See also: IRC log
<trackbot> Date: 08 June 2010
<mhausenblas> any scribe volunteers around? :)
<mhausenblas> Chair: mhausenblas
<mhausenblas> victimnick: cygri
<mhausenblas> Agenda is at http://lists.w3.org/Archives/Public/public-rdb2rdf-wg/2010Jun/0016.html
PROPOSAL: accept minutes of last meeting, http://www.w3.org/2010/06/01-rdb2rdf-minutes.html
<juansequeda> +1
RESOLUTION: accept minutes of last meeting, http://www.w3.org/2010/06/01-rdb2rdf-minutes.html
<mhausenblas> http://www.w3.org/TR/rdb2rdf-ucr/
ericP: FPWD is public
mhausenblas: let's publicize it widely! tweet, blog, show to your grandma
juansequeda: who announces it on
semantic-web?
... also, i'll write a blog post for semantic universe
mentioning it
ericP: purpose of WD is to
communicate with outside world
... the more we communicate now, the less unpleasant surprises
later
mhausenblas: also, it's advertising for the WG
<MacTed> *heh*
mhausenblas: Harry said that more or less everything is prepared, Ivan offered to chair the panel
juansequeda: who is going to be on the panel?
<mhausenblas> http://semtech2010.semanticuniverse.com/sessionPop.cfm?confid=42&proposalid=3019
mhausenblas: see here for list of panelists: http://semtech2010.semanticuniverse.com/sessionPop.cfm?confid=42&proposalid=3019
juansequeda: DanM also wants to be on the panel
<mhausenblas> ACTION: Eric to send list of SemTech panel members to Harry [recorded in http://www.w3.org/2010/06/08-rdb2rdf-minutes.html#action01]
<trackbot> Created ACTION-61 - Send list of SemTech panel members to Harry [on Eric Prud'hommeaux - due 2010-06-15].
mhausenblas: not sure wether Ahmed is still going to be on the panel; probably just replace Ahmed with Dan in the list on the semtech site
<mhausenblas> ACTION-56?
<trackbot> ACTION-56 -- Michael Hausenblas to create a wikipage for face-to-face -- due 2010-06-03 -- OPEN
<trackbot> http://www.w3.org/2001/sw/rdb2rdf/track/actions/56
<Souri> venue?
mhausenblas: someone needs to take care of local organization -- ericP?
ericP: any volunteers for housing the F2F?
mhausenblas: harry mentioned Stanford
ericP: if everyone ponies up some
bucks, we can get a room and network from the hotel
... anyone staying in a particularly large hotel room?
... i can do local organization as long as it doesn't mean
footing the bill
<mhausenblas> ACTION: eric to take care of local organisation [recorded in http://www.w3.org/2010/06/08-rdb2rdf-minutes.html#action02]
<trackbot> Created ACTION-62 - Take care of local organisation [on Eric Prud'hommeaux - due 2010-06-15].
juansequeda: ask ashok wether we can do it at oracle?
Souri: i can check with Oracle management -- don't know how it works on a Sunday
mhausenblas: ok, ericP is process owner
ericP: Souri, please check that we can get network to the outside and don't have to sign an NDA to get into the room
Souri: location would be 45min outside of SF -- can take train, and then it's one more mile
<alexdeleon> Apologies, Boris and I wont be able to attend
<Souri> time? 9-5pm?
<mhausenblas> Who will be at F2F?
<Seema> not me
<ericP> +1
<juansequeda> Dan and Juan
<Souri> me
<MacTed> I don't expect to ... sadly
juansequeda: Harry likely won't be there
<whalb> no, unfortunately not
<Souri> If Tony can do it, that will be the best (no commute)
<ericP> eric, richard, dan, juan, souri, ashok
low bar: 6
<Souri> sure
mhausenblas: so looks like 6 min, 10 max
<Souri> not much difference
<Souri> options: 1) Tony 2) Oracle 3) Stanford 4) paying for it at a cheap hotel
<mhausenblas> http://www.w3.org/2001/sw/rdb2rdf/directGraph/
<ericP> -> A Direct Mapping of Relational Data to RDFhttp://www.w3.org/2001/sw/rdb2rdf/directGraph/
<mhausenblas> http://www.w3.org/2001/sw/rdb2rdf/wiki/Main_Page
mhausenblas: i thought it's a good idea to have a seed editor's draft to discuss, but there can be wiki page as well
<mhausenblas> http://www.w3.org/2001/sw/rdb2rdf/track/
mhausenblas: please raise issues
in the tracker
... this allows us to create actions for the issues
<mhausenblas> Michael: if you mention issues or actions in mails (ISSUE-XX) or ACTION-YYY
<mhausenblas> This is ACTION-XXX
ericP: i tried to create simple+conservative foundation
mhausenblas: can you sum up what's the overall goal of this doc?
