[prev in list] [next in list] [prev in thread] [next in thread] 

List:       mozilla-rdf
Subject:    reification; algebra for RDF
From:       Dan Brickley <Daniel.Brickley () bristol ! ac ! uk>
Date:       1999-08-12 6:41:39
[Download RAW message or body]


Fwded from the RDF Developers list (RDF-DEV  
http://www.ilrt.bris.ac.uk/discovery/rdf-dev/).

The algebra may appeal to David. Regarding reification, the spec seems
to suggest that an API would allow us to ask for reificiations of any
fact stored, without care for whether they're "really" in the DB.
Storing a unique ID per triple would be enough to deliver this I guess.
However it does not say that the we can infer presence of a triple from
presence of a reification (scare quotatation) of said triple.

For current Mozilla purposes, reification can probably be set aside;
trust issues for eg might (IMHO) be more usefully dealt with through dig
sigs. Reification of complex claims is also a bit of an unexplored can
of worms: all the examples I've seen are of simple atomic triples.
Reifying a graph into a set of linked rdf:Statement nodes seems
reasonable but hairy and of in many apps dubious merits. I prefer to
ascribe properties to RDF databases themselves, ie. have a chunkier
grain of granularity.

Dan



---------- Forwarded message ---------- Date: Thu, 29 Jul 1999 19:02:55
-0700 From: Sergey Melnik <melnik@db.stanford.edu> To: RDF mailing list
<rdf-dev@mailbase.ac.uk> Subject: reification; algebra for RDF

There are two points that I seem to have missed:

1) Reification

The RDF M&S specification does not say anything about the existence
dependencies of statements and their reified counterparts;

Example:

 (A) (Tom, hunts, Jerry)

 (B) (rdf:type, X, rdf:Statement)
     (rdf:predicate, X, hunts)
     (rdf:subject, X, Tom)
     (rdf:object, X, Jerry)

Does an RDF model containing (A) implicitly contain (B)?
Does (B) imply (A)?

2) Did anyone ever present a set-theoretical model (algebraic structure)
for RDF models? Now that people start reasoning and inferencing using
RDF having a common algebraic foundation becomes important.

Here is a draft of how such an alg. structure may look like; please
comment on it:

http://www-diglib.Stanford.EDU/diglib/ginf/WD/rdf-alg/rdf-alg.ps OR
http://www-diglib.Stanford.EDU/diglib/ginf/WD/rdf-alg/rdf-alg.pdf

Regards,
Sergey

-- 
E-Mail:      melnik@db.stanford.edu (Sergey Melnik)
WWW:         http://www-db.stanford.edu/~melnik
Tel:         OFFICE: 1-650-725-4312 (USA)
Address:     Room 438, Gates, Stanford University, CA 94305, USA

[prev in list] [next in list] [prev in thread] [next in thread]