Add definitions for triples and triple patterns

[kjetilk]

Addressing #348 .

I decided to paraphrase, as the formal definitions in particular for SPARQL, did not seem appropriate to put our document. Please check that out. I also decided to constrain to URIs, as we generally have done, but see #347 .

I'm not particularly confident this is the right thing to put in there, but the normative references are there for those who need them.

[damooo]

URI restriction must be only about identifiers assigned to information-resources managed by solid server, independent of triple definition.
It must no way restrict general definition of triple to have it's terms to be only uris. Thus it just bars away much of linked-data.
According to spec, there cannot be triple <<dbpedia:rama> dbpedia:wikipage wikipedia:राम>>. in a resource <exampl.org/about-rama.ttl>.
It is contrary to rdf-spec. subject, predicate, object identifiers can be any iri in a triple.
Only identifiers we assign to solid-resources is constrainable, whether they are used in a triple in any position or not.

[kjetilk]

Yes, indeed. But if we defined it in terms of a IRI, we'd have to define IRI too. Perhaps we shouldn't introduce definitions in our document at all, but rather reference the relevant specifications in the places we use those terms?

[kjetilk]

Solid Editors resolved to close this and instead merge #351 .