W3C home > Mailing lists > Public > public-poiwg@w3.org > November 2010

[minutes] POI WG minutes for 10 Nov 2010

From: Matt Womer <mdw@w3.org>
Date: Tue, 16 Nov 2010 14:00:57 -0500
Message-Id: <55A66455-A6A3-4E15-A990-037F799C6DE3@w3.org>
To: public-poiwg@w3.org
Hi all,

The minutes for last week's teleconference are here:

And as text below.

My apologies for getting these out so late.  I was hoping to sort out who was present, as you can see there were a few unidentified callers.  Please contact me off-list if you're not listed, or own one of the obfuscated numbers listed.

The action items were:
ACTION: jacques to collect triples on the mailing list [recorded in http://www.w3.org/2010/11/10-poiwg-minutes.html#action02]
ACTION: mliebhold to put out a definition of a 'thing' to the mailing list. [recorded in http://www.w3.org/2010/11/10-poiwg-minutes.html#action01]



      [1] http://www.w3.org/

                               - DRAFT -

            Points of Interest Working Group Teleconference

10 Nov 2010


      [2] http://lists.w3.org/Archives/Member/member-poiwg/2010Nov/0009

   See also: [3]IRC log

      [3] http://www.w3.org/2010/11/10-poiwg-irc


          +3539149aaaa, +39.011.228.aabb, marengo, +31.30.230.aacc,
          jdesmit, JonathanJ, Matt, +1.650.464.aadd, vinod,
          Mike_Liebhold, +1.919.599.aaee, Andy, +44.750.800.aaff, Gary,
          +1.617.848.aagg, cperey, +1.617.764.aahh, +1.773.575.aaii,
          Alex, Raj, +1.347.661.aajj, Carl, +1.312.894.aakk, k,
          +1.617.642.aall, Marco_Marengo




     * [4]Topics
         1. [5]Agenda
         2. [6]Discussion framework
         3. [7]POI data format
         4. [8]Use Cases
         5. [9]AOB
     * [10]Summary of Action Items

   <trackbot> Date: 10 November 2010

   <scribe> Scribe: Matt

   <cperey> np


   <ggale> There's 3 open actions according to

     [11] http://www.w3.org/2010/POI/track/actions

   andy: Christine, thanks for editing the agenda. We'll put stuff
   there, try to do it two weeks out. I was thinking this might be a
   failed experience, but we can keep it going a little bit.


   <trackbot> ACTION-9 -- Karl Seiler to add a persona/perspective type
   field to use cases. -- due 2010-11-03 -- OPEN

   <trackbot> [12]http://www.w3.org/2010/POI/track/actions/9

     [12] http://www.w3.org/2010/POI/track/actions/9

   close ACTION-9

   <trackbot> ACTION-9 Add a persona/perspective type field to use
   cases. closed


   <trackbot> ACTION-10 -- Gary Gale to codify on the WIki the "What is
   an Point of Interest" discussion -- due 2010-11-03 -- OPEN

   <trackbot> [13]http://www.w3.org/2010/POI/track/actions/10

     [13] http://www.w3.org/2010/POI/track/actions/10

   <cperey> so Action10 is closed?

   <k> am in the call

   andy: As a group, we'll see what we can put there and see if we can
   live with it going forward. Let's look at Gary's definition here,
   discuss it and then give it a week and see if we can live with it.

   <k> k = k seiler

   <jdesmit> +1

   gary: Fine.

   close ACTION-10

   <trackbot> ACTION-10 Codify on the WIki the "What is an Point of
   Interest" discussion closed

   <k> karl seiler joined

   andy: [Reviews agenda...]

   <cperey> +1

   <mliebhold> +1

   andy: Agenda good, let's move forward.

Discussion framework

   <jacques> +1.347.661.aajj is jacques

   <cperey> [14]http://www.w3.org/2010/POI/wiki/Main_Page

     [14] http://www.w3.org/2010/POI/wiki/Main_Page

   cperey: My suggestion because I've been watching the traffic on the
   list is that sometimes we change topics in the threads and it gets a
   little harder to follow.
   ... I was hoping you could take a look at the wiki page, and before
   Karl or someone added the definition of POI to the main page, there
   was only the very top and bottom portions.
   ... We have a lot of topics floating in and out, I would like to
   capture them on the wiki page appropriately.
   ... The topic that seems messiest is the data model one.
   ... I think we should try to keep the discussion on the list to one
   of these four topics.
   ... If they're not sufficient, create a new topic and page to
   correspond to it on the wiki page.

