See also: IRC log
<matt> scribe:matt
<trackbot> Date: 12 January 2011
cperey: I'd like to work on a position paper for the AR standards meeting.
Andy: We're driving to release a draft this month, I think even with a small quorum we'd be best served to work on the tasks of the group. I misread your mail, thinking you wanted to know thoughts about a paper, rather than the content of the paper.
cperey: I wanted to capture some thoughts real time. Just a few pages.
... Just enough to have a meaningful discussion at the meeting.
Andy: I'd like to spend some time working on the core things, maybe we can look as we look at the core draft we can ask which things would be good for a position paper.
... For example, if the relationship primitive is good for an AR standards paper.
... As we go through each of these, the charter for this group is to do the POI work, but also look at the various AR pieces. With or without another standards workshop coming up, we might drift off that a bit, so it's good to keep in mind.
gary: I think we need to keep working towards the draft document. Everything else comes after it. If we have time to distill out the wisdom of the group for the position paper, then that's great, but we should be working out the final details.
cperey: The position paper isn't for me, it's on behalf of the POI WG. I'm proposing to pitch in, but it's not my position paper. It's a paper from the group to stimulate the discussion at a multi-organizational level.
gary: That's fine, but my comment still stands. Have to have a draft document in order to work on the position paper. I still think the draft Recommendation has to take precedence.
Alex: I'd agree with the sentiment that we need to keep momentum going and that's what we should be doing.
http://www.w3.org/2010/POI/wiki/Core/Draft#Relationship_primitive
Andy: This had a lot of discussion at the f2f, but not a whole lot in the core draft.
gary: We've talked about how one POI can have lots of relationships with others.
... We need to determine the canonical set of those relationships, and what they mean.
... We've got "contains", "contained-within", "adjacent-to".
... We could make things a lot richer and more interesting if we include contained-by and contained-within relationships.
alex: What about inherits-from?
... There seems to be a concern about organizational relationships, why not be explicit in the relationship?
gary: For every hierarchical system you have that works well within one market, there are an equal number of cases where it just falls apart.
... The inheritance side is worth looking at, maybe use cases. See if it gives us the flexibility to say to the world at large: here is the POI WG view and their relationships, and have people pick it up and see if it works in their implementation.
alex: Are you suggesting something concrete?
gary: We need to make the draft Recommendation as concrete as we can, but not constrain it to the way in which it could potentially be used.
... I had a paper come over my desk talking about the need for standards in social media companies. People bemoan having standards that aren't being used, so we need to be flexible enough to make sure it works on a global basis.
<Andy> +q
alex: Why don't we brainstorm on some possible relationships and suggest that these are some of the possible ones that fit some particular use cases and other than that, leave it open.
<scribe> ACTION: gary to formalize relationships and create some use cases [recorded in http://www.w3.org/2011/01/12-poiwg-minutes.html#action01]
<trackbot> Created ACTION-25 - Formalize relationships and create some use cases [on Gary Gale - due 2011-01-19].
andy: Is there a need for relationships beyond inheritance? e.g. Silicon Valley contains these areas and S.V. is in California. Do we need that contains within or can we do just inheritance?
... Do you need a relationship that implies some sort of copy of the data, or do you want just the geographic closeness?
gary: I believe you do.
cperey: There may be metadata attached to a POI, and it has a relationship to the POI and the object that's being displayed, and authority that published information about the relationship that may be relevant.
... Do you have a primitive for the source?
alex: I think this helps stimulate something we need to get our heads around.
... This comes back to the POI as a descriptor for everything debate.
... This comes back to what's in scope. For instance are governmental relationships in metadata about POIs, or is it an important fundamental to POIs?
gary: Clarify a bit, are you talking about administrative relationships or something beyond that?
<cperey> a relationship, in my opinion, does not have to be related to a government
alex: Are these relationships something all POIs are going to have or not?
