Hi Sergio, below I could give you some clarifications about how I interpreted W3C's draft,
On 10 June 2013 20:19, Sergio Fernández <wik...@apache.org> wrote: > Hi, > > I'd not like to get this discussion dying without getting some conclusions > from all members of the project. > > > On 04/06/13 19:39, Raffaele Palmieri wrote: > >> I would have said container, but the meaning gived to this term in ldp >> draft's W3C is different by the actual concept of context in Marmotta. >> Though I think that there are intersections between context and ldp >> container, conceptually they are two ways to assemble triples, the >> latter introduces some strong constraints for membership and is the >> medium to create and access resources. >> > > Conceptually maybe, but not formally: > > - a context/graph is a quad as <g,s,p,o> > > - a ldp:Container is a triple as <c,m,s> > where m is the membership predicated (rdfs:member by default right now). Not properly, a container is a conformant LDPR, not a single triple. Not necessarily the membership subject will be the LDPC resource itself. A container has always a state represented by the set of member triples <ms,mp,mo> where <mo> may be additionally subject of other triples included in container's representation. > Actually my first implementation of LDP was under the assumption that every > container was also a context/graph. Although at the end I arrived to the > conclusion that it may be too complex, this could be true, but not the > other way (not all graphs would be a ldp:Container). Therefore, since LDP > is not defined in terms of SPARQL, personally I'd prefer not to use > container for describing contexts/graphs. Though, there are not normatives about use of SPARQL, LDPC may be associated to a sparql endpoint in application scenarios, and LDPR may be created, updated, deleted using SPARQL Update. > Cheers, > > P.S.: this refreshes that I'd need to put effort again on getting > seriously back to LDP stuff in Marmotta; I hope to have some more time in > July... > > > for the moment +1 for graph. >> > > Good; thanks for your opinion on this. > My vote for "graph" depends on the fact that in Marmotta "context" has a different meaning than "container", and so "graph" is the only good alternative. > > Cheers, > > -- > Sergio Fernández > Cheers, Raffaele.