   <ggale> +q

   ggale: I agree, it would be lovely to have this a little more
   structured. Are we not at risk of keeping the wiki beautiful, rather
   than digging into the nuts and bolts.
   ... The email is a better medium for discussing things. Then maybe
   each week someone volunteers to clean up the topic areas.

   cperey: I think that's fantastic and agree.
   ... I just wanted a discussion framework.

   matt: I was thinking the front page would be more like our home
   ... if there's lots of topics, I wouldn't want it to be obstructing
   the rest of the info on the page.

   <ggale> If the wiki gets more than 2 screenfuls on the home page
   then it needs tidying

   cperey: Yeah, I don't like Resources as a name. And I liked upcoming

   <ggale> We're verging on that now


   matt: I think we all like it, let's move the discussion offline.

POI data format

   andy: Temporal POI was the first sub topic.
   ... POI associated with refugee camps moving for instance.

   <k> +q

   <ggale> +q

   karl: Going back to ggale's definition of POI, if we can separate
   location from the definition... I think it'd be great for it to have
   a start and end.

   <ggale> "POIs have the shelf life of bananas" ... quote of the day
   so far

   karl: Then there are temporal attributes of POIs. There's ones that
   exist at a time and cease, and another temporal aspect of "movies
   are from this time to this time", the theater doesn't end.

   ggale: Aren't we in a chicken/egg situation. If we're trying to
   decide on a data format for something that we haven't decided on a
   definition of what it is.

   mliebhold: I have a suggestion: it's easy to make things way more
   complex early on and then spend months digging out from premature
   complexity. It's either information about a place or a thing.

   <ggale> +q

   mliebhold: I think for information about a place, there's only one
   characteristic that's critical: the degree of accuracy of the x,y,z
   coordinates. A POI could be an ornament on the build, which would
   need millimeter accuracy. For a thing, place isn't so relevant, and
   the only way I know how to do that is through a marker or trigger, a
   visual search, so a URI pointing to this. Everything else can be
   handled in RDF and triples handling any metadata they need.

   ggale: I agree that we can make this arbitrarily complex then spend
   months digging ourselves out of the hole.

   <jdesmit> +q

   ggale: Maybe I can ask if we're happy with the three tier model I
   put on the wiki, and if so, have others flesh each out on the wiki.
   ... Then let's see if we can use existing codifications/standards.
   ... See if we can merge them back into a single entity.
   ... Then check the use cases, see if we can cover those.

   karl: I haven't had a chance to look at the wiki yet.
   ... I'm working on the terminology page. Then we'll try to make the
   vocabulary consistent. This goes back to defining a model that's
   grounded in existing standards and definitions. If you think I'm
   introducing complexity let me know.

   jdesmit: I like your approach, and I like Mike's comment about a
   ... ?? place definition on the wiki ??

   karl: Let's write up the place definition on the mailing list and
   get people to make edits on the wiki.

   <cperey> good proposal

   Karl: Have three or four people drive the definition down.

   ggale: I'd rather have the debate on the mailing list.

   <k> matt that last was Karl S

   ggale: rather than the wiki

   <jdesmit> +1

   <ahill2> +1 on mail over wiki


   <k> so what is the w3c method to form work teams

   <cperey> thing (cute name)

   <cperey> an object?

   <k> i volunteer to be on the POI detailed definition work team

   <scribe> ACTION: mliebhold to put out a definition of a 'thing' to
   the mailing list. [recorded in

   <trackbot> Created ACTION-13 - Put out a definition of a 'thing' to
   the mailing list. [on Mike Liebhold - due 2010-11-17].

   <cperey> another action was proposed: to break up into teams

   andy: I heard a suggestion that we work in smaller teams.

   Carl: I think we should get the definition framework, then break up
   into teams to further refine the definitions, then work on the

   <jdesmit> +1

   Andy: Sounds like a good approach to me.


   Raj: I have a comment on the data model discussion before we leave
   ... Looking at the use cases, they seem to be largely around search
   and imply a services framework, yet we're only defining in our
   discussion an information model. How do we see this information
   model being used? Are services harvesting around the internet and
   then we do our own queries or...?

