Re: normative statement that xsd:string subclass of plain literals

I think it's not entirely up to us to make this statement normatively, as 
it depends upon the definition of xsd:string which is not in our 
remit.  There was also a small matter (I think) that some valid plain 
literals were not valid xsd:string, or vice versa, which would kill the 
entailment as simply stated (or is that now behind us with the NFC 
revision?) -- as hinted at by 'for suitable values of "aaa"'.

I suggest the entailment should follow logically from, rather than be 
asserted by, our specification together with the xsd specification, which I 
think it does.  And maybe reinforced by a suitable test case.

#g
--

At 10:35 09/10/03 +0100, Brian McBride wrote:

>Dave Reynolds was asking me about the relationship between plain literals 
>and xsd:string, i.e is a plain literal without a lang tag (modulo some 
>funny characters) an xsd:string.
>
>I adopted the policy of referring him to the spec and indeed he came back 
>and said "found it, its in the bit about entailment rules."
>
>"Bu**er", I said, that section is now informative!
>
>I just wanted to check whether we have a normative statement in the specs that
>
>_:a eg:prop "aaa" .
>
>xsd:string entails
>
>_:a eg:prop "aaa"^^xsd:string .
>
>for suitable values of "aaa".  I should check this myself, but I'm burned 
>this morning and don't want to forget ...
>
>Anyone know?  Does it matter?  Is there anything else we might have lost 
>by making that section informative?
>
>Brian

------------
Graham Klyne
GK@NineByNine.org

Received on Thursday, 9 October 2003 06:53:18 UTC