W3C home > Mailing lists > Public > www-webont-wg@w3.org > March 2002

Moving forward

From: Jonathan Borden <jonathan@openhealth.org>
Date: Thu, 14 Mar 2002 16:30:11 -0500
Message-ID: <00ba01c1cb9f$6fe7b7e0$0a2e249b@nemc.org>
To: "WebOnt WG" <www-webont-wg@w3.org>
Cc: "Jos De Roo" <jos.deroo.jd@belgium.agfa.com>, "Jeremy Carroll" <jjc@hplb.hpl.hp.com>, "pat hayes" <phayes@ai.uwf.edu>, "Peter F. Patel-Schneider" <pfps@research.bell-labs.com>
It is really hard for me to believe that these semantic/layering problems
are not solvable.

I have two simple ways of syntactically 'tagging' unasserted triples, one
with a small change to the RDF syntax itself, the second with a new
rdf:parseType similar to "daml:collection" but called "daml:sentence" (or
pick another name -- I don't care) -- pick your poison :-)

First a couple questions:

1) Will allowing "unasserted triples" or "colored triples" solve an
otherwise intractable problem with OWL?

2) Can this problem be otherwise solved by a) LynnS' logic (tm :-), or b)
are unasserted triples _also_ required?

If the answer to (1) is "yes" and/or (2b) is "yes" then we should move
forward with a proposal to RDFCore for a syntax that provided for
'unasserted triples'. Changes to RDFCore at this point will not be welcomed,
and so making such a proposal should not be taken lightly i.e. I want to do
this _only if_ it is essential to OWL (or provides a benefit to WebOnt that
arguably outweighs any delay caused to RDFCore).

There are two simple ways of doing this:

1) allowing an embedded <rdf:RDF> collection of triples to be the object of
a statement. Such an embedded collection might not be asserted. Along the
lines of "daml:collection" such a 'context' or 'bundle' of statements might
be represented in the current RDF as either:

This might  be converted into a collection of reified statements i.e.
<rdf:Bag><li><rdf:Statement .../></li><li> ... </rdf:Bag>

2) One could also create a new parse type _like_ daml:collection but
specifically to create a list of 'reified' statements e.g.

<oneOf rdf:parseType="daml:sentence">
    <Pineapple rdf:resource="#a"/>
    <Mango rdf:resource="#a"/>
</oneOf>


this might parse into a list of reified statements e.g
<daml:List><daml:first><rdf:Statement />...</daml:List>

Of course option (1) could expand into a daml:List, and option (2) could
expand into an rdf:Bag.
The  details of how either (1) or (2) might get converted into RDF triples
that depends on the details of how RDFCore deals with reification, and I'll
leave that alone for the moment.

Jonathan
Received on Thursday, 14 March 2002 16:33:07 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:57:48 GMT