10 Mar 2006

See also: IRC log


karl, karlUshi




<Dave> Vincent Quint talking about content adaption in the ubiquitous web.

<Dave> q: your distributed architecture diagram - transformation servers are valuable

<Dave> or did you intend all the transformations to take place in the proxy?

<Dave> a: the proxy will be used only when there isn't adequate transformation on one or the other side.

<Dave> there may be more than one proxy and proxies may be specialized to different purposes.

<Kazuyuki> Dave, do you want to record the name of the commenter?

<Kazuyuki> ok

<Dave> q: (Johan) the device description problem is much high in the SIP community, what is the reason you only focus on HTTP?

<Dave> a: most clients are using http, but I am not claiming that all clients must use http, especially for streaming media.

<Reto> it is fine now

<Dave> Johan: SIP offers more negotiation possibilities than http and that expands the architecture.

<Dave> q: (Liddy) search for proxies to handle specific transformations is also important

<Dave> I am also concerned about privacy for user/device profiles.

<Dave> a: profiles can also be stored on the device and passed to the proxy upon demand.

<Dave> In answer to the first question, I mentioned about DI Select and other means to make the transformation sensitive to the context.

<Dave> q: (Don R) are you aware of any device profiles that you can use?

<Dave> a: the adaptation may not provide sufficient information via http.

<Dave> q: are there any public repositories of device information?

<Dave> a: yes there are some.

<Dave> Dave notes that the is an open source repository and W3C's MWI is working on the APIs and mechanisms to support repositories.

<Dave> q: doesn't the descriptions need to cover rules for what transformations are appropriate

<Dave> q: why do you need to negotiate, why not just send everything?

<Dave> a: it is too much

<Dave> Dave thanks Vincent

<Dave> (applause)

<Dave> Reto Krummenacher presents on triple spaces for a ubiquitous web of services.

<Dave> q: this is a distributed blackboard, right?

<Dave> a: yes, but we are combining with the Semantic Web as semantic databases.

<Dave> The sematic web will enable the combining of large amounts of distributed information.

<Dave> q: I am curious as to why it is called triple spaces?

<Dave> a: the first idea was to combine RDF and tuple space, RDF uses triples, hence triple space.

<Dave> alternative names include semantic space.

<Dave> q: (Franklin) in tuple spaces you can change values, is that the intent, to allow you to navigate and make changes?

<Dave> a: currently we publish RDF graphs and consuming them.

<Dave> (shows additional slide with write, read, take, query operations)

<Kazuyuki> Dave, I'm not sure why, but yesterday's log was: http://www.w3.org/2006/03/08-ubiweb-irc

<chaals> It's because RRSAgent gets the time based on GMT or Boston time or something, so if you live in Japan or Australia it gets confused about the date.

<Dave> Presentation from University of Deusto.

<karl> [http://www.w3.org/2006/03/08-ubiweb-minutes]

<Dave> q: (Johan) what about scalability, have you tested with more than a few nodes?

<Dave> a: we want to expand the range of scenarios to address that.

<Dave> the centralized nature of the orchestrator is also a weakness of the current approach that we would like to address

<Dave> q: (Franklin) have you tested for usability?

<Dave> a: not formally as yet.

<Dave> q: you showed us a diagram for time vs number of APs. can you explain it?

<Dave> a: constraint processing can take time due to its cost ...

<Dave> q: can you give me an example of reasoning for adaptation?

<Dave> a: a rule that knows that if you are in a given room then you are in a given building with a given location

<Dave> Johan Hjelm presents "building the ubiquitous applications environment"

<Dave> q: (Reto) is there a website for SPiCE?

<Dave> a: not yet

<Dave> q: (Franklin) your project is addressing a number of security related desires. But aren't there real usability problems for configuring that?

<Dave> a: we are using some widgets to simplify this, and one approach is based on sets and another is based upon recipe cards

<Dave> a third technique is based upon constructing a matrix

<Dave> people don't want comprehensive solutions, and prefer solutio

<Dave> solutions to particular problems instead.

<Dave> q: (Dave) can you show us an example of policy language for privacy management?

<Dave> a: not right now, see the website.

<Dave> q: for fuzzy properties have you consided machine learning for understanding the context (e.g. for nearness)?

<Dave> a: yes

<Dave> we are using bayesian techniques for this.

<karl> Scribe: karl

Break out session

Chaals: is the semantic web a too heavy approach to a context of small devices?
... The SW might be too hard or overkill.
... but in other way, it helps to maximize interoperability
... what is the middle ground?

dave: I think you have to go back to use cases to be able to find the relevant technologies.

