W3C home > Mailing lists > Public > www-rdf-logic@w3.org > January 2001

RE: Reification quoting in RDF/N3 was: A note comparing Conceptua l Graphs and RDF/Semantic Web

From: Bill dehOra <BdehOra@interx.com>
Date: Fri, 19 Jan 2001 17:33:45 -0000
Message-ID: <23CF4BF2C499D411907E00508BDC95E131F888@ntmews_01.interx.com>
To: "'Sean B. Palmer'" <sean@mysterylights.com>, www-rdf-logic@w3.org

:Most people could easily understand that Notation3... but a 
:machine doesn't
:cogitate it, it just parses (and possibly remembers) the triple, and
:assigns "Sean B. Palmer" (a string literal) as the :hasAuthor 
:property of that URL. 

Cogitation by machines or humans is irrelevant here. I'm not making myself
clear. I'll try again: what mechanism is to be used to bind a semantically
empty string to a property, on the web, using rdf language? Binding that
string to a URI isn't adequate in itself and I think it's an error to
conflate a unique entity with a meaningful one.


In other words, what "Sean B. Palmer"?, the machine 
:doesn't know
:anything about that string beyond that fact.
:However, if I tied it in with something unique (a URI):-

Only useful if that URI in turn has associated semantics; in RDF this occurs
presumbly by having properties over that URI. That still doesn't indicate a
binding mechanism in the first instance. 

:
:> So given that predicates such as loves will impose ambiguity,
:
:Not really, it's just a property. If I searched for any :bill 
:that :loves
::scotch, then I could find it on an SW engine, couldn't I? 
:That's enough...
:the machine doesn't have to understand what "love" is :-) 

I never said it did or should or can. It has to be able to ground the term,
determine possible semantics and compute which semantics are to be applied
within a process that can start inferencing over the subject and object
bound in turn to that property. I grant you that "namespaced:loves",
"tomahto:tomato", "tomayto:tomato" and suchlike really gets you out of a
hole with controlling any such ambiguity (on the assumption that programs
which share terms share processes). 


:But also, don't
:forget that the Semantic Web is a machine processable Web intended for
:humans... 

Carts and horses. When the web is presented to a person that person will
impose semantics. That doesn't in any way make a web with machine
processable semantics. I'm sorry if I'm being a PITA about this, but it's
important to avoid invoking generalised AI handwaving and automagicking
demons.

-Bill
Received on Friday, 19 January 2001 12:34:23 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:52:38 GMT