<trackbot> Date: 09 March 2011

<karls> i did draft the category primitive, still generating representative examples today, will pub via email tonight

<karls> welcome

<matt> Agenda

<matt> Scribe: Alex

<karls> i have to exit after 1/2 H, sorry

ID Primitive

matt: lets do id talk first since Karl has to leave soon

<matt> ID Primitive

<matt> Alex: In the discussion it all sounds very reasonable, about an ID primitive. My only concern: does a URI go against the grain of efficiency, and against what people are thinking of with IDs.

<cperey> sorry mute myself

<cperey> +q

<matt> karls: URIs helps protect us from clashing IDs. The same model should be used for both places and locations though.

matt, I've got it

<karls> +1 to cperey

<Ronald> +1

christine: certainly other groups have have tackled this issue; seems a bit premature for the 5 of use to weigh in solidly

<cperey> definitely do not have the knowledge in this domain

<cperey> I know that for personal identity there are hundreds of people who work on it continuously

matt: the web is an existing system of decentralized system of identifiers and is working pretty well

+1 to caputre in wiki

<karls> so we need a lead or small to team to research ID systems to adopt or align with

<Luca> +1

<vinod> +1

<cperey> it's not in my scope +1

<matt> ACTION: Matt to put ID requirements in Wiki doc [recorded in http://www.w3.org/2011/03/09-poiwg-minutes.html#action01]

<trackbot> Created ACTION-31 - Put ID requirements in Wiki doc [on Matt Womer - due 2011-03-16].

matt: proposed an action item to add the results of this ID to the wiki

christine: I suggest (has worked in the past) to invite an expert into our meeting on these topics we don't have experience with

vinod: I have some resources here that I can contact who know quite a bit about URI

matt: I'll put together some resources from the current e-mail thread

<matt> ACTION: cperey to invite Henning after Matt has put ID requirements in the wiki [recorded in http://www.w3.org/2011/03/09-poiwg-minutes.html#action02]

<trackbot> Created ACTION-32 - Invite Henning after Matt has put ID requirements in the wiki [on Christine Perey - due 2011-03-16].

<cperey> we can have the two speakers during the same session so that they can hear one another

<matt> ACTION: vinod to invite DERI expert after Matt has put ID requirements in the wiki [recorded in http://www.w3.org/2011/03/09-poiwg-minutes.html#action03]

<trackbot> Created ACTION-33 - Invite DERI expert after Matt has put ID requirements in the wiki [on Vinod Hegde - due 2011-03-16].

christine: Vinod and I can then hand this document off to Henning and Vinod's contacts
... suggests we schedule these two expert opinions on the same day

<matt> Alex: We might need some consensus on where these ID are going, are they on every primitive?

alex: can Karl clarify if he think each primitive needs an ID?

karl: place certainly needs it, location can be overloaded so it might be nice to have an ID so it can be shared

<cperey> Category IDs is Henning's specialty

karl: category identifiers help make sure text descriptions refer to the same category

<cperey> the IETF has been working on this as part of a WG

<cperey> probably GeoPriv?

<Ronald> +1 to have the same id mechanism where ids are needed

karl: I would move that we make our IDs share common traits among the place, location and category IDs

christine: we should use somebody else's category ID

<cperey> do objects have IDs? what about the Internet of things?

christine: what about locations that change?

<matt> Alex: Is it possible that a location can have an ID whose values are changing?

<matt> Alex: Maybe this is in some ways indicating a source of location information, not that it is static?

<cperey> I dom't undertand

<matt> cperey: What good is it for it to be fixed in space if the ID is changing?

<Ronald> +q

<matt> Alex: I was suggesting the opposite, if the ID of a car or something is changing then it is maybe reflective of it's location changing.

<matt> cperey: That's an object

<matt> Alex: I agree it's different than a fixed place like a building, but I could imagine different authorities managing different facilities, e.g. one for a building and another for a GPS tracker.

<matt> cperey: Wouldn't that be the location primitive?

<matt> cperey: I thought an ID was fixed.

<matt> cperey: I thought with the ID I could share the object between apps.

<matt> karls: We're confusing that a place and object has an ID with the fact that the ?? has a ...

<matt> vinod: The URI could be the same but dereferrencing the URI would give you the changing data.

<matt> karls: Here's the use case: five shoe stores in the same shopping mall, each has it's own place URI, but they share a location ID.

<matt> karls: Then there's the case of a delivery truck that has a unique object ID and a location ID that renders where it is right now.

<matt> vinod: If a location ID is a URI, then the service can always update the data at the URI.

<matt> Ronald: I tend to agree with Alex that for sensors the ID of the location primitive makes sense.

<matt> Ronald: The location primitive itself was not meant to be only for fixed locations.

matt, I'll

get it

<matt> Alex: Christine, do you feel this discussion satisfies what she's looking for in a POI?

<karls> have to sign off, have a good day

<matt> cperey: I'll have to see it in writing, but my sense is that there are sub IDs

<matt> Alex: One ID is about the POI, one is about the location (e.g. a sensor, a source, an authority)

<matt> cperey: So they are nested.