Johan: One of the issues is that people add a lot of things when discussed about their desires. So you end up with a universal knowledge representation system.
... It's not a context representation problem. You should not start by representing everything, but more by representing small things available to you and then create your context.
... Users don't create ontologies. They want something which solves their problem.

<chaals> Karl: RDF is a way of representing things. People create folksonomies that are dynamic ontologies, but there is no value in a big framework that requires people to build their ontologies beforehand...

dave: there are people who make money from building ontologies. But yes we could have an approach were many people annotate things, give information

Johan: The issue is how to represent the user.
... The problem is how to find something which is relevant, or appropriate for the user which is middle ground.

dave: example of use cases

JOhan: Footbal games ?

<karlUshi> [12:33] <karl> dave: there are people who make money from building ontologies. But yes we could have an approach were many people annotate things, give information

<karlUshi> [12:34] <karl> Johan: The issue is how to represent the user.

<karlUshi> [12:34] <karl> ... The problem is how to find something which is relevant, or appropriate for the user which is middle ground.

<karlUshi> [12:35] <karl> dave: example of use cases

<karlUshi> [12:35] <karl> JOhan: Footbal games ?

<karlUshi> [12:35] * karl is missing the articulation of Johan

<karlUshi> [12:36] <karl> Johan: we transmitted files as CSV files because it was a lot faster than RDF files

<karlUshi> [12:37] <karl> SpeakerX: Use RDF when you need, or XML, or text. It really depends on the context of your use.

<karlUshi> dave: question_missed

dave: something interesting this morning
... was the notion of annotation for devices
... what is needed for annotations

Johan: it's not only about having to send your data, but also the status of interaction between devices. A negotiation future protocol should include this?

SpeakerA: yes

Johan: when negotiation is happening, there should be the possibility of dealing with multiple events, so a notification system

dave: multimodal context interface are dealing with this. Type of papers, quantity etc.

SpeakerA: The mechanism needs to be exposed

dave: that comes back to what kind of mechanism and languages you should use for that.

DonWright: All of these problems have been solved
... but not on a large scale basis

dave: web services?

DonWright: it's too hard to do. Not too hard in the sense in writing of the software, but too hard in terms of deployment.
... problems of firewall, problems of physical structures of companies
... etc.

dave: what about admin costs?

DonWright: many companies have proposed this kind of business but none has been successful
... because of the deployment issue

SpeakerA: The status of events is not always relevant.

dave: what kind of apis, web services, ontologies (common vocabularies) should be built to deal with this?

<UEDA> <break dur="45min"/>

<scribe> scribe: karl

Lower case ubiquitous web architecture and its implementations by Satoru Takagi

Satoru: establishment of ubiquitous computing environment as social features
... Ultra Tiny Computers everywhere: software and hardware
... Introduction of the hardware: T-Engine, micro T-Engine, P-Engine
... creation of Ubiquitous ID "ucode"
... UCR = UCode Relation (compatible with RDF) it's an URI
... it has properties like RDF
... Real ucode: in dice, active IR-tag, 2 dimensional bar-code like QR-code
... based on social infrastructure
... -> http://www.w3.org/2006/02/UbiquitousWebUNL.pdf
... Embeddability, scalability
... Demo of a concrete architecture
... User context information like CC/PP, FOAF
... lower case ubiquitous web platform configuration
... application of semantic web to GIS
... SVG Map with spatial metadata
... UCR-spacial metadata
... UCR-spacial network
... (DEMO live)
... using an XML Parser 30kb and RDF/XML Parser and Object Manager with Query Interface = 28 Kb

-> http://www.uidcenter.org/english/uidtags.html

-> http://www.uidcenter.org/english/uid.html


Franklin_Reynolds: I'm not sure I understand the difference with other IDs systems
... ucode:..... it can be treated as an URI.

karl: I put a link in the minutes to the FAQ. It seems it's a structured code.

Franklin_Reynolds: it seems to be a centralized system

<chaals> s/centralised/decentralised/

satoru: yes

dave: missedthequestion

Access Position Paper

by Yamakami Toshihiko

Toshihiko: Technology Provider since 1984 - 70 million of mobile users
... Access is specialized with Embeeded non pc network software
... software engineering in restricted environment
... participation to consortia
... presentation of use cases

<Kazuyuki> is afraid that this log has public permission...

[Kazuyuki the log is public, and there's no issue with that]

scribe: presentation of the issues and requirements
... Easy setup vs Visibility/controlability for users
... diverysity vs interoperability
... who will take the lead?


