W3C home > Mailing lists > Public > w3c-rdfcore-wg@w3.org > February 2002

Re: Datatyping Summary

From: Sergey Melnik <melnik@db.stanford.edu>
Date: Mon, 04 Feb 2002 14:51:49 -0800
Message-ID: <3C5F1085.CF45A2A4@db.stanford.edu>
To: Brian McBride <bwm@hplb.hpl.hp.com>
CC: Dan Connolly <connolly@w3.org>, Patrick Stickler <patrick.stickler@nokia.com>, RDF core WG <w3c-rdfcore-wg@w3.org>, Jeremy Carroll <jjc@hplb.hpl.hp.com>
Brian McBride wrote:
> 
> I have been encouraged by Jeremy to repeat some remarks which might help
> clarify this part of this discussion.  There has been an ongoing debate
> about whether "1984" has a -consistent global meaning-.  I suggest some of
> the confusion comes from different use of language amongst members of the
> WG and suggest the following simple model might be helpful.
> 
> Consider an implementation model where we have an application built on a
> generic RDF processing tool.
> 
>                   Application
>             ---------------------  the line
>             Generic RDF Processor
> 
> Lets take S idiom B:
> 
>    <book> <dc:Title> "1984" .
>    <mary> <age>      "1984" .
> 
> Below the line only processing which conforms with the RDF model theory is
> sanctioned.  Below the line, both occurences of the string "1984" denote a
> string.
> 
> This does not preclude an application applying above the line knowledge
> that the value of the <age> property is the lexical representation of a
> integer, get the value of the <age> property of Mary, which is "1984",
> transform it to an integer and then do whatever it likes with that
> value.  But than happens above the line.
> 
> When Dan says that "1984" has a single consistent global interpretation, I
> take him to mean, that applies below the line.  Below the line, only the
> model theory applies.   Nothing I have seen DanC write has suggested to me
> that an application is in some way prohibited from interpreting "1984" as
> the lexical representation of an integer.
> 
> When Patrick says that it has been clearly demonstrated that "1984" cannot
> have a single global intepretation, I believe he is referring to the above
> the line interpretation.
> 
> Thus Patrick and Dan are talking about different things and they could both
> be right.
> 
> Brian

Brian, nice summary! Frankly, I believe S-B is intrinsically flawed,
exactly for the reasons you put forward. It sort of solves the problem
by delegating it to the application layer above. Jeremy also discovered
this "bug". Personally I think of S-B as a simple transitional
workaround whose primary justification is backward compatibility
(specifically, CC/PP and DAML).

Sergey


> At 09:44 30/01/2002 -0600, Dan Connolly wrote:
> >On Tue, 2002-01-29 at 03:19, Patrick Stickler wrote:
> >[...]
> > > I believe that Jeremy's recent 1984 example (in
> > > addition to other examples provided over the past
> > > few days) clearly demonstrates that a literal
> > > does not have consistent global meaning.
> >
> >No, you have not established that as fact.
> >I accept it as your preferred design choice,
> >and I accept that you find S unacceptable
> >in various ways, but S is a coherent design
> >wherein "1984" does have a consistent global
> >meaning.
> >
> >--
> >Dan Connolly, W3C http://www.w3.org/People/Connolly/
Received on Monday, 4 February 2002 17:34:54 EST

This archive was generated by hypermail pre-2.1.9 : Wednesday, 3 September 2003 09:45:04 EDT