Re: Proposal for ISSUE-12, string literals

On 13/05/11 16:12, Richard Cyganiak wrote:
> I'd like to add a paragraph somewhere that says, approximately: “A
> datatype rdf:PlainLiteral has been defined in [THAT SPEC]. It's for
> compatibility with systems that require everything to have a
> datatype, or don't have anywhere to put a language tag, and hence
> couldn't represent RDF graphs otherwise. An implementation that
> supports real plain literals MUST use them and MUST NOT use
> rdf:PlainLiteral.”
>
> Would that address your concern about mentioning equalities for
> rdf:PlainLiteral?
>
> Or do you think it would be better not to mention rdf:PlainLiteral at
> all in RDF Concepts?

If it's not needed, then I prefer to not add it but that approximate 
text is OK - reiterate the fact it should not appear in RDF exhcnaged 
between systems.

Generally, I think that adding a third form into the mix when none of 
them are complete solutions seems to be making it harder if/when a 
proper, complete solution in the future.  I have no real evidence just a 
instinct that an extra form can only make migration harder later.

	Andy

Received on Friday, 13 May 2011 15:37:13 UTC