Meeting minutes
discussion on profiles (many-to-many transparent and many-to-one opaque) and their exact definition
<AndyS> <<( :s :p :o )>> rdf:subject :s ; rdf:property :p ; rdf:object :o .
<niklasl> We also have the question about what the rdf:type of rdf:reifies is. owl:ObjectProperty in OWL Full, I presume...?
<AZ> What did AndyS write. I was not connected to IRC
<thomas> <<( :s :p :o )>> rdf:subject :s ; rdf:property :p ; rdf:object :o .
<AZ> thx
<AZ> in RDF 1.1, one can write: (:s :p :o) rdf:subject :s; rdf:predicate :p; rdf:object :o . Would you say that this is redundent, since :s is obviously the first element of a list of 3, so it must be the subject?
<AndyS> Can already infer "non-triples" in D-entailment.
<AZ> <<(:s :p :o)>> rdf:_1 :s; rdf:_2 :p; rdf:_3 :o . is this equally redundent as rdf:subject, etc.?
<doerthe> I am surprised, but I really agree with you thomas :D
<niklasl> Like others have repeatedly said, a literal "<s> <p> <o>"^^rdf:Triple is opaque...
<doerthe> niklasl: but that is not queryable
<niklasl> No.
<AZ> Literals and their inner structure are querying to the extent that SPARQL provides datatype-specific functions or keywords (e.g. REGEX for strings)
<thomas> but could be made queryable i reckon
<doerthe> mmm, granted, AZ, yet, not elegant
<AZ> absolutely, thomas!
<enrico> <<(:s :p :o)>> rdf:_1 :s; rdf:_2 :p; rdf:_3 :o . is this equally redundent as rdf:subject, etc.?
<enrico> :x rdf:_1 :s; rdf:_2 :p; rdf:_3 :o .
<enrico> entails <<(:s :p :o)>> a rdfs:resource
<enrico> <<(:s :p :o)>> owl:same-as :x.
<thomas> but :x is just a list ?!
<AZ> I don't understand, but doerthe is in the queue
<thomas> sorry, a bag
<doerthe> yes, but I want you to answer as well :)
<niklasl> it's a bag container yes, nothing more.
<AndyS> One of bag/seq/alt
<doerthe> it's fine if you answer first, then I can think about this discusion
<niklasl> ah true, we don't know the rdf:type.
<doerthe> I do not want to stop the flow, I am interested in this discussion :)
<enrico> <<( :s :p :o )>> a :TripleTerm .
<enrico> :TripleTerm owl:hasKey (rdf:subject rdf:predicate rdf:object) .
<AZ> one possibility is indeed going towards modal logic
<AZ> but there could be other ways
<niklasl> ... kind of like when a singleton property is used as a predicate in an asserted triple .... ("kind of")?
<AZ> are you sure this would be isomorphic (even when considering RDFS etc)?
<doerthe> to decide, I want to fully understand, and that already helped :)
<niklasl> +1 exploring these concerns is clarifying
<AZ> There are logics that are not modal but allow multiple local interpretations (e.g. DFOL, DDL, and anything with local model semantics)
<doerthe> mmm, and would you want to have multiple local interpretations? As an option? Based on the predicate for example?
<doerthe> (currently I am just enjoying playing with the ideas ;) )
<niklasl> ...but what is the domain of rdf:edge ?
<niklasl> And where is the set of use cases for those?
<enrico> graph ::= ( triple | reifier rdf:reifies tripleTerm | reifier rdf:edge opaqueTripleTerm )*
<enrico> <<:a :b :c>>
<enrico> <<":a :b :c">>
<thomas> <<( :a :b :c )>> instead of << :a :b :c >> I assume?
<AndyS> What is <<:a :b :c>> translated to?
<doerthe> thank you all for the explanations, that really helps me