See also: IRC log
<trackbot> Date: 06 April 2011
<cperey> sorry to be late
<scribe> Scribe: matt
-> http://www.w3.org/2010/POI/track/actions/open Open Actions
Andy: I thought it would be good to review actions.
<cperey> calling in over skype
ahill2: Christine and Matt talked about generating a summary of the F2F, did that happen?
cperey: I have been traveling.
matt: Me neither. I can take an action, or Christine?
cperey: It'd be useful!
<scribe> ACTION: Matt to write up summary of face to face meeting. [recorded in http://www.w3.org/2011/04/06-poiwg-minutes.html#action01]
<trackbot> Created ACTION-52 - Write up summary of face to face meeting. [on Matt Womer - due 2011-04-13].
ACTION-17?
<trackbot> ACTION-17 -- Karl Seiler to add samples to flesh out the Anchor definition in the data model -- due 2010-12-08 -- OPEN
<trackbot> http://www.w3.org/2010/POI/track/actions/17
karls: Not done, I need to catch up on the F2F first.
ACTION-18?
<trackbot> ACTION-18 -- Alex Hill to flesh out the 'extent' section of the data model -- due 2010-12-13 -- OPEN
<trackbot> http://www.w3.org/2010/POI/track/actions/18
ACTION-18 due next Wednesday
<trackbot> ACTION-18 Flesh out the 'extent' section of the data model due date now next Wednesday
ahill2: Also not done.
ACTION-23?
<trackbot> ACTION-23 -- Gary Gale to scrub wiki and terminology pages to reflect consensus notion of location and place terms -- due 2010-12-22 -- OPEN
<trackbot> http://www.w3.org/2010/POI/track/actions/23
ACTION-23 due next Wednesday
<trackbot> ACTION-23 Scrub wiki and terminology pages to reflect consensus notion of location and place terms due date now next Wednesday
action-25?
<trackbot> ACTION-25 -- Gary Gale to formalize relationships and create some use cases -- due 2011-01-19 -- OPEN
<trackbot> http://www.w3.org/2010/POI/track/actions/25
Andy: I know there were discussion of relationships at the f2f, were there use cases created around them?
-> http://www.w3.org/2011/03/31-poiwg-minutes.html#item09 Relationships Minutes from F2F
cperey: It was treated at the f2f on the third day. I would say close the item and create a new one that reflects the minutes.
karls: Does the same apply to the anchor and extent definitions?
cperey: We didn't treat those subjects at the f2f.
close ACTION-25
<trackbot> ACTION-25 Formalize relationships and create some use cases closed
-> http://www.w3.org/2010/POI/wiki/Core/Draft#Relationship_primitive Relationship Primitive in Core
matt: I think this is my duty as temporary editor to go through f2f minutes and make sure the draft reflects that now.
cperey: We had a great discussion on this, comparing it to OSM. At OSM they're not clustered, but can be sequenced, etc.
ahill2: I think we got a real sense of the difference between the hammer that is OSM and what people think of as POIs. How that turns into a reformulation of relationships is a TBD.
<scribe> ACTION: Matt to update core draft Relationships Primitive to reflect what we learned at the f2f [recorded in http://www.w3.org/2011/04/06-poiwg-minutes.html#action02]
<trackbot> Created ACTION-53 - Update core draft Relationships Primitive to reflect what we learned at the f2f [on Matt Womer - due 2011-04-13].
cperey: We have an opportunity to create relationships that don't break OSM's policies. Not saying how, but the knowledge from that subject is there.
ahill2: I concur that yes, I think one of the outcomes of this is that we develop the notion of relationships in POI that is sensitive to the types of relationships that OSM considers and that we can be clear on delineating differences there, rather than before where we would have just stumbled into it.
karls: Do we have a member of the group who has that kind of insight to OSM?
