can refer to this. We also refer to DCCI. We also refer to DIAL to produce web pages in line with device capabilities
<richardschwerdtfe> wonsuk: we refer to RDF to and SCXML (to maintain state of the devices)
<richardschwerdtfe> wonsuk: these are to achieve seamless device coordination
<richardschwerdtfe> dsr: we should include tie in with services
more specifically, it would be useful to explain how device coordination relates to work on web services, and how it differs.
<richardschwerdtfe> rich: we need to tie in the total picture with what the mainstream mobile community will be doing such as DPE. We refert to CC/PP here but CC/PP has limited uptake.
for example device coordination may involve web services but that will not be true in all cases.
<richardschwerdtfe> KangChan: we will work on that for the next revision
<richardschwerdtfe> wonsuk: the entire picture of device coordination is not easy
<richardschwerdtfe> wonsuk: I also we need a good picture of this in the W3C point of view
<richardschwerdtfe> dsr: If people have different ideas of device coordination what are they?
<richardschwerdtfe> dsr: We need to look at what question they might have
<richardschwerdtfe> rich: when I talk to web 2.0 mashup people the discussion of device compatability never arises. The don't even know what that would mean
<richardschwerdtfe> rich: collecting those questions are hard
<richardschwerdtfe> dsr: what questions they have is hard
<richardschwerdtfe> dsr: you are familiar with UPnP. What are their perspectives
<richardschwerdtfe> dsr: This does not have to be big. We need use cases.
<richardschwerdtfe> wonsuk: The main issue is what is the communication protocol - UPnP, Bluetooth, etc.
<richardschwerdtfe> wonsuk: UPnP is heavy. A simple and generic standardized to web-based communication
<richardschwerdtfe> wonsuk: need coordination mechanism and need to address security and privacy
<richardschwerdtfe> fabio: bluetooth is not a communication protocol
<richardschwerdtfe> KangChan: from the point of view of device coordination we need to consider it
<richardschwerdtfe> fabio: these are really two different issues. one is the low level hardware communication layer but then their is the application communication layer.
<richardschwerdtfe> KangChan: there is a logical communication model.
<richardschwerdtfe> KangChan: we are not planning to consider network protocols.
<richardschwerdtfe> dsr: you need to address the coordination layers
<richardschwerdtfe> KangChan: in web services there are servics to describe the services
<richardschwerdtfe> KangChan: device description is critical for device coordination
<richardschwerdtfe> dsr: we need to discuss how devices information is routed
<richardschwerdtfe> fabio: It is also possible to have partial device coordination.
<richardschwerdtfe> wonsuk: the use cases include short descriptions.
<richardschwerdtfe> dsr: you are going to cover the use cases?
<richardschwerdtfe> wonsuk: yes
<richardschwerdtfe> wonsuk: we have a definition of terms
<richardschwerdtfe> KangChan: we found no real definition of device coordination
<richardschwerdtfe> fabio: I can see one application spread across multiple devices
<richardschwerdtfe> KangChan: we spoke about the level of device coordination. Strong, loosely, and un - coupled coordination are the different scenarios
<richardschwerdtfe> fabio: what is the goal - that application or the device
<richardschwerdtfe> dsr: the devices may only be there to support the coordination of services
<richardschwerdtfe> dsr: strongly coordianated: all events funnel through one coordinator
<richardschwerdtfe> dsr: loosely coordinated: talk briefly to a gate keeper and then you can use the device
<richardschwerdtfe> dsr: in order to operate safely you need to make use of a high level protocol.
<richardschwerdtfe> dsr: the idea is a overlay with rich description of devices at the highest layer
<richardschwerdtfe> dsr: may develop an application with devices employing different technologies. We would then abstract them at the higher layer
<richardschwerdtfe> andy: there are different use cases here
<richardschwerdtfe> dave: a high level goal to make it easier to build applications that run on a wide range of devices
with a mix of networking technologies, vendors and product generations.
<richardschwerdtfe> wonsuk: first is home networking where we have automatic volume control, home-theater systems, home appliances
<richardschwerdtfe> KangChan: from the interoperability we need to consider these types of use cases
<richardschwerdtfe> fabio: in this case: I am watching the tv on my mobile device and then on my digital TV
<richardschwerdtfe> fabio: so I might want to start on one device and continue on another
<richardschwerdtfe> andy: if you make your shopping list on a laptop you are a pretty sad person
<richardschwerdtfe> KangChan: there are several scenarios but what is the appropriate use case to describe the intent of the device coordination
<richardschwerdtfe> s/kanghen/kangchan/
<richardschwerdtfe> ACTION: Kangchan create revision of Device Coordination Note [recorded in http://www.w3.org/2008/07/25-uwawg-minutes.html#action03]
<trackbot> Created ACTION-97 - Create revision of Device Coordination Note [on Kangchan Lee - due 2008-08-01].