Johan: Question about Time Consuming Spiral slide
... question1

Toshihiko: Content of quality, integration ....

Miroslav: How the quality of devices impact your business?

Toshihiko: Technology progresses, more power, more space, we are specialist of restricted environments.
... Microsoft is a specialist of big environment but have difficult for tiny environment
... more power means our business will be more difficult

Tokuda: NTT Docomo controls everything?
... so how do you create the future if they control? How do you see it?

Toshihiko: We will just respond to our customers request wherever it is.
... Happy old days are gone.
... We have customers and they choose what they want.
... We have to be careful, what is the W3C role?
... we have to deal capabilities of devices, devices are very dynamic, so meta operation could be done at W3C

<Kazuyuki> Dave: what do you expect to W3C?

The Ubiquitous User Agent - Charles McCathieNevile and Hallvord R. M. Steen

Hallvord: The web for us is already ubiquitous.
... (demonstration of using the Web as it is right now)
... email, calendar, contact list, etc, interacting


Kanchan_Lee: is there an activity inside W3C like Mobile OK?

chaals: Yes and I'm one of the editors. It's a mobile best practices
... It's broad guidelines
... the next step will be mobile ok, if the best practices are successful.
... We like this from Opera perspective.
... if you do web sites which don't respect things it's harder.
... but if you do it right, it's benefitial for everyone.

karl: what about ubiquitous authoring tools?

chaals: it's very difficult.
... accessibility guidelines were a step. Authoring tools are crucial
... There are no authoring tools developpers at this workshop.
... It's very important.

dave: There are languages which are developed right now to input at W3C
... davequestion?

chaals: to avoid securities issues, you make a standardized UI, but you don't want to do that. That's the same for security.
... but right now each vendor is doing its own framework with not a lot of interop
... we would like to promote interop in security.
... The securities policies of browsers vendors are different
... even if the techniques are quite similar.
... we are ready to cooperate, and have been cooperating with other browser vendors already on this. But it's a complex work

dave: pop up messages are really an issue for all kind of devices.

chaals: yes it's an issue. We would like to have a framework. but a security model is not yet done.

Franklin_reynolds: Do you think that there should be a WG for basic security framework for the Web

chaals: It's one of the area, where I would like to see the W3C push an activity
... for now, there is SSL, but I don't know if I can trust what I get from the other end, or not.
... (answering a comment of Franklin Reynolds)

Event-Driven Coordination Rule of Web Services enabled Devices in Ubiquitous Environment - Kanchan Lee

-> http://eng.mic.go.kr/ The IT839 Strategy/

<karlUshi> scribe: karlUshi

Franklin_reynolds: I was not clear about WS_ECA. I didn't understand. What's the connection between events conditions and actions

Kanchan: ECA supports logic expression.
... all events come from Web services

Session of discussions

<karl> scribe, karl

<karl> scribe: karl

Dave: I would like we discuss about the notion of events in terms of Web services

dave and franklin reynolds have a discussion on distributed computing with events.

Dave: Distributed applications would be easier to write with simple markup

Johan: We have been hearing a lot of things in this workshop
... which are different.
... If the W3C should do something, what should be addressed.
... We haven't scratched enough the surface to know what we want to do.
... It's a very interesting workshop. But I would be concerned to start an activity on this.
... We don't have enough use cases right now with practical problems to solve.
... research, industrial developments, etc have been presented.
... it's not time to start an activity
... at least until we have defined problems

dave: W3C is already working on many of these activities
... it's more about knowing how all these pieces fit together.

Johan: There is a lot of work going on everywhere not only at W3C.

Chaals: I have the same work assumptions than Johan.
... Is the problem clearly defined?
... There are many groups at W3C covering already some of the work.
... If there are 4 groups doing the same thing, We can't allocate resources.
... and we might not participate
... What I would like to see from this workshop
... is that to discuss with WGs to push them in the right direction.

Dave: something specific?

chaals: CDF crossover with multimodal, API groups, etc.
... There are already groups which have already deliverables on these topics
... We don't want to have more and more Mobile focused groups
... SVG do the right thing somehow. There is no special WG for mobile
... but they care of doing the things right for mobile.

Dave: it's the case for many WGs.

chaals: To an extent yes
... Don't make more mobile groups
... make more technology which are mobile friendly.

Johan: Good problems definition helps a lot in this case.
... Activities which don't have a good problem definitions, for example, Mobile Best Practices, and Device Independence and Device description WG.
... some WGs at W3C have overlapping agendas.

