Semantic Web Coordination Group Teleconference

<trackbot> Date: 04 April 2012

<ivan> minutes of last meeting

Ivan: Minutes approved?
... Approved
... Next Meeting?

<sandro> http://www.theregister.co.uk/2012/04/03/ibm_db2_10_infosphere_10/ "Graph data is stored in a special format called Resource Definition Framework (RDF), and you query a data store with this data using a query language called SPARQL."

Ivan: Sandro will chair in 2 weeks due to Ivan being at WWW
... Total lack of self congratulation

<ivan> Paul's mail

<sandro> (but will cancel if nothing interesting is happening)

Ivan: Provenance -v- GLD issues

<ivan> Regret+ MichaelH

Paul: Had a look at GLP FPWDs. Organisation and People have agents; looks like FOAF.

Paul: Also have notion of agent within provenance ontology; redefined and not just direct from FOAF (although modelling style is similar).
... Too many different agents; need mappings or advice as to which one to use.

<bhyland> The GLD ED is here https://dvcs.w3.org/hg/gld/raw-file/default/people/index.html

<pgroth> https://dvcs.w3.org/hg/gld/raw-file/default/people/index.html

<pgroth> http://dvcs.w3.org/hg/gld/raw-file/default/org/index.html

Sorry, but I don't know who is speaking.

<pgroth> +q to respond

Bernadette: Hasn't been responsible for relevant work and so can't really comment.

Sandro: Not sure exactly what is going on but thinks there is plan for merger of vocabularies.
... Expects result to be less like FOAF; maybe no agent concept; has to bridge between foaf and scheme.org; will use rdf:subclass to link; could do same to link with provenance

<Zakim> pgroth, you wanted to respond

Bernadette: Suggests coordinating with Michael and/or Dave Reynolds; make sure fundamentals such as People vocab dovetails nicely.

<bhyland> I agree with you pgroth, people just want direction on the one to use, make it simple to use.

David: seems odd to have so many uris for the same concept; maybe shouldn't have agents in provenance

<pgroth> attribution

Ivan: Are there special req. for provenance justifying having own vocab rather than just referring to FOAF; would prefer using unique uris in W3C land.

<Zakim> pgroth, you wanted to respond

Paul: Prov. has agents and people because attribution is a key req -- need to know who did what; shouldn't use "non-standard" namespaces?

<pgroth> Luc, was the reasoning right?

<Luc> we needed person, organization and softwareAgent. SofwareAgent is not in FOAF.

Ivan: Grey area, but FOAF already used in other standards

<sandro> SEMIC

Bernadette: Agrees with Ivan; May be some problems with FOAF, e.g., w.r.t. definition of European names; let's see if alignment is possible; having a conversation about this would be good.

<sandro> IanH: It may actually be a GOOD thing to have multiple names for the same thing, linked together with subClass or equivalentClass

<pgroth> it is a good thing…technologically… but not from an adoption perspective

<sandro> ... otherwise you're trying to mandate a standard upper ontology, and efforts to do that have historically always come to grief

<sandro> ... because orgs end up have important differences between their upper level concepts, like People

<sandro> ... and even if they mean the same thing, people are used to using different terms for the same thing.

<sandro> ... so why NOT let that all go ahead, and tie it together with subclass/equivalentClass assertions

<sandro> luc: FOAF didnt have software agents, so we needed a new class

<Zakim> sandro, you wanted to say I don't know which is simpler -- making people use a term in another namespace...?

<sandro> luc: I take Ian's view that it's up to the users to decide if this is the same as foaf:agent or not

Luc: provenance had to distinguish persons, organisations and software agents; foaf didn't have the latter; decided to create own classes and leave it to others to decide equivalence or otherwise with foaf

Sandro: Agrees with Ian; worries about history of failed upper ontology efforts; also avoids mixing/importing namespaces

<sandro> ( clearly there a split here based on how people feel about ontology reasoning... )

David: Agrees that strength of RDF is flexibility and lack of upper level ontology, but sees benefit of vocab alignment and (re-)use of (defacto) standard vocab. Biggest problem of warehousing and data alignment is use of multiple vocab. Middle ground is encouraging re-use and only inventing new vocab if needed.

Sandro: did consider and reject due to lack of software agent

David: Don't reject an entire vocab due to lack of one or two terms; try to re-use as much as possible

<sandro> sandro: I'm not saying to throw out the vocab -- of course, relate the two

<Luc> ... and even more nuances once it comes to 'agent' !!!!!

Bernadette: People vocab in GLD wasn't available; let's see if this is sufficient for provenance; don't make people handle complexity of multiple vocabs; surely for something as basic as person we can get some agreement

<bhyland> for the record, my suggestion is that the Prov ED authors speak with mhausenblas and dave reynolds to nut out if People Vocab can be extended to encompass Agent from Prov Data model.

Ivan: Agrees that standard vocab is good; this is how schema.org started; finding "right" vocab was too difficult; would like to explore alignment between two groups

<Zakim> sandro, you wanted to say I don't object to trying to merge the two, but don't make it a requirement

<Luc> +1 to sandro, the topic of agent is a source of never-ending discussions in prov-wg

Sandro: good to try, but setting bar too high; could take a long time; reasonable to let them try but don't push too hard.

<bhyland> I'm proposing coordinating a 1 hr phone call to see if this can be achieved. If so, great. If not, we may punt to two namespaces to handle Agent and People.

Ivan: agreed that some discussion will take place between groups, say in next 2 months

<pgroth> myself and luc

<sandro> There's no question that there WILL be two namespaces. The only question is whether Agent is also in Prov, or only in People.

Ivan: Paul and Luc plus Bernadette and Michael should participate

<scribe> ACTION: Bernadette to coordinate phone call [recorded in http://www.w3.org/2012/04/04-swcg-minutes.html#action01]

<trackbot> Created ACTION-8 - Coordinate phone call [on Bernadette Hyland - due 2012-04-11].

<sandro> action-8?

<trackbot> ACTION-8 -- Bernadette Hyland to coordinate phone call -- due 2012-04-11 -- OPEN

<trackbot> http://www.w3.org/2001/sw/CG/track/actions/8

Ivan: AOB?

<sandro> action-8?

<trackbot> ACTION-8 -- Bernadette Hyland to set up phone call between GLD People vocab people and PROV Agent vocab people -- due 2012-04-11 -- OPEN

<trackbot> http://www.w3.org/2001/sw/CG/track/actions/8

Ivan: W3C management gave green light to linked data platform charter

<bhyland> @ivan, can you please paste link in so we can see what came together in the end?

Ivan: Goal is to present newly started group at SemTech (group start 1st June)

<sandro> http://www.w3.org/2012/ldp/charter

Sandro: Can't have WG meeting at SemTech, but hope to organise informal getting to know each other meeting.

Ivan: SemTech panel about WG on Thursday(?)
... Meeting Closed

<bhyland> my final point was there is a possibility that I won't be at SemTech SF to deliver my talk on Publishing Open Gov't Data Worldwide if a competing conference accepts a presentation (Health Datapalooza in DC same week).