<Kangchan> Hi, krscmith.. it's coffee break. ^^
<johnbreslin> Skype seems to just work one way krcsmith as the mikes don't pick up everyone at the F2F - so you can talk to them but it may be difficult to hear in the other direction
<krcsmith> No problem, I'm happy to dial in via mobile.
hi Kevin, we don't have a phone here, but can easily call your computer if you have a skype client?
<krcsmith> Hi Dave, ok - let me set it up.
Jose presents on delivery context evolution. (slides to follow)
Jose notes to Rich http://mymobileweb.morfeo-project.org
[Kevin joins via skype]
<scribe> Scribe: dsr
Jose: Protege isn't a perfect tool, e.g. it has problems with large ontologies. This is one reason to split the ontology into smaller modules.
Protege makes it easy to define classes and ranges.
He shows us how to drive Protege on the screen.
Andy asks about the namespaces for vocabularies.
Jose: each namespace has a URI and local prefix
Putting properties into different namespaces makes it practical to modularize the ontology.
Protege can be extended with plugins, and we could in principle use that to manage a modular ontology.
Rich: have you looked at Altova SemanticWorks?
Jose: this is available free for use by W3C working groups along with XML Spy.
Andy: I have already been working
with protege importing owl, but found it tricky, just as Jose
says is the way not to use it.
... I found protegy is easier to use than SemanticWorks.
Jose: Swop is another tool, but
the development of it has now stopped.
... general issues. Protege is unmanageable for large
ontologies with around 100 classes and properties.
<andy> Andy says: sorry for not expressing well Dave: - what I'm saying is I have the CORE AccessForALL that we agreed in some other standards bodies and projects already in protege. I modelled the vocabs as separate OWL files but DID import them - which Jose says is not a a good idea.
We may want to reconsider how we deal with units to make this more convenient.
proposed actions include refactoring classes and properties in a number of namespaces.
Jose describes the proposed sub-ontology modules.
We need to provide guidelines on how to create new modules that are compliant with the ontology.
Jose notes that the current ontology doesn't make it clear that cm are a unit of length, but rather defines them as a unit of measure, which is wrong.
(or rather incomplete)
He shows us the MUO ontology for units. see http://idi.fundacionctic.org/muo
Fabio: what is the relationship to WURL?
Jose: WURL is a database for different devices and would conform to the ontology. The ontology describes the vocabularies, not the specific devices.
Fabio: the bluetooth profile doesn't seem to fit at the same level as other parts of the ontology.
Jose gives a quick explanation ...
Diego Berrueta is the author for the units ontology and will talk with Ivan Herman to see about creating a W3C spec for that work.
Dave notes that we can't normative reference the units work if that isn't a standard.
Jose we could make the units into a module in the Delivery Context Ontology.
But first we need to get the ideas reviewed to check that they are okay, hence the contact with Ivan.
The proposal is to use an RDF literal node for the units. see the above link for details.
Dave asks Jose when we expect to get feedback from Ivan and other RDF experts.
we could perhaps arrange a joint teleconference with Ivan to discuss units.
Given that we are in the Summer vacation period that may take some time to arrange.
Dave asks Jose how long it will take before a modularized version of the ontology will be ready for WG review?
Jose: around October 15, to give people a chance to study it prior to the Tech Plenary.
Dave: we may also want to arrange a joint meeting at the Tech Plenary?
Jose: perhaps with the Semantic Web Best practices WG?
<scribe> ACTION: jcantera to contact Ivan to arrange a joint teleconference on units and see about possibility of a joint meeting during the Tech Plenary. [recorded in http://www.w3.org/2008/07/25-uwawg-minutes.html#action04]
<trackbot> Created ACTION-98 - Contact Ivan to arrange a joint teleconference on units and see about possibility of a joint meeting during the Tech Plenary. [on Jose Manuel Cantera Fonseca - due 2008-08-01].
Jose: the other work item is to write a primer and here is a possible table of contents.
This would be ready by the end of 2008.
It would cover an overview, diagrams, relationship with APIs, relationship with existing vocabularies (e.g. DDR API), how to create a new compiant vocabulary, etc.
Hi Kevin, just tried calling you but it seems we heard you but you didn't hear us. will try again
Jose is just powering up the projector with your slides.