<cperey> proposed meeting to foster relationships with OSM commnity in September
<cperey> Jacques provided a tutorial on OSM on third day. Helped to frame the context o OSM
matt: I think Jacques is closest to the data. Plus we've got on the list of potential F2F locations as one that's co-located with the OSM meeting in September (also c o-located with OGC)
<cperey> good little primer and a lot of people better understand the situation with OSM
ahill2: On the third day we got a primer on OSM from Jacques. He would say it's simple, what else is there to say.
<cperey> we could ask him to provide the tutorial as an invited topic?
matt: With OSM we should be careful though, as we have different concerns. We can just update our database if we change something, we're distributed, etc.
Raj: Was there any talk of linked data and OSM?
ahill2: It's a single database, and people are building on top of it, but when we asked if people are linking across data, he didn't think that was something that happened.
... e.g. you might build an app on top of OSM that has a level of abstraction on top of it, but he didn't really have a good answer for linking between data bases and referring to IDs outside of the one OSM database.
-> http://linkedgeodata.org/About Linked Geo Data
matt: We did talk about linkedgeodata, which is extracting OSM data and making it linked data friendly.
ahill2: They also provide a SPARQL API.
matt: And a REST API.
Raj: So OSM is successful but people when building their apps are what, copying the data and building on top of it?
karls: How do you mean copying?
Raj: For instance, Google is supposedly using OSM data, but I doubt they're hitting OSM servers, they're downloading the data presumably.
... In other words apps aren't using it online, they're copying it and building apps on top.
ahill2: I think that's a fair description.
matt: Certainly on large apps.
karls: I can understand Jacques pause. It relates to ID management outside of the database.
... Our stuff will be going across many data sources. We got to unique URIs.
ahill2: One topic was how to engage the OSM community. It's organized in a different manner than say OGC.
... Jacques was encouraging us to get in the forums, etc. But that doesn't really give us someone to engage with, to invite into the group and collaborate with us.
<cperey> +1 to this recommendation/and the lack of someone to contribute
<scribe> ACTION: Andy to find someone within OSM to collaborate with [recorded in http://www.w3.org/2011/04/06-poiwg-minutes.html#action03]
<trackbot> Created ACTION-54 - Find someone within OSM to collaborate with [on Andrew Braun - due 2011-04-13].
<cperey> I'm just preparing the "how to invite experts to contribute to our domain knowledge" (one of my action items)
Raj: We need more people at the table, Facebook, Gowalla, etc. Talked with someone from Simple Geo, and he had problems with this. The barriers are high right now.
Andy: I talked to both Gowalla and Foursquare and I don't think either of them have the bandwidth to participate in creating the documents.
... Foursquare is interested in reviewing. As we get drafts it's worth sending it to them, but I don't think they have the bandwidth.
Raj: I think we can get over bandwidth hurdles if we can get real commitment and interest.
ahill2: We had some discussion of this at the F2F about participation.
-> http://www.w3.org/2011/03/29-poiwg-minutes.html#item05 Some participation discussion
ahill2: I've heard some motivation of being able to brand data as W3C compliant, but I can't imagine that's a strong motivating factor to smaller companies.
... We have to be sensitive to some of those motivations.
fons: I just had a discussion with someone at TomTom, and it seems that they are interested in at least participating in the WG. I tried to get them to become a W3C member first, but if they can't do that, then it might be interesting to have them in the WG as an invited member.
karls: From Nokia/NAVTEQ/MS, I think anyone in the map business is very interested in facilitating local search. The hinge point there is on POI definitions. We're all looking to build the largest most robust POI database.
... We need standards for that to ensure the widest participation. We're very interested in making this work.
Andy: I think there is lots of interest from big guys, but smaller ones don't have the same motivation to build the largest data set. Their stuff is community driven, etc.
karls: We might start with commercial data, but our end goal is to have community.
Raj: That's a great statement. I think we don't need every one of them, but I think we need more sides of the coin here. Otherwise, we'll keep going around in circles about the core data definition.
karls: Makes sense to me.
ahill2: Karl, can you elaborate a bit? You've got a proprietary database of POIs and you would like to have a product that is drawing POIs from other sources on the fly and that that would be a good motivation for a standard?