Don: To some extent, I will agree with Johan and chaals
... we spent a lot of time on mobile but not on non traditional devices.

<karlLapUshi> I dropped again

<karlLapUshi> Karl: maybe we should try to coordinate more inside W3C, maybe an outcome of Workshop would be to create a big picture, identify technologies

<karlLapUshi> ... and their interactions.

<karlLapUshi> Retallack: Is there one world wide web?

<karlLapUshi> chaals: one of the thing we didn't the network info-structure

<karlLapUshi> ... we build browsers on many different info-structures: phone, web, aircraft, etc.

<karlLapUshi> ... In the mountains, there is the satellite phone.

<karlLapUshi> ... People in some regions will share phones

<karlLapUshi> ... Old machines will still work on.

<karlLapUshi> ... it is important, W3C has made an important step in that direction.

<karlLapUshi> ... Africa is one of the place where W3C had no meetings yet, but there will be a Web API meeting there later this year

<karlLapUshi> ... There is a lot of work which should be done by W3C, maybe not everything

<karlLapUshi> ... like sometimes phone protocols

<karlLapUshi> ... so scoping the appropriate domain.

<karlLapUshi> ... is important

<karlLapUshi> Johan: The requirements are very high in this workshop compared to places in the world like Bangladesh.

<karlLapUshi> ... There are many people with illiteracy but they are able to use the technology.

<karlLapUshi> ... It's part of our responsibility to take care of these countries.

<karlLapUshi> ... and these people too.

<karlLapUshi> Liddy: I just come back from Cambodia where a wireless motorbike is passing get data and bring back answers later on.

<chaals> Johan: The idea of web interaction without people involved - applications working with each other, is an interesting area of development, which could have some well-defined problems

<karlLapUshi> ... telemedicine is working that way there.

<chaals> [liddy's example of bandwidth moving at 50mph, literally... ]

<karlLapUshi> Liddy: The value of the web is the information and the ability to transmit information from one place to the other for saving someone life is pretty cool.

<karlLapUshi> Franklin: Filters of events are very important

<karlLapUshi> Dave: multimodal group is working on that.

<karlLapUshi> Franklin: service discoveries, we haven't talked about them.

<karlLapUshi> ... Standardize vocabularies and structures

<karlLapUshi> ... Names (ids) are fragile when we move.

<karlLapUshi> ... It would be great to solve this issue

wrapping up

<karlLapUshi> (applause)

<karlLapUshi> ADJOURNED

Summary of Action Items

[End of minutes]

Minutes formatted by David Booth's scribe.perl version 1.127 (CVS log)
$Date: 2006/03/10 08:11:20 $

Scribe.perl diagnostic output

[Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.127  of Date: 2005/08/16 15:12:03  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Succeeded: s/GMT/GMT or Boston time or something/
Succeeded: s/spatial/spacial/
FAILED: s/centralised/decentralised/
Succeeded: s/SpeakerB/Miroslav/
Succeeded: s/develop/developed/
Succeeded: s/it or not/what I get from the other end, or not/
Succeeded: s/about to cooperate/ready to cooperate, and have been cooperating with other browser vendors already/
Succeeded: s/meetings yet/meetings yet, but there will be a Web API meeting there later this year/
Found Scribe: karl
Inferring ScribeNick: karl
Found Scribe: karl
Inferring ScribeNick: karl
Found Scribe: karlUshi
Inferring ScribeNick: karlUshi
Found Scribe: karl
Inferring ScribeNick: karl
Scribes: karl, karlUshi
ScribeNicks: karl, karlUshi

WARNING: No "Present: ... " found!
Possibly Present: Dave Don DonWright Franklin Franklin_Reynolds Hallvord Hirotaka JOhan Kanchan Kanchan_Lee Kangchan Kazuyuki Liddy Miroslav Retallack Reto Reto_ Satoru SpeakerA Tatsuki Tokuda Toshihiko UEDA chaals chibao chibao_ hallvors herve herve_ karl karlLapUshi karlUshi mimasa quint quint_ sakakura
You can indicate people for the Present list like this:
        <dbooth> Present: dbooth jonathan mary
        <dbooth> Present+ amy

WARNING: No meeting title found!
You should specify the meeting title like this:
<dbooth> Meeting: Weekly Baking Club Meeting

WARNING: No meeting chair found!
You should specify the meeting chair like this:
<dbooth> Chair: dbooth

Got date from IRC log name: 10 Mar 2006
Guessing minutes URL: http://www.w3.org/2006/03/10-ubiweb-minutes.html
People with action items: 

[End of scribe.perl diagnostic output]