See also: IRC log
<trackbot> Date: 25 August 2011
<robman> hey karl_
<cperey> hello
zkaim, aabb is karl_
<cperey> cperey waves to ahill2
<karl_> did you all get my draft edit pass?
-> http://lists.w3.org/Archives/Member/w3c-archive/2011Aug/0443.html Karl's message (mangled)
<scribe> Scribe: Matt
-> http://www.w3.org/2002/09/wbs/45386/POI-F2F-2011-2/ Registration Page
-> http://www.w3.org/2010/POI/wiki/index.php?title=Face_to_Face_Meetings/September_2011 Logistics
-> http://www.w3.org/2011/08/04-poiwg-minutes.html#item01 Last talk of F2F
ahill2: First day has this ARML ad hoc meeting. Do we want to be a part of that?
cperey: A regular WG we cannot attend unless we are members of OGC, while a Domain WG is open to other people, no intellectual property there.
<cperey> +1
<cperey> I will attend it
karl_: It's 4 o'clock on Monday? Seems like a good one.
matt: Let's do it.
<andy> +1
PROPOSED RESOLUTION: WG will attend ARML ad hoc meeting at 4pm on Monday
<cperey> Mass Market DWG
<cperey> I'm going to attend the 3D Summit
ahill2: There's a Mass Market Domain WG that meets from 1-3 on Wednesday
<cperey> http://www.opengeospatial.org/event/1109tcagenda
<cperey> http://www.opengeospatial.org/projects/groups/massmarketdwg
<cperey> Google
ahill2: Seems germane.
cperey: The primary task seems to
be the care and feeding of KML.
... It's the door through which ARML entered.
... The chair is Ed Parsons.
<cperey> http://www.opengeospatial.org/event/11093dim
[[Discusion currently centers on the following topics: … KML …GeoRSS Geo in relation to Web 2.0…]]
ahill2: Any others?
cperey: 3dim, see above URL
... These people talk about things of interest. They talk about
high density environments. They speak about and apply things
like CityGML. I believe building information management comes
in here too.
... Indoor 3d models, campus models, etc.
ahill2: And the time?
cperey: All day Tuesday.
... I believe they'll publish an update to their program
soon.
... CityGML is on Wednesday
ahill2: Realistically, what do we
think about CityGML… *audio drop outs*
... Don't know how we connect with CityGML.
cperey: It's a spec WG anyway, we
can't attend.
... CityGML and it's applications will be part of the 3d summit
on Tuesday.
<ahill2> please repost 3d summit link
karl_: Maybe we can get feedback from the chairs?
-> http://www.opengeospatial.org/event/11093dim 3D Summit
cperey: I'll do that, and forward new agenda to the list.
<scribe> ACTION: cperey to keep POI WG up to date on 3D Summit agenda, publish it to POI WG list [recorded in http://www.w3.org/2011/08/25-poiwg-minutes.html#action01]
<trackbot> Created ACTION-100 - Keep POI WG up to date on 3D Summit agenda, publish it to POI WG list [on Christine Perey - due 2011-09-01].
ACTION-32?
<trackbot> ACTION-32 -- Christine Perey to invite Henning after Matt has put ID requirements in the wiki -- due 2011-03-16 -- CLOSED
<trackbot> http://www.w3.org/2010/POI/track/actions/32
cperey: I wrote to Henning, but
he was busy, he wrote me back just this week.
... He asked if it was still relevant and if we wanted his
input. I said sure and asked him for which Thursdays he might
be able to attend.
<andy> am i the only one have audio problems
cperey: I don't think we finished the topic, but punted. I'd like to at least punt on it knowingly.
<ahill2> Christine was cutting out on me, too
cperey: He's very into categories and IDs.
karl_: Since there's a proposal we have now, we could run it by him.
cperey: Exactly.
... I asked him to take us on a quick tour of IDs, categories,
etc.
... Asked him for input on the spec,e tc.
... Asked him if a URI is sufficient or appropriately robust
for our purposes.
<karl_> great
matt: Let's go for it.
cperey: When he replies, I'll copy you, Matt.
matt: Copy Andy and Alex too please.
ahill2: Conference call or f2f?
cperey: I actually invited him to
the f2f too.
... I also asked him to submit something for the AR standards
community meeting too.
... He wanted to start with a conference call.
matt: Thank you!
<cperey> yes
<cperey> no other suggestions
<cperey> when are we going to work?
matt: Any other suggestions for things on OGC?
ahill2: I think that's good enough. Set aside some time on Tuesday for the 3d summit, and ARML on Monday.
karl_: Outside of those, we'll probably sweep the draft. What about hot topics? Let's poll the team on weakest points or outstanding topics?
ahill2: Sounds good, karl_?
karl_: IDs are the most germane.
Link too, very empowering, that we allude to with a handful of
predicates, but we need to shore that up.
... Decide what links are for and not for, relative to
relationships, etc.
... We're not going to be able to specify all of the
predicates, so how do you handle the open-endedness of link
predicates?
ahill2: How do we go about avoiding a situation where there are a lot of us in the room but we don't have good answers to these things?