   <k> lost him?

   ggale: Sounds like the question Raj was asking was: are we defining
   the reference model or defining the services that expose the
   reference model? Or are we building the reference model.

   karl: I think the root of it all, the value, is for information
   exchange between systems.
   ... That is fundamental.

   <ahill2> +1 looking for a data exchange model

   ggale: I see us looking at how this is exchange rather than building

   <jdesmit> agreed. services are what we use eventually, but I think
   data exchange format should be the focus

   <ahill2> +q

   andy: We don't want to be in the business of building services.
   We're looking at building something that exists today. What use
   cases do we need to satisfy? But I don't think we need to be in the
   business of defining how the sevices work.

   <k> +1

   <cperey> services development are how companies differentiate

   ahill2: Those of you with more experience building services around
   data models, can anyone comment on, if what we want to do is build a
   data exchange model, is there advice on what to look out for? Or can
   we just ignore services when we're constructing a data model?

   Carl: If we wrap ourselves around the services that make the data
   model sing, we'll be wrapped around the axel.

   ahill2: That sounds like a reasonable suggestion.

   <jdesmit> +1

   <k> Karl seiler will intro himself as "seiler"

   jacques: Services will be useful to test the data model, but
   ... I want to make a list of triples in XML or RDF that could be
   useful, they're on the mailing list.

   <ahill2> +1 for jacques action item

   <cperey> jacques volunteered Action Item to write up some triples
   according to suggestion on the mialing list

   <jacques> ok

   <scribe> ACTION: jacques to collect triples on the mailing list
   [recorded in

   <trackbot> Created ACTION-14 - Collect triples on the mailing list
   [on Jacques Lemordant - due 2010-11-17].

   Carl: It's a worthy effort, perhaps cart before the horse. There's a
   lot of standards in this area.

   <cperey> a new thread in discussion framework


     [17] http://www.w3.org/2010/POI/wiki/Related_Specifications

   <ahill2> +q

   ahill2: As we're fleshing this out, could we have examples


   <k> seiler - good idea

   <jdesmit> +1

   <cperey> +1 to Alex suggestion to have examples

   Carl: Seems to me that until we have a data model or agreed
   definition of POI, then evaluating specific encodings is premature.

   <jacques> +1 to examples

   matt: Maybe we do this after we have data model and use cases in

   ahill2: When it does happen, I'd vote for having things inline,
   rather than separated out.
   ... There's a tendency to have definitions that lead other places.

   jdesmit: As we try to agree or converge on a data model then let's
   introduce examples.

   ahill2: What we really want to do is hash out the data model first
   and get ourselves happy with that before we move on.

   Karl: Isn't the data model work bounded by the definitions work?

   ahill2: I maybe meant definitions. Definition of POI and the four
   tiers that have been mentioned.
   ... The three original ones and 'thing' that has been suggested.

   <cperey> +1 on the discussion of order of focus but, isn't this
   going to be an iterative process?

   matt: So, we'll work on definition of POI, data model, then use
   cases, then standards.

   rsingh2: I'd like to make a radical proposal that a POI is just a
   location and time. We're not going to be able to agree on the other
   things for now.
   ... I can see people having databases of a few thousand POIs in
   these use cases. And then most of the sophisticated stuff could be
   done in the client.
   ... A definition that has more than location, geography and a start
   and end time is going to have us going in circles.

   Carl: I just added a third definition of POI that the IETF uses.
   ... So, it comes down to location and time, everything else are
   optional properties.

   <mliebhold> +Q

   <ahill2> -q

   <jacques> location, time and content

   <cperey> I'm going to drop off, will dial back in

   <ahill2> +q

   ahill2: When we say UUID, are we talking about the location, or the
   ... Is it the place or the location?

   karl: I think it's a POI. I think the definition of places, which
   are often bounded by polygons, can be prescriptively defined.
   There's a lot of standards we can comply with or meld.
   ... Same thing with addresses.
   ... I think we can define location, there's a lot of definitions for
   that, places, polygon structures, then... our definition of POI
   should be simple, very open ended. I like the IETF definition.