<andy> Scribe:Andy
<richardschwerdtfe> richardschwerdtfe: scribe: rich
<richardschwerdtfe> [07:07am] richardschwerdtfe: kevin: created an aggregation integration industry to create an moderator in the middle
<richardschwerdtfe> [07:07am] richardschwerdtfe: kevin: If you spend time integrating with a lot of operators and more time innovating that would be better
<richardschwerdtfe> [07:08am] richardschwerdtfe: kevin: This designed to help startups deliver mobile solutions quicker
<richardschwerdtfe> [07:08am] richardschwerdtfe: kevin: The status quota is such that network capabilities are walled off and huge authentication is in place
<richardschwerdtfe> kevin: the user does not get the same content from the desktop web
<richardschwerdtfe> kevin: need to pay for it
<richardschwerdtfe> kevin: There is an increase in traffic which would drive revenue opportunity. If there is a commonly support set of network apis which developers could use and sign up for they could be called and reused from server side, native apps, etc.
<richardschwerdtfe> kevin: the realization would be a simple set of apis which could commonly operated across operators
<richardschwerdtfe> kevin: this and opportunity for people to start using these apps. and reusing them
<richardschwerdtfe> kevin: we chartered the long tail of content producers
<richardschwerdtfe> kevin: on the project team we have 19 operators (big European, kelstra, etc.), aggregators
<richardschwerdtfe> kevin: we are working to leverage existing aggregator capabilities
<richardschwerdtfe> kevin: we are are looking to use the best of ParleX
<richardschwerdtfe> kevin: looking at popular development communities to refactor google gears, etc. to use common apis
<richardschwerdtfe> kevin: looking to take this back into the OMA
<richardschwerdtfe> kevin: will have more API definitions over the summer
<richardschwerdtfe> kevin: operators have done a poor job of this over the years
<richardschwerdtfe> kevin: trying to make this all easier for people
<richardschwerdtfe> kevin: we want open apis for people to be able to use mobile network systems within their applications
<richardschwerdtfe> kevin: the rest of the presentation is to reach out to operators
<richardschwerdtfe> kevin: we are stating we want to embrace open apis
<richardschwerdtfe> kevin: the response has been very positive but a lot more work to get apis into operators
<richardschwerdtfe> kevin: This ties us into the UWA ontology
<richardschwerdtfe> kevin: we have an api on the user as well
<richardschwerdtfe> kevin: this is basic delivery context information
<richardschwerdtfe> kevin: We need to make sure the GSMA uses the delivery context ontology
<richardschwerdtfe> kevin: I should take some questions
<richardschwerdtfe> Andy: it is not really a question but more a point. On the user parts of the ontology (whether tye are on a pay phone or not), this is missing in different standards which I am working on. I would like to work with you on this.
<richardschwerdtfe> Andy: if we could harmonize with what we are doing with ISO that would be great.
<richardschwerdtfe> kevin: that would be very welcome. Anything we could reuse would be great.
<richardschwerdtfe> Andy: this is an ISO SC36 proposal but it is not public yet
<richardschwerdtfe> Andy: the functional individual requirements will be out in September. I need to get approval for posting. If not I will send you via email
<richardschwerdtfe> dsr: are we looking at browser bindings
<richardschwerdtfe> kevin: not yet but we could in the future
<richardschwerdtfe> kevin: we may have multiple proprietary ways to get information like location
<richardschwerdtfe> dsr: w3c is looking to form a working group as is Apple
<richardschwerdtfe> dsr: we will need to try to track
<richardschwerdtfe> kevin: at vodaphone we have a co-chair on that group
<richardschwerdtfe> kevin: there would need to work on two different methods at that time and would need to see how the projects would work together
<richardschwerdtfe> dsr: 911 and E112 are proving for server-based location. It will be interesting to see if the browser vendors will collaborate on this
<richardschwerdtfe> dsr: if you are using time to arrival techniques .... you don't want to have to depend on GPS
ToKevin: there is the ISO/IEC JTC1 SC36 Individualised Adaptability that will be an ISO standard Sept. 08 - its described in the presentation Rich and I did yesterday which will be posted to the list shortly. The other one is that we are doing work in ISO/IEC JTC1 SC36 on a Metadata Schema element that describes delivery context requirements that need to be associated with content, such as sreen size, audio codecs, frrame rate etc. - some derived from MPEG 21 wor
<richardschwerdtfe> kevin: if you have a non-gps handset ...
<richardschwerdtfe> dsr: can you develop a roadmap for how this impacts UWA?
