- From: Seth Russell <seth@robustai.net>
- Date: Tue, 21 Nov 2000 17:38:51 -0800
- To: Jonathan Borden <jborden@mediaone.net>
- CC: Sergey Melnik <melnik@db.stanford.edu>, RDF Interest Group <www-rdf-interest@w3.org>
Jonathan Borden wrote: > I think you are mixing up the triple and the reification of the triple. The > triple is unique but its reification need not be unique. > > (Bush wonThe Election) > > (SethsStmt type Statement) > (SethsStmt subject Bush) > (SethsStmt predicate wonThe) > (SethsStmt object Election) > > (ECStmt type Statement) > (ECStmt subject Bush) > (ECStmt predicate wonThe) > (ECStmt object Election) Well I think you're right, that solves this particular problem and adhears to the current M&S. However I still have two troubels: 1) Doing inference: The inference paths on refied statements are different than on non reified statements. If I want to reason in the context of Seth's statements, I cannot also reason in the context of what MyMemory believes with the same algorithms. But I realize that this is more just an implementation nightmare than a problem with the spec. 2) Assigning statements to contexts: Again another implementation nightmare if we are to adhear to the reification technique of M&S. In MyMemory every statement can belong to any number of contexts. Saying a statement belongs to a context is a statement about a statement and thefore would require the set of four statements. I will get lost in my underwair if i want to browse around my different contexts. topic: Seth Russell needsHelpWith: MyMemory topic: MyMemory abilityTo: (and: (read: RDF) (write: RDF)) abilityTo: (browse: Context) topic: MyMemory abilityTo (browse: Context) if: can be designed then: will exist topic: (MyMemory abilityTo (browse:context)) if: (can be designed) is: difficult
Received on Tuesday, 21 November 2000 20:36:51 UTC