<cperey> GOOD QUESTION!
<cperey> I've been in that room!
ahill2: Is there a Dan Brickley or someone else that can be at this F2F to help us out? Can OGC identify someone maybe?
<cperey> yes, but... who?
<cperey> Carl Reed is the best
ahill2: It would be neat if people who were there at the meeting could be invited to come join us for relevant parts.
matt: That's part of the idea of collocating. We should figure out who want and when.
karl_: Links, we need help there.
<cperey> but he's fully deployed during the TC
karl_: That's got to be a common thing across many specs. There has to be an expert out there.
<cperey> I haven't identified anyone in OGC informed on links...more than Raj
ahill2: Let's figure out who we want to invite over.
<robman> I think my audio isn't loud enough...but i did discuss links in detail with Dan and he seemed happy with my proposal
matt: Raj probably knows who to talk to off the top of his head, but failing that we can go through the presenters, etc, and see who we want.
<ahill2> robman, I avn't heard you yet
cperey: Raj is an expert, as is Carl.
<robman> maybe my mic has died
matt: I suppose we should confirm with Raj.
<robman> doh
<scribe> ACTION: Raj to determine when he, Carl and others from OGC and OGC members who might want to attend F2F [recorded in http://www.w3.org/2011/08/25-poiwg-minutes.html#action02]
<trackbot> Created ACTION-101 - Determine when he, Carl and others from OGC and OGC members who might want to attend F2F [on Raj Singh - due 2011-09-01].
<robman> headset
cperey: There's another way we
can approach this, there's an OGC mailing list from the TC.
When people are looking for something specifically, a press
release, etc, it gets issued to this mailing list.
... If we can specify the area of expertise that we're looking
for, maybe Carl or Raj can send something to that list.
<ahill2> +1
cperey: e.g. "We're seeking a few people with knowledge in X, Y, Z to attend the POI WG"
ahill2: I think that's a great
idea.
... Incumbent on us to write up a list that OGC members can
look at and they can say "I can contribute here".
cperey: Like Carsten who attended the POI WG in Amsterdam.
ahill2: Shall we table the specifics of our agenda until next meeting?
matt: I think between Karl's suggestions on link and id, plus OGC stuff, we've got two days full. I don't think we'll be lacking things for the third day.
karl_: Let's poll the POI mailing list.
cperey: Step 1, send poll to mailing list, step 2 have Raj send to TC list topics we're seeking input on.
<robman> the link model that's already proposed and that was in todays agenda would be good
<scribe> ACTION: ahill2 to send mail to public mailing list to poll for agenda topics for F2F [recorded in http://www.w3.org/2011/08/25-poiwg-minutes.html#action03]
<trackbot> Created ACTION-102 - Send mail to public mailing list to poll for agenda topics for F2F [on Alex Hill - due 2011-09-01].
<robman> yes
<robman> 8P
<karl_> lets just irc it
<robman> faster
<robman> so what aspect of the CG did we want to talk about
ahill2: I gather everyone has
read the emails related to the CG. It's not really that we need
to rehash that discussion, but I wanted to see if everybody was
on the same page about how that influences what the POI WG is
doing.
... There's been discussion about the AR notes, and some
frustration about how we've been handling AR. There has been
important issues with the POI side though. And some issues
around link proposal from robman.
... I want to make sure we're all on the same page about where
the effort into AR should go.
... Will we spend more time on AR in the calls, or focus on the
core?
<robman> the CG is really just for the broader AR discussion as far as i'm aware
<robman> and should not impact the POI stuff...just have input
<robman> it's really just an engagement thing
<robman> yes...but there's not really enough time is there
ahill2: There have been
suggestions that we need to be dedicating more time in these
meetings to AR related content. My feeling is that if we focus
on the next agenda item, the linked data stuff, then maybe
that's addressing a lot of the AR concerns.
... The way it's going to connect up to other content has been
a big concern.
<robman> if we add an AR profile to that link proposal...it is very simple
<robman> just adding orientation etc to certain elements
<robman> sorry
matt: I agree with Rob said. AR profile would be simple in that scenario.
ahill2: The proposal robman made, I agree with 100%. The POI spec should be lightweight and leverage other existing media types and effectively link to those.
<robman> +1
ahill2: That part of the discussion really satisfies a great part of the AR requirements, then the rest can be an AR profile. I'm excited to see that we've come around to this point.
<robman> i was aiming to have a draft RDF/etc version for discussion for next week (if i have time)
ahill2: We just need to do the hard work of figuring out to what extend do we do things like time, labels, location, etc in there.
<robman> yep
karl_: The conversation centers around links, and we should make sure we're talking and working on it in such a way that it's flexibly expansible and sufficient for AR use case linkages. It'll help us flesh out the linking model if it can support AR.
ahill2: Exactly.
<robman> it should not have to just be rdf though
<robman> 8)
ahill2: It also needs to support the concerns of people like yourself and others who come from industry. We don't want to turn it into an RDF ball of spaghetti that no one wants to deal with.
+1 to not just RDF
karl_: Labels, categories and links: if we discuss what's allowable where, we can still keep the spec lightweight with just those constructs.