<richardschwerdtfe> kevin: we use sharepoint which you could monitor to see what we are doing on apis
<richardschwerdtfe> kevin: we are looking to demonstrate at Makal in Asia
<richardschwerdtfe> kevin: the main cross-over is the data connection profile which will be restful
<richardschwerdtfe> kevin: this should make a simple piece of work - 2 methods
<richardschwerdtfe> kevin: if we are talking about web developers getting an easier route into mobile, we could use DISelect to define what is apporpriate for different devices
<richardschwerdtfe> kevin: the parser could exclude what is not relevant
<richardschwerdtfe> dsr: we were talking about the ontology before lunch with a last call draft in time for the technical plenary
<richardschwerdtfe> kevin: I don't see a big problem
<richardschwerdtfe> dsr: we are not going to have telecons for the sake of having telecons - there needs to be some clearly identified items
<dsr> scribe: dsr
<jcantera> http://www.w3.org/2007/uwa/wiki/Expression_language_for_accessing_the_delivery_context
Jose presents a proposal for a generic function to access any property value in the delivery context.
The function also could be bound to the properties and aspects in the DDR API.
Kevin: the presentation I just sent to the list covers prebuilt functions, but Jose's proposal looks good.
[we wait while Jose opens up Kevin's presentation]
Kevin: should we make this
generic and extensible rather than a fixed set of accessor
functions?
... I have to be on another call later, and hence want to ask
about DIAL now.
Proposes means to support XSLT based upon filtering nodes by ID or ROLE according to the delivery context.
Jose: I totally agree with the
value of that approach.
... I will be talking a little later about redefining DIAL to
support XHTML 1.1.
XHTML2 is slowing down the adoption of DIAL.
Kevin: we can't ignore the huge market experience with HTML and XHTML1. The dependency of DIAL on XHTML2 has been a problem.
Jose: we can borrow some specific features from XHTML2 and place them in our own namespace.
We can then progress DIAL to a CR/REC much faster.
Kevin: I agree with that and look forward to helping with that.
Jose offers to help Kevin with editing the DIAL spec.
Kevin: that would be great as I am very busy with the GSMA work. Being able to generate the spec from the schema sounds interesting.
Rhys used to have an XSLT for XML-Spec to XHTML, but that's the other way around.
<krcsmith> I support a Group Resolution to create a new namespace for DIAL and decouple from XHTML 2.0
Jose: I am very busy with the ontology and DIAL, when would you be able to help with work on the access functions?
Perhaps Rotan could help? The end of the year should be do-able.
<krcsmith> ACTION: Kevin to assess reuthoring of XPATH access functions to consider new generic syntax. [recorded in http://www.w3.org/2008/07/25-uwawg-minutes.html#action05]
<trackbot> Created ACTION-99 - Assess reuthoring of XPATH access functions to consider new generic syntax. [on Kevin Smith - due 2008-08-01].
<jcantera> dsr: evolution of DDR-API?
<jcantera> kevin: We support the DDR Simple API
<jcantera> dsr: We will discuss it in a future call
[slides to follow DIAL reexpression]
XHTML2 last draft is July 2006, and it looks as if it will take a long time to reach REC status.
DIAL is perhaps too complex to implement, e.g. requiring the full set of XForms actions.
DISelect can be emulated with XSLT.
The table module can cause problems for rendering on small screens.
Multicolumn or nested tables are difficult to render on mobiles.
Johannes: these may be valuable though on other devices with larger screens. The language should support richer UI than mobile.
Jose: what is the point of DIAL in that case?
Rich: asks Jose if the XHTML2 WG put out another draft, are you basically saying that you will ignore it?
Jose: if we want to move DIAL to
REC in the near future, we can't rely on XHTML2 being ready in
time.
... we could import some parts of XHTML2 e.g. section element
without delaying matters,
Rich: in my understanding the XHTML2 WG is looking for where they could add value. They have some really good people and a lot to offer for mobile. Perhaps they should merge with UWA?
Jose: I don't have a good feeling for the scope that the XHTML2 WG are working to.
Rich: XHTML2 hasn't worked out in the client and this is causing people to reconsider, e.g. to refocus on authoring framework rather than delivery formats.
Jose: we should discuss this further with the XHTML2 WG.
The idea isn't to drop XHTML2, but rather to move a device independent format forwards in a shorter time frame.
Jose (in response to a question from Rich) we have already included the XHTML role module.
Jose explains his proposal for DIAL modularization.
This includes so,e XHTML 1.1 modules plus extensions, some XHTML2 and modules, XForms UI and data module.
