- From: Brian McBride <bwm@hplb.hpl.hp.com>
- Date: Mon, 11 Nov 2002 19:58:27 +0000
- To: Graham Klyne <Graham.Klyne@MIMEsweeper.com>, Dave Beckett <dave.beckett@bristol.ac.uk>
- Cc: Jeremy Carroll <jjc@hplb.hpl.hp.com>, w3c-rdfcore-wg@w3.org
At 17:51 11/11/2002 +0000, Graham Klyne wrote:
>At 12:31 PM 11/11/02 +0000, Dave Beckett wrote:
>>Hmm, the EBNF we are using from
>>http://www.w3.org/TR/REC-xml#sec-notation can't express the length
>>restrictions of RFC3066 on the primary-subtag and subtag.
>>
>>so at best we can have:
>>
>> language ::= [A-Za-z0-9]+ ('-' [A-Za-z0-9]+ )?
>>
>>or if we go for lowercase only
>>
>> language ::= [a-z0-9]+ ('-' [a-z0-9]+ )?
>>
>>I'm prefering the latter I think; with pointers to the RFC3066
>>section above. The current N-Triples language definition is too far
>>away from the RFC3066 etc. version.
>
>I don't have strong feelings here, but I note that RFC3066 explicitly
>allows upper- and lower-case. That doesn't mean we can't be more
>restrictive in N-triples. I think either of the above is OK.
The abstract syntax is restrictive to a single case. I suggest we want the
simplest possible mapping between n-triples and the abstract
syntax. Hence, parses are expected to normalize the language code.
Brian
Received on Monday, 11 November 2002 14:57:07 UTC