Re: ldp-ISSUE-62 (siblings): Creating Sibling Containers [Linked Data Platform core]

hi John, 

> > > Every LDPR has a representation with a number of triples. 
> > > An LDPC is a subset of those triples - same-subject, same-predicate. 
> > > That's what I mean by 'inside' - maybe I should've used 'subset'. 
> 
> > Well, they are related by definition because in the same-subject, 
> > same-predicate definition of a LDPC it has a single LDPR resource as
> > the same-subject. I would hope that HTTP client can discover this - 
> > using the membershipSubject triple for example. 
> 
> same-subject, same-predicate is NOT the definition of a LDPC.  It is the definition of Membership Triples.  an LDPC has additional triples, e.g. <memb-subj, a, ldp:Container> 

That's clear. 

> 
> I do not see where in the spec (I checked Editor's draft, but I'm pretty sure none his would be affected by the choice) it says that the membership subject MUST be an LDPR.  Indeed, doing so would (definitionally) preclude the use of hash URIs for membership subjects, since the spec DOES define an LDPR as an HTTP resource (which, to render an assumption explicit, I read to mean that the LDPR is identified by its HTTP request-uri, and those NEVER include a fragment component in 2616). 

I take your point about hash URIs - that was just me being sloppy. 
What I wanted/intended to talk about was: what does it mean if the membershipSubject is a LDPC ? 

regards,
Roger