Meeting minutes
discussion on what properties can or should link to triple terms? Issue #127
<tl> https://
<Souri> :r rdf:reifies <<( :r2 rdf:reifies << :s :p :o )>> )>> .
<Souri> :r rdf:reifies :r2 . :r2 rdf:subReifies <<( :s :p :o )>> .
<tl> :r1 rdf:reifies :r2 .
<tl> :r2 rdf:reifies :r3 .
<tl> :r3 rdf:reifies <<( :s . :p :o )>>
<tl> :r1 :a :b .
<tl> :r2 :x :y .
<tl> :r3 :ho :hum .
<tl> << << << :s :p :o >> :ho :hum >> :x :y >> :a :b .
:r rdf:reifies <<( :r2 rdf:reifies <<( :s :p :o )>> )>> .
:r rdf:reifies <<( :r2 rdf:reifies _:x )>> .
_: x owl:sameAs <<( :s :p :o )>> .
<AndyS> Counting changes in SPARQL, which may or may not matter.
<tl> annotation syntax is one level. that's easy, and will hopefully be popular. i see no reason why that should not provide enough simplicity, and why we shouldn't stillallow more complex things in other syntax
<AndyS> Moving the burden onto the query authors - there are more query authors than other concern areas
Conclusion 1: we discussed about restricting (or suggesting to restrict) the usage of triple terms only as object of rdf:reifies, but we believe that their usage should be unrestricted.
<gkellogg> +0
<enrico> +1
<AndyS> +0
<doerthe> +1
<tl> +1
<TallTed> ±0
<niklasl> +0
Conclusion 2: we discussed restricting the usage of nested triple terms, by looking at approximations involving only non-nested triple terms, but we feel that this would be depending too much on the use cases. A rewriting of nested triple terms using owl:sameas would be always possible, but beyond RDF.
<tl> +1
<enrico> +1
<gkellogg> +0
<AndyS> +0.5
<TallTed> ±0
<niklasl> +0.1
<doerthe> +1 (but I think we should give souri the chance to further discuss his arguments)
<enrico> +1 to doerthe
I propose we bring these two issues to the WG
<niklasl> I sympathize based on at least how hard it is to think about propositions of propositions of propositions.