ericP: provide common vocabulary
for discussing mappings
... and define a simple and straightforward mapping that
multiple vendors can implement
Dan: we did a survey paper on
this
... there's existing work. we shouldn't restart from the
beginning
ericP: this document should
define terminology so that we can say, "this existing paper is
direct mapping plus many-to-many plus xyz"
... i don't want to start over, but documenting connections to
all the existing stuff is not usually role of w3c document
Dan: are you looking for some small nugget, some core intersection that everyone who has ever done this agrees to?
ericP: not really; wanted to start with a simple map
Dan: my first thought was, rewrite this in a different notation so i can compare it to that other thing
ericP: i'm agnostic about notation
Dan: notation is important.
charter says we should be looking at RIF
... we could start with W3C notation (RIF), or DB notation
(datalog)
... for myself, either of these is fine
<Souri> As I mentioned in last meeting's post-adjourn session, it is critical to add to the formal part an informal decscription of RDB2RDF mapping (most people will read it)
mhausenblas: who is the target audience? WG or wider world?
Orri: target audience should be
DB people
... can someone outline difference between RIF and datalog?
Dan: substantially the same
Orri: if target audience would be
information architects. they know SQL, but not necessarily
datalog
... and not RIF either
mhausenblas: isn't this targeting
the working group?
... surface syntax of the language is something different
Dan: we should agree on our internal foundations. so internal facing documents with small audience can help
ericP: strong vocabulary also helps the outside to talk about our mappings. write papers about it etc
Dan: let's do something that makes us happy
mhausenblas: do you want to go rec track?
ericP: section 2 is informal
description of the transformation
... section 3 defines notation
... section 4 defines relational model and rdf model
... are constraints (pk etc) usually in the relation or in the
header?
<Souri> foreign key is defined with the relation
juansequeda: creating the
instances from the DB is usually trivial
... but you have to create the ontology first
... we have this all in our paper
... once we have the ontology, getting the instances is just
2-3 datalog rules probably
ericP: if we say, "this relational schema implies that ontology", i don't see how that tells us everything we need for translating the instances
juansequeda: in what you have
now, the ontology is implicit
... the properties have domain and range, you just don't state
them explicitly
ericP: happy to have that
Dan: local ontology has the contents of the database catalog
Souri: translating relational
schema to rdf, first thing is to generate rdf schema. it's
implicit.
... every column becomes datatype prop, every fk becomes obj
prop
... that should be included
<juansequeda> +1 to Souri
<ericP> cygri: i'm trying to understand what you're trying to capture in the local ontology (directed at dan and juan)
<ericP> ... i beleive that the local ontology misses some stuff in the data semantics
<ericP> ... not sure this is needed for the direct graph defn
<ericP> ... not sure that re-stating the constraints is necessary
<ericP> ... is the goal to state the constraints or just the data?
cygri: are we trying to translate the data to rdf? or are we trying to translate the constraints?
<Souri> direct mapping (in a sentence): table => class, column => datatype property, foreign-key constraint => object property
alexdeleon: why not generate a default mapping from the schema, rather than an rdf graph?
<mhausenblas> Michael: juansequeda's comments will be last; closing queue now
juanseqeuda: people often are asking for the ontology on the d2r list, people need this
ericP: that doesn't mean we should do it here
Dan: whenever i write something
in OWL, it has a representation in RDF
... it would be conevneint to have machinery on the OWL level
to be able to operate on things
... should we make everything explicit with OWL if we can?
<Souri> The generated RDF schema has to be exposed to the SPARQL query writer, otherwise no one can write a meaningful query.
Dan: if you don't want to read the OWL definitions, then you don't have to, just interpret it as RDF
<juansequeda> +1 Souri
<juansequeda> and also the need to map it to other ontologies if needed
meeting adjourned
<ericP> http://dvcs.w3.org/hg/stemGraph/file/edc6c17a664d/src/main/scala/Main.scala
This is scribe.perl Revision: 1.135 of Date: 2009/03/02 03:52:20 Check for newer version at http://dev.w3.org/cvsweb/~checkout~/2002/scribe/ Guessing input format: RRSAgent_Text_Format (score 1.00) Succeeded: s/scribe/victim/ Succeeded: s/ericP/mhausenblas/ No ScribeNick specified. Guessing ScribeNick: cygri Inferring Scribes: cygri Default Present: mhausenblas, +1.512.471.aaaa, +1.781.273.aabb, juansequeda, MacTed, +1.603.897.aacc, Seema, +43.316.876.aadd, whalb, Souri, EricP, +49.133.6.aaee, cygri, orri, alexdeleon, boris, nunolopes Present: mhausenblas +1.512.471.aaaa +1.781.273.aabb juansequeda MacTed +1.603.897.aacc Seema +43.316.876.aadd whalb Souri EricP +49.133.6.aaee cygri orri alexdeleon boris nunolopes Regrets: Ashok Soeren Marcelo Li Ma Harry Agenda: http://lists.w3.org/Archives/Public/public-rdb2rdf-wg/2010Jun/0016.html Found Date: 08 Jun 2010 Guessing minutes URL: http://www.w3.org/2010/06/08-rdb2rdf-minutes.html People with action items: eric[End of scribe.perl diagnostic output]