<robman> but aren't categories and labels just links too
<robman> i think the less primitives we have the better
<robman> but location is definitely key
<robman> muzak
karl_: About categories and labels just being links, well, that centers in my mind around "do you want it to just be a ball of links?" I think that's not where you want to go.
ahill2: Because of overhead and complexity?
<robman> if links can be inlined that can make it more consumable
<robman> links are just for the canonical src reference
karl_: The consumability of it. You want people to be able to easily concoct a POI and share it. Names, locations, with out off board linkages.
<robman> yes...duality is good
<robman> definitely not exclusively external links
karl_: If everything is forced down the role of links that need to be dereferenced, I think that's too far.
<robman> peep
<robman> kml does work like that now-ish
ahill2: Rob has both serialized
and referenced links. I'm not sure how that works in practice.
If I get serialized links, at what point do I go to the
authoritative version?
... How does that serialization affect you karl_?
karl_: Perhaps.
... I try to look at the specs in terms of readability.
... I'd have to look at an example.
<robman> 8)
karl_: That seems easy.
<robman> shoot
<robman> it has the inline
<robman> but not the ref
<robman> correct...not dual in one use...but it uses both patterns
ahill2: About KML: I'm not sure I agree with KML working like that. It has links to external data that have refresh properties and things like that, but they don't have this sort of dual mode where you -- unless you're referring to just doing a link or declarative data.
<robman> but in different places
<robman> html5 is that links are all external...but that's just a cultural choice
ahill2: Wouldn't the W3C/HTML5 be
that we'd have markup with primitives and link people to
those?
... Either it's external HTML or it's inline in a more
traditional matter.
... As opposed to reinventing the link.
<robman> that's a fair comment
<robman> but there's nothing logically stopping the dual use
ahill2: This seems central to us moving forward with linking.
<robman> http://code.google.com/apis/kml/documentation/kmlreference.html#link
ahill2: I get that we've got consensus that we want to go that way, but we've got to do it in a way that makes it both easy and flexible.
<robman> see that example ahill2
<robman> it has an ID in the link tag
<robman> but there's also body content wihtin the link
<robman> so if the id were an href/url src
<robman> yes...but if we use url's for ids
<robman> then we get that
<robman> i have a script typing for me
matt: But that ID is to identify the link itself, not external info, right?
<robman> well i think any useful link could be searialised
karl_: You want to conflate using the body content and the link, right robman?
<robman> link ref should be the canonical location
<robman> but using the serialised version is useful
<robman> ahill2 is right...what i've proposed is not completley in use now
<robman> but it is logically possible
ahill2: There is body content within the link in KML, but that content isn't what comes across in the link. The link is to other KML content. There are some elements that can go into the link element, but KML doesn't actually have this model of serializing some of that data inside the link element.
<robman> and i've only proposed it to make consumption easy
<robman> hehe
ahill2: But, they do have a link
element with content.
... And they do combine that with hrefs to external
content.
<robman> then we could use normal http cache control etc for ref control
karl_: The way I interpreted it: labels are descriptive text, categories are triples, and links are references.
<robman> but in html5 links are really just external predicates
karl_: You could argue that I could address what labels do with categories, or do it with links, and I don't that's a bad thing.
<robman> but do we have to define all the categories etc. if they're primitives
karl_: Labels for descriptive things, categories are triples and links are links. I think that's good.
<robman> where if we use links they can be extended
<robman> ok
<robman> that's my big fear
<robman> 8)
karl_: We're not going to define
categories. We can't get into the business of describing the
universe. A category is just a means to an end.
... It's a way to describe things in the spec.
<cperey> +1
karl_: We can give some rough guidance on what labels and categories (maybe not that word) and links are for, but people are free to use them as they need.
<robman> but if you look at what i proposed to dan
<robman> that was using links as triples
<robman> well part of a triple
ahill2: I'm getting the impression that we have this existing proposal with some primitives that we can leverage when appropriate, serialize inline, and be able to alternately use links.
<robman> yep
ahill2: Some of robman examples are asking questions about elements and attributes, etc, so let's use elements with inlines, without having to add complexity of links per se. I don't know if that's possible, but there's part of me to which that sounds appealing.
matt: Me too.
<robman> i'm sure there's lots of devil in the detail ...but its' a start
<karl_> yes, i have to run
<cperey> I'm signing off
<cperey> bye for this week!
ahill2: Let's finish the call and meet again next week.
<karl_> adios
<robman> cool
<robman> cya all
This is scribe.perl Revision: 1.136 of Date: 2011/05/12 12:01:43 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 Default Present: Matt Present: robman andy karl_ cperey Matt ahill2 Regrets: Ronald Agenda: http://lists.w3.org/Archives/Member/member-poiwg/2011Aug/0004 Found Date: 25 Aug 2011 Guessing minutes URL: http://www.w3.org/2011/08/25-poiwg-minutes.html People with action items: ahill2 cperey raj WARNING: Input appears to use implicit continuation lines. You may need the "-implicitContinuations" option.[End of scribe.perl diagnostic output]