Identity of reified statements

• From: Sergey Melnik <melnik@DB.Stanford.EDU>
• Date: Sun, 14 Nov 1999 19:12:28 -0800
• To: RDF Interest Group <www-rdf-interest@w3.org>
• Message-ID: <382F7A1C.E70CA946@db.stanford.edu>

Jonas example of his DB schema (see previous postings) led me to the
following question:

is the identify of reified statements unique?

If I use reification (e.g. "John said that Mary is wicked"), the fact
"Mary is wicked" is not stated, so it doesn't matter *who* does not
state it. Therefore, if I have multiple reifications of the same fact,
they must have the same identity. Consider if I add "Sergey  denied that
Mary is wicked". Obviously, John's and Sergey's statements must refer to
the same thing independent from where they came from.

So, what about generating fake URIs for reified statements? Must
applications translate every such URI into something unique to answer
queries consistently? If so, wouldn't it be nice to have a standard
algorithm for that, e.g.

URI = http://..#Mary | http://..#is | http://..#wicked

or so?

Sergey

Received on Sunday, 14 November 1999 22:07:59 UTC