... POIs have all sorts of valid information that we want, but is it valuable to a point of interest, is it in scope?
gary: I think it is, not every POI will have that, but one of the real world use cases is looking at schools. We wanted to be able to search for school POIs which fell within the two forms of governmental administrative areas, both of which should be POIs.
... Could we find out how to do that? We could with some low level data hacking.
Alex: I think that's a nice point. Devil's advocate though: we're talking about a myriad of relationships. Some are inferential though. There's an area that has a geographic footprint, isn't this what databases do already? Does it need to be codified? If relationships have been left out, how do we determine which are in and out?
gary: My view is that everything is a POI, and if we have a model that allows you to model everything and a set of relationships that allow you to model slightly more ??y relationships, then I think we've built something of real world use.
alex: No argument with that.
Andy: Where did we land on the question of provenance as a relationship?
... I think it makes sense to bring in some of these things and hack them out at another point.
alex: Gary made a statement about being cautious about what we prescribe, we want to hit a sweet spot where it's useful for people. I think we should give examples of relationships that we think are useful (including provenance), and yet not over prescribe.
... I think provenance is useful, and I think if we had a usecase to go alongside it.
gary: What's the difference between source, freshness, etc, as an attribute, and provenance as the relationship?
cperey: I believe provenance has a role in a POI. A relationship seems like a valid place, it ties in a person place or thing with it's source.
... If it doesn't what is the relationship primitive for?
... My chair doesn't exist in any government database, but my house does and my chair in my house.
... If there's a relationship primitive, it needs to be able to not be fixed to a geospatial grid.
alex: Relationship is quite a generic term and as the provenance or authorship -- a lot of the things we're ascribing to POIs are relationships. We need to figure out what is the dividing line. What are some of the canonical things we have from a relationship standpoint?
... We should be aware that we're talking about a powerful term when we talk about relationships.
<gary> Almost everyone is echoing ... even to myself
alex: I want to pare it back to things that are directly related to POI. A few of the things brought up sound like meta-data to me.
gary: Plus one to that.
alex: I think the challenge is to figure out what are these things we have to have in the POI.
<Andy> +1 to Cperey comment on it not being strictly tied to geo
gary: I need to reiterate things we mentioned in Atlanta. I think we agreed that we want to achieve here is not to describe every single thing in the world, as that's a monumental task, but at the same point in time we don't want to make the model so rigid and brittle that it cracks when someone tries to use it in a way we didn't expect.
... This is why we came up with this set of primitives that have a strong geo location and geo spatial bias, but the fact is that the work being done out there in industry and in the community is at the moment predominated by geospatial/located constructs.
... I haven't heard anything yet that keeps us from having short temporal things, or things in motion.
<scribe> scribe: matt
gary: We don't necessarily have to put that in the first draft.
<Andy> scribe: Andy
<matt> alex: Well put.
<cperey> I am sorry, I need to sign off
<cperey> bye!
<gary> +1 to that Alex
<matt> Alex: Can't we use this meeting as a way to round out our draft and submit it?
<matt> matt: Will someone be there to present it?
<matt> andy: Christine.
<matt> andy: My travel schedule is a little crazy right now.
This is scribe.perl Revision: 1.133 of Date: 2008/01/18 18:48:51 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: RRSAgent_Text_Format (score 1.00) Found Scribe: matt Inferring ScribeNick: matt Found Scribe: matt Inferring ScribeNick: matt Found Scribe: Andy Inferring ScribeNick: Andy Scribes: matt, Andy ScribeNicks: matt, Andy Default Present: Matt, Andy, +44.750.800.aaaa, Gary, cperey, +1.773.575.aabb, Raj?, Alex Present: Matt Andy Christine Alex Gary Regrets: Karl Ronald Found Date: 12 Jan 2011 Guessing minutes URL: http://www.w3.org/2011/01/12-poiwg-minutes.html People with action items: gary[End of scribe.perl diagnostic output]