Re: A concern with the Datatyping proposal

I should have been clearer in my previous message [1]:

   http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2002Feb/0525.html

At 05:47 PM 2/19/02 +0000, Graham Klyne wrote:
>With reference to:
>
>   http://www.coginst.uwf.edu/users/phayes/DatatypeSummary3.html
>
>I have a concern with the use of subproperty declarations to "switch 
>idioms"; e.g.

[...]

>And from the subproperty assertion:
>
>   _:a2 rdf:dtype exa:octalInteger .
>
>which has introduced a datatyping conflict.

I meant "datatype clash", in the sense that Pat uses in section 4 of his 
proposal document.

In the sense that we now have this:

   _:a2 rdf:value "10" .
   _:a2 rdf:dtype exa:decimalInteger .
   _:a2 rdf:dtype exa:octalInteger .

which, in effect, gives two inconsistent ways to interpret the literal.  Is 
the value of node _:a2 8, or 10, or something else?

I'm not trying to say this is necessarily invalid RDF, but rather that it's 
a very undesirable position to be in, as Pat says:

[[[
     Exactly how such conflicts would be detected and what the resulting 
behavior
     would be is not fully defined, and might depend on applications 
external to
     RDF. Possibilities include posting an error condition, or choosing one of
     the datatyping conventions randomly. In any case, such a clash of 
datatype
     assignments is a Bad Thing.
]]]

(And I also should have said:  I don't think this is a problem with most of 
the proposal, just one particular part of it that I don't think is 
especially vital.)

#g

[1] http://lists.w3.org/Archives/Public/w3c-rdfcore-wg/2002Feb/0525.html



------------------------------------------------------------
Graham Klyne                    MIMEsweeper Group
Strategic Research              <http://www.mimesweeper.com>
<Graham.Klyne@MIMEsweeper.com>

Received on Tuesday, 19 February 2002 14:24:01 UTC