<matt> Alex: Yes. I was asking Karl about that.. There will still be this primary ID that is the POI id.

matt: OK, moving on to the next topic

vinod, how do you spell your institution/

AR Landscape

<vinod> Digital Enterprise Research Institute

thanks

<matt> AR Landscape Browsers

<cperey> first I've looked at it

matt: Jonathan has added some info about the AR browsers that are out there

<cperey> review it now

can we get an overview of what was added?

jonathan: can we move this to the last topic?

Administrative stuff

<matt> Next F2F Registration

<cperey> Matt, I have an item to add to admin topics

matt: the registration form is up, please fill it out whether attending or not - hotel, days attending

<matt> F2F meeting page

matt: reminder 29th-31st of March, agenda is entirely TBD

<cperey> for the agenda, is it appropriate to call for papers?

<cperey> OK

cperey: is a call for papers appropriate?

<cperey> +q

matt: like in December, we will keep this focused on a working meeting

<matt> DST reminder

matt: US & Canada moving clocks forward on the 13th and then Europe 2 weeks later
... for the next 2 weeks the meeting is 1 hour different

cprerey: is it 5 or 7 horus now?

cperey; right now we are 6 hours apart and for those two weeks we are only 5 hours apart - so1 hours earlier for Europe

<cperey> +q

cperey: in Barcelona there was the AR Standards Meeting but nobody from the W3C meeting was there
... we didn't have any contribution from the POI group
... the next AR Standards Meeting is likely to be between 12th and 18th of June
... in Taiwan

cprerey: I am making an invitation to the POI group to present their work at this meeting

+1 to move topic to next week

matt: sorry for mangling the link, lets move this topic to next week

+q

Linked Data

<matt> Linked Data discussion starts here

vinod: in the PDFs I attached I am trying to highlight the kind of complex queries that are possible with linked data
... we need to have semantics and structure to handle these complex queries

<matt> Vinod's advantages to Linked Data use in POI

vinod: the linked data allows extended data

+q

<matt> Alex: We've been trying to address extending the POI data, and so I want to make sure we're talking about the same thing.

<matt> Alex: Is this the same thing, or does linked data have greater extensibility?

<matt> vinod: Having a standard structure so that queries can be made. I'm talking about the content, rather than the delivery.

<matt> Alex: When you say structure, I'm thinking you can take on any data to an XML representation. Are you saying that isn't structured because we don't have a structure or schema and so then you can't do complex queries on it?

<matt> vinod: Correct, that's what I'm saying.

<matt> vinod: If you consider a person as a POI, he has different attributes than an historical POI.

<matt> vinod: If complex queries have to be supported then there must be semantics.

<matt> Alex: Not arguing against linked data, but trying to understand what it gives. Can't you add data that comes along with it's own schema through XML and namespaces? As long as the fundamental unit, a POI is simple enough, then can't the namespace facilitate that?

<matt> vinod: You can do some conversions, but linked data makes complex queries possible. What tools we use, e.g. relational database to RDF.

<matt> Alex: I tend to agree that linked data helps you query sources for additional information, that's the power of it.

<matt> vinod: If you don't have a standard structure then there are different names for attributes, etc.

<matt> Alex: So, through things like an XML schema with a namespace, can't you get some of this functionality out of XML?

<matt> vinod: There's ontologies, representing data in an object.

<matt> Alex: This process can occur without linked data though, right?

<matt> vinod: If we have a standard ontology adopted, we can support complex queries.

<matt> Alex: For those of us who aren't familiar with linked data, when I looked at the linked data representations, they look kind of complex. What is the most simple POI case?

<matt>

<matt> vinod: If you consider a person, you represent him in the FOAF ontology and then you can consider him in an AR context by adding POI attributes.

<matt> Alex: Basically linking up data with location as indicated in fifth slide.

<matt> vinod: Anything could be a POI, but we could have AR metadata outside of the POI core.

<matt> vinod: The AR attributes are for rendering in an AR browser.

<matt> Alex: It seems to me that it might be useful for the group to have an example that is similar to the use cases that we're used to.

<matt> Alex: e.g. the fast food restaurant in this municipality in this mall is a restaurant.

<matt> Alex: Contrast it to a typical XML representation or an RDB format.

<matt> ACTION: vinod to represent one of our common use cases in linked data friendly way [recorded in http://www.w3.org/2011/03/09-poiwg-minutes.html#action04]

<trackbot> Created ACTION-34 - Represent one of our common use cases in linked data friendly way [on Vinod Hegde - due 2011-03-16].

+1 to closing topic

Hotel reservations

<matt> Alex: Did anyone get a response from the hotel email?

<matt> Ronald: The hotel has received some reservations.

<cperey> bye

<Ronald> bye

<JonathanJ> bye

This is scribe.perl Revision: 1.135  of Date: 2009/03/02 03:52:20  
Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Succeeded: s/RDB/relational database/
Found Scribe: Alex
Inferring ScribeNick: Alex
Present: Vinod Luca

WARNING: Fewer than 3 people found for Present list!

Regrets: Jens
Agenda: http://lists.w3.org/Archives/Member/member-poiwg/2011Mar/0004
Found Date: 09 Mar 2011
Guessing minutes URL: http://www.w3.org/2011/03/09-poiwg-minutes.html
People with action items: cperey matt vinod

WARNING: Input appears to use implicit continuation lines.
You may need the "-implicitContinuations" option.