W3C home > Mailing lists > Public > w3c-rdfcore-wg@w3.org > October 2001

Re: literals must be self-evident

From: Brian McBride <bwm@hplb.hpl.hp.com>
Date: Tue, 23 Oct 2001 07:49:53 +0100
Message-ID: <3BD51311.601@hplb.hpl.hp.com>
To: Pat Hayes <phayes@ai.uwf.edu>
CC: Graham Klyne <Graham.Klyne@MIMEsweeper.com>, w3c-rdfcore-wg@w3.org
Hi Pat,

Literals formed the main technical discussion at the telecon last Friday.  My 
understaning, (haven't checked the minutes yet) is that we agreed that there is 
an important choice to be made between on the one hand your choice 2, where the 
type is 'part of' the literal, and other choices where the type information is 
represented in the structure of triples.

I'll probably get this wrong from logical point of view, but I've got another 
terminological problem.  I wonder if it would help to talk about representing 
instances of concrete data types, such as int, date, float etc, thus 
distinguishing between what RDF currently calls literals and these other beasties.

Brian

Pat Hayes wrote:

>> At 12:43 PM 10/17/01 -0500, Dan Connolly wrote:
>>
>>> That is: it's essential that the interpretation of
>>> an RDF document is a function of the document alone,
>>> and doesn't vary according to the contents of other
>>> documents.
>>
>>
>> I think I agree with the thrust here, but I'd like to clarify 
>> something:  it may be that access to another document will provide 
>> more detailed information about what is stated in a document (e.g. 
>> knowing the domain/range of a property from a separate schema may 
>> allow one to make additional inferences about resources used in an 
>> otherwise stand-alone document).
>>
>> The key requirement here seems to be that the interpretation of a 
>> document in isolation cannot be invalidated when some external 
>> document is also consulted.  (This seems to be a kind of monotonicity.)
>>
>> Thus, I think what you are asking is that the truth under any given 
>> interpretation of some RDF document is invariant;  consulting another 
>> document may restrict the interpretations that are considered to be 
>> models.
> 
> 
> Right, exactly. Thats how the extended MT would work. A datatyped 
> intepretation is an interpretation (of the vocabulary) and a datatyping 
> scheme (of the nodes) which together satisfy some pretty obvious 
> semantic constraints (that the class extension of a datatype name is 
> that datatype, and that every literal occurrence labels a node which is 
> datatyped consistently with the class membership of the literal value.) 
> Then it all works smoothly in the usual way, where the more you know, 
> the smaller the set of typed interpretations gets. Since the typing is 
> applied to nodes, it's impossible for typing in one graph to directly 
> influence that in another (ie except via their mutual constraints on the 
> shared vocabulary used on those node labels.)
> 
> Pat
> 
> 
>> #g
>>
>>
>> ------------------------------------------------------------
>> Graham Klyne                    MIMEsweeper Group
>> Strategic Research              <http://www.mimesweeper.com>
>> <Graham.Klyne@MIMEsweeper.com>
>> ------------------------------------------------------------
> 
> 
> 
Received on Tuesday, 23 October 2001 02:54:26 EDT

This archive was generated by hypermail pre-2.1.9 : Wednesday, 3 September 2003 09:41:09 EDT