   <mliebhold> +q

   <ahill2> -q

   mliebhold: I sent mail, we can start with them associated only with
   location, but you can frankly have a POI attached to a physical
   object. We might postpone that discussion, but we'll have to get to

   <jdesmit> +1 to mike

   <ahill2> +1 4 dimensions + POI

Use Cases


     [18] http://www.w3.org/2010/POI/wiki/Use_Cases

   <mliebhold> +Q

   ahill2: Whenever the subject of visual recognition and temporal
   moving objects comes up, I would make the initial suggest that I'm
   not sure the data model needs to necessarily... location can be
   transient... moving objects, or things, be folded into location, so
   that we don't have to keep track of whether something is a moving
   object or distinguish it from a fixed place

   <jacques> + for tracker information in the data model

   mliebhold: That would require every object to disclose it's
   location. That's probably a burden that isn't appropriate for
   everything. A POI as a loaf of bread.

   ahill2: I think we should take this to email.'

   rsingh2: I don't think it's an edge case. There's things like lunch
   trucks that serve every day or hour serving different people.

   ahill2: Not an edge case necessarily, but from a practical
   standpoint how to fold it into the definition.

   <ahill2> -q

   <ggale> I need to sign off ... next meeting in the day job calls

   matt: On use cases, any volunteers for this? Particularly checkin

   <jacques> yes

   <k> thx

   <cperey> can I make announcement (re: upcoming meetings)?

   karl: I've been doing the bulk of the use cases, don't want to
   dominate them.

   matt: I'll send out a call for more use cases.


   cperey: I've got a meeting for AR standards -- not just POI -- at
   MWC. The AR standards community members will be 17 Feb starting in
   the afternoon and all day 18 Feb and ending at noon on the 19th.
   ... We have a host, and I wanted to make sure you have the details.
   Those who are interested should join us.
   ... Another meeting proposed at UPC, the campus used for W3C AR. I
   won't be using that space, but it's also available during mid Feb
   ... And there are OGC meetings happening world wide, and they want
   to post meeting announcements as well.

Summary of Action Items

   [NEW] ACTION: jacques to collect triples on the mailing list
   [recorded in
   [NEW] ACTION: mliebhold to put out a definition of a 'thing' to the
   mailing list. [recorded in

   [End of minutes]

    Minutes formatted by David Booth's [21]scribe.perl version 1.135
    ([22]CVS log)
    $Date: 2010/11/10 17:31:08 $

     [21] http://dev.w3.org/cvsweb/~checkout~/2002/scribe/scribedoc.htm
     [22] http://dev.w3.org/cvsweb/2002/scribe/

Scribe.perl diagnostic output

   [Delete this section before finalizing the minutes.]
This is scribe.perl Revision: 1.135  of Date: 2009/03/02 03:52:20
Check for newer version at [23]http://dev.w3.org/cvsweb/~checkout~/2002

     [23] http://dev.w3.org/cvsweb/~checkout~/2002/scribe/

Guessing input format: RRSAgent_Text_Format (score 1.00)

Succeeded: s/??/Karl/
Found Scribe: Matt
Inferring ScribeNick: matt
Default Present: +3539149aaaa, +39.011.228.aabb, marengo, +31.30.230.aa
cc, jdesmit, JonathanJ, Matt, +1.650.464.aadd, vinod, Mike_Liebhold, +1
.919.599.aaee, Andy, +44.750.800.aaff, Gary, +1.617.848.aagg, cperey, +
1.617.764.aahh, +1.773.575.aaii, Alex, Raj, +1.347.661.aajj, Carl, +1.3
12.894.aakk, k, +1.617.642.aall
Present: +3539149aaaa +39.011.228.aabb marengo +31.30.230.aacc jdesmit
JonathanJ Matt +1.650.464.aadd vinod Mike_Liebhold +1.919.599.aaee Andy
 +44.750.800.aaff Gary +1.617.848.aagg cperey +1.617.764.aahh +1.773.57
5.aaii Alex Raj +1.347.661.aajj Carl +1.312.894.aakk k +1.617.642.aall
Agenda: [24]http://lists.w3.org/Archives/Member/member-poiwg/2010Nov/00

     [24] http://lists.w3.org/Archives/Member/member-poiwg/2010Nov/0009

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

Found Date: 10 Nov 2010
Guessing minutes URL: [25]http://www.w3.org/2010/11/10-poiwg-minutes.ht
People with action items: jacques mliebhold

     [25] http://www.w3.org/2010/11/10-poiwg-minutes.html

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

   End of [26]scribe.perl diagnostic output]

     [26] http://dev.w3.org/cvsweb/~checkout~/2002/scribe/scribedoc.htm
Received on Tuesday, 16 November 2010 19:01:01 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 16:48:26 UTC