... Or do you want it easier for you to purchase POIs or acquire it that it comes in in a palatable way for integration into the database?
karls: There's an arc we are projecting. We're trying to build out platforms that we can put in a consortium space that is a single unified repository for places, locations and relationships. We're going to feed that content with anything we can license to get to critical mass as fast as possible -- you have to achieve a tipping point in volume for it to attract users -- but once that is hit, the intent is that it is open.
... We could extend the base records we have with licensed data for instance.
... But that last part is not what we should strive towards. A standard that drives us toward participation is what we need.
... We've got islands of information that we want to bridge.
Raj: That sounds like a great business use case.
... Something like that could drive the core data format. e.g. we could have where the location is in the standard, but maybe the open hours changing could be a proprietary extension.
karls: Could be. We really also want to facilitate in a wikipedia way the repair of the database.
... The standard needs to allow to happen or provide governance on change management.
... POI have much shorter shelf life than maps.
... Extensibility and flexibility are important too.
Raj: Change management to me requires unique IDs to me. What if change management came out of updating those URIs? Maybe it doesn't take a huge data model just a few key things.
karls: If we could figure that out it would be a real accelerant for the whole thing.
ahill2: Can you elaborate on that a bit Raj? A simple example?
Raj: We can't predict ahead of time everything that could go into a POI, but say we put 90% of our energy into the unique ID. A point location, a name, a start and end time. Then when people want to collaborate they know they are talking about the same object because they have a unique ID.
karls: We've talked about DNS IDs.
Raj: if the unique ID were in there, the rest could fall out from that.
ahill2: We had a lot of discussion with Thomas Wrobel and Dan Brickley about semantic entities like dbpedia, etc that are providing unique URIs to known things.
... e.g. a URI for a horse, cow, dog, etc. I took away from that that instead of a single registry there might be a couple, e.g. dbpedia, library of congress, etc.
... Each managed in some sense like wikpedia, etc. Then there's a more rigid model of DNS like. Which direction should we go?
Carsten: I think you can push the wider model open by saying you don't have a registry, just the Web itself. Lifecycle management is harder on the Web, but strong on DNS.
... I think this is the right direction: minimal data model, effort on unique ID and making sure the data can be linked up.
karls: I agree with everyone that the ID, any work we can do around that is the anchor. Rich location definition is important too. Relationships, being open ended, is important too.
... I don't think we can go far beyond ID, location, relationships and attributes.
Raj: In the DNS model if we had some governance people could get their unique ID for a POI and then it gets propagated around the world.
karls: That's what Navteq is building in their POI registry. They intend to not own that in the long haul, that it should go into a consortium controlled space.
ahill2: In DNS you get competition for different domains, etc. Can anyone comment on what they think on that?
... We want a model where it's not owned, right?
Raj: Right. Who owns DNS? ICANN? And Wikipedia is governed by an organization, and has a lot of governance. If you put a bad article up there they won't publish it.
karls: Content moderation is key, current and accurate is immensely valuable.
... I think the standard should support the basic construct of changes, owners, etc.
... Timestamps to know that it is current and accurate.
<cperey> +1 on that
<cperey> moving from a data-oriented discusion to a services oriented discussion
matt: We talked a lot at the f2f about talking in terms of HTTP and XML. I think we won't have to define lots of stuff around this, but provide guidance, e.g. "if POI data changes, when the URI is dereferenced, this HTTP header is produced to redirect to the updated one, cache tags are used, etc."
... This stuff we need to do to make it fit in the Web architecture needs to be figured out, not just the data.
Raj/karls: This sounds like we're moving from data defining to service defining.
karls: If we can't imagine that the standard doesn't facilitate the creation of a consortium that has governance over the quality, then I think we haven't hit the mark.