Rich: we would like consistency between mobile and desktop.
Dave: would like more details on the precise problems with XForms you want to avoid.
Rich: it would be good to align with ARIA.
Rich asks Jose to include tabindex.
Dave thinks there is a risk of misunderstanding with the XHTML2 and XForms WGs.
To avoid that we need to use the HCG to set expectations for what we are doing and plan for a joint meeting to review the proposal when it is ready.
<scribe> ACTION: jcantera to investigate adding tabindex into DIAL for use with any element and also adding values for zero and minus one. [recorded in http://www.w3.org/2008/07/25-uwawg-minutes.html#action06]
<trackbot> Created ACTION-100 - Investigate adding tabindex into DIAL for use with any element and also adding values for zero and minus one. [on Jose Manuel Cantera Fonseca - due 2008-08-01].
See HTML5 for more details.
Dave notes that this is a valuable addition and used it as part of HTML Slidy for LI elements on expand/contract of list items.
Rich: asks about SVG?
SVG was considered for DIAL but is dependent on progress on CDF.
Rich: how many companies are looking at DIAL?
Jose: Volantis implemented a subset of DIAL as XDIME2
Dave notes that some features of XHTML2 proved problematic, e.g. any element can be a link or replaced by linked content.
The proposed changes to DIAL would make it easier to implement in full.
Dave: could you expand on what the issue is with AVTs (including what they are).
AVTs are replaceable expressions within attributes in DIAL, and borrowed from XSLT.
Jose: I haven't looked at them in detail, but they seem to be hard to implement.
Dave wonders what we need to do to explain DIAL to a broader audience e.g. ordinary web developers who are very sceptical about XHTML2 and XForms.
Jose: we will update the primer after we have agreed the changes to DIAL.
Dave: can we come back to the tables question. Is DIAL a lowest common denominator format? Having to put tables in a conditional section adds extra effort to authoring.
Jose: DIAL combines abstract and concrete UI.
Dave: this touches upon authoring convenience, sometimes it is very convenient to include conditionals within markup, but in other cases, it is more convenient to use a separate XSLT style sheet to transform generic data markup into tables or whatever best suits the delivery context.
Jose: We decouple DISelect from DIAL and make it applicable to any XML language.
You could still use DISelect with DIAL when appropriate.
Dave: how then can you validate DIAL with DISelect?
Jose: that is impossible in general, but you can at least check it is well formed.
Dave asks Jose if he thinks we will have a sufficiently well developed proposal in time for joint discussions with the XHTML2 and XForms WG at TPAC08?
Jose: yes. Dave will look at the TPAC schedule to see what may be practical. We would need a proposal by early october.
<jcantera> we need some way to encourage participation
<jcantera> dsr: i don't like people not to have agenda items
<jcantera> jcantera: mailing list
<jcantera> dsr: someone for one area one responsibility who takes care of scheduling a conference call
<jcantera> dsr: e-mailing the chair when there will be a conference call
<jcantera> groups: Rotan or Kevin for convenient access functions
<jcantera> (proposal)
<jcantera> Kangchan and Wonsuk for Device Coordination
<jcantera> Jose for DIAL
<jcantera> Andy for Personalization
Jose for the ontology
Dave to harvest minutes for the roadmap for next 6 months
<andy> Personalization Roadmap
<andy> Andy and Rich prepare first draft of W3C note on the harmonization of Access For All personalization with Ubiquitous Web Activites by September 5.
<andy> Present to IMS and ISO related standard bodies by end of September and gain informal agreement
<andy> Develop First Draft of Device Context Ontology harmonized with Access For All preferences by January 30 09
<andy> http://isotc.iso.org/livelink/livelink.exe?func=ll&objId=1056984&objAction=browse&sort=name
<andy> http://isotc.iso.org/livelink/livelink.exe?func=ll&objId=806742&objAction=browse&sort=name
Face to face planning. Rich to see if Austin is a possibility for March 2009.
<jcantera> I will look into have a F2F by the end of June in Spain
We need to start thinking about re-chartering. Our current charter will expire on 31 March 2009.
Hopefully, we will by then have a clear idea of the scope for future work.
<scribe> ACTION: dsr to contact XHTML2 and XForms WGs on updated DIAL proposal and ideas for a joint profile. [recorded in http://www.w3.org/2008/07/25-uwawg-minutes.html#action07]
<trackbot> Created ACTION-101 - Contact XHTML2 and XForms WGs on updated DIAL proposal and ideas for a joint profile. [on Dave Raggett - due 2008-08-01].
.... end of meeting ...