Draft minutes 2008-06-24

http://www.w3.org/2008/06/24-awwsw-minutes.html

and in plain text below...  Since all four of us wanted to talk, not  
scribe, the transcript is sort of spare.

24 Jun 2008
See also: IRC log

Attendees
Present+1.781.643.aaaa, jar, DBooth, Alan_Ruttenberg, skw,  
+1.617.253.aabb, alanrRegretsChairSV_MEETING_CHAIRScribejar286
Contents
	• Topics
	• Summary of Action Items

no one is scribing
talking about rdf graph / owl axioms vs. interpretation
dbooth calls this two-level
alan: why does what pat says matter?
... may be interesting, but in practice what does it mean?
dbooth: semweb arch can only talk about graphs ("1st step")
jar: why is identity important?
dbooth: without identity, we descend into a tarpit (esp. w.r.t. IRs)
... an IR can have mass
skw: can you give an example?
alanr: i thought we had consensus that IRs don't have mass, why  
speculate otherwise?
... if anything can be true of something, then the thing is vacuous
skw: q: webarch says 200 is for IRs, so question is, when can I use a  
200?
dbooth: identity on semweb is entirely sets of assertions
jar: what's broken, what needs to be fixed? in the normative docs?
... why are you bringing this up?
dbooth: because ir discussions have been unproductive...
... talking about identity is the way to make progress
jar: so what is an example where talking about identity helps?
... awwsw was supposed to be about http semantics, not semweb  
architecture
alanr: no, charter was supposed to be broader
dbooth: denotation of a URI is determined by a particular set of  
assertions
<skw> I think that RDF Model Theoretic semantics are one set of  
constraints on an interpretation, however....
discussion of rdf and owl semantics, trying to relate to dbooth's view  
of 2-level architecture
<skw> ... the normative appeal to the URI specs. also (IMO) induce the  
constraints that arise from the use of URI as refering names...
<skw> so... that on the semantic web, many URI are not free to be used  
to denote other things.
jar: how does uri declaration idea bear on this discussion?
... there cannot be an algorithm to determine meaning
dbooth: what about rdf:type? isn't there an algorithm for determining  
its denotation?
<dbooth> There *must* be an algorithm for determining the first step  
in the mapping from a URI to what resource it denotes.
alanr: at first i thought that, in absence of CN, when do a GET you  
get the IR. (whether you can ever know what something denotes?)
skw: what you end up having locally is a different resource than what  
you [accessed]
... this is re the PDF file conundrum
jar: denotation is determined by the interpretation sensu RDF  
semantics / OWL model theory, not by the graph
dbooth: to find out what rdf:type means, you read the spec - the prose.
<skw> Ahah... first one i found in RDF semantics is: rdf:type rdf:type  
rdf:Property .
alanr: what are the consequences of looking at the spec?
dbooth: you can think of the prose as being a set of assertions
... then you interpret the assertions
alanr: next step - I test my understanding using a validator and a  
reasoner
<skw> 3.3 rdf:type
<skw> rdf:type is an instance of rdf:Property that is used to state  
that a resource is an instance of a class.
<skw> A triple of the form:
<skw> R rdf:type C
<skw> states that C is an instance of rdfs:Class and R is an instance  
of C.
<skw> The rdfs:domain of rdf:type is rdfs:Resource. The rdfs:range of  
rdf:type is rdfs:Class.
<skw> http://www.w3.org/TR/2004/REC-rdf-schema-20040210/#ch_type

Received on Friday, 27 June 2008 22:21:29 UTC