<cperey> I cannot extend
<cperey> no problem
-> http://www.w3.org/2010/POI/wiki/Face_to_Face_Meetings/2011_Future_Face_to_Face_Meetings Options for F2F
-> http://www.w3.org/2002/09/wbs/45386/POI-F2F-2011-2-choices/ F2F choices
matt: details options
cperey: In addition, there is an AR Standards meeting in Taiwan, on June ??. Karl will be there, and Martin, maybe Jonathan and Jim?
... My hope is that regardless of where it is, having a meeting hosted in conjunction with another event is highly desirable.
matt: I updated the f2f poll to reflect AR standard smeeting too.
cperey: When/where does the group want to work on the drafts, and when/where does the group want to interface with other groups and share?
Raj: Is there an option for four?
matt: Sure. We say 2-3 in the charter, but that's to estimate how much time people can spend, if the group wants to meet more, sure.
ahill2: It has been unclear up until now how the format will be used. This discussion about registry makes a lot of sense.
... I think the service idea is quite compelling.
<Carsten> Got to go as well, bye
matt: Maybe someone can take an action to look at current proposals for registry type stuff? And maybe another for how this stuff can fit in the Web architecture?
Raj: Sounds very open ended?
matt: Just starter stuff.
ahill2: Could someone report back on registries out there? I know they are known to people, but maybe a listing of various dbpedias, library of congress and other attempts to create registries of POIs? Maybe some commentary on policies? How they manage change? Centralized authority? Editorial board? Vote up/down?
karls: I'll take that action. I'm building out a content management group as we speak, and so have had some look at things similar to this.
<scribe> ACTION: karls to report to group on what community oriented registries exist and what their governance policies are [recorded in http://www.w3.org/2011/04/06-poiwg-minutes.html#action04]
<trackbot> Sorry, couldn't find user - karls
<scribe> ACTION: karl to report to group on what community oriented registries exist and what their governance policies are [recorded in http://www.w3.org/2011/04/06-poiwg-minutes.html#action05]
<trackbot> Created ACTION-55 - Report to group on what community oriented registries exist and what their governance policies are [on Karl Seiler - due 2011-04-13].
<scribe> ACTION: Raj to look into DNS based POI registry models due two weeks [recorded in http://www.w3.org/2011/04/06-poiwg-minutes.html#action06]
<trackbot> Created ACTION-56 - Look into DNS based POI registry models due two weeks [on Raj Singh - due 2011-04-13].
ahill2: We've talked about having experts come in to discuss things we don't understand fully. Creating a service, I don't have the experience for that. Maybe we can get an expert in for that?
<scribe> ACTION: matt to work on finding an expert to discuss with WG about service oriented details [recorded in http://www.w3.org/2011/04/06-poiwg-minutes.html#action07]
<trackbot> Created ACTION-57 - Work on finding an expert to discuss with WG about service oriented details [on Matt Womer - due 2011-04-13].
<scribe> ACTION: Karl to work with matt on ACTION-57 [recorded in http://www.w3.org/2011/04/06-poiwg-minutes.html#action08]
<trackbot> Created ACTION-58 - Work with matt on ACTION-57 [on Karl Seiler - due 2011-04-13].
karls: I'll help Matt with that one.
ahill2: I assume OSM is one of the names involved in this, at least for consideration.
matt: Sure.
-> http://www.w3.org/2010/POI/track/actions/open Open Actions
matt: Everyone please review action items! Close them if need be, or open new ones or whatever.
<scribe> ACTION: Alex to start email thread on possible new call time [recorded in http://www.w3.org/2011/04/06-poiwg-minutes.html#action09]
<trackbot> Created ACTION-59 - Start email thread on possible new call time [on Alex Hill - due 2011-04-13].
<karls> adios amigos
action-50?
<trackbot> ACTION-50 -- Christine Perey to determine which OGC WGs are meeting at OGC TC in June -- due 2011-04-07 -- OPEN
<trackbot> http://www.w3.org/2010/POI/track/actions/50
trackbot, end meeting