W3C home > Mailing lists > Public > semantic-web@w3.org > July 2010

Re: IKL example in N3 - (was Subjects as Literals)

From: Henry Story <henry.story@gmail.com>
Date: Fri, 2 Jul 2010 17:15:33 +0200
Cc: Reto Bachmann-Gmuer <reto.bachmann@trialox.org>, Jeremy Carroll <jeremy@topquadrant.com>, Semantic Web <semantic-web@w3.org>, public-cwm-talk@w3.org
Message-Id: <328D2490-15FF-40D4-AA90-90C7CE6530CB@bblfish.net>
To: Pat Hayes <phayes@ihmc.us>
Sorry Pat,

I just read your second presentation. So I see what you are getting at now....

On 2 Jul 2010, at 16:36, Henry Story wrote:
> [4] :Joan rel:siblingOf :Mary;
>          foaf:knows :Bill;
>          rel:hasMet :Bill;
>          :believes { :Bill foaf:name "Robert" } .

What you are suggesting is something is to use literals as names, tied to belief contexts.
That makes sense, and somehow could be written in N3 out like this

?b = :Bill .

:Mary :believes [ = :b
                  = { "Robert"^^:b = ?b;
                           rel:hasMet :Joan . } ] .

Where "Robert"^^:b is Mary's way of thinking of Bill. 
We need the = ?b to tie the speakers belief that her reference points to ?b. Or perhaps that should
be place in the outer context.

So yes then it makes sense to have literals as predicates too. If Mary believes that
rel:hasMet means something a lot more precise in mary's mind then perhaps we would need to write

:Mary :believes { "Robert"^^:b "http://purl.org/vocab/relationship/hasMet"^^:b :Joan } .

And one could come up with some nice syntactic sugar to make that more writeable.

So thanks for that thought. Of course when people mention literals in predicate position I always thought of pure string literals.... like this it makes sense :-)

Received on Friday, 2 July 2010 15:16:19 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 19:48:11 UTC