Re: PROPOSAL: CHANGES for CURIEs

looks o.k. to me (I might not be at the meeting tomorrow)

Ivan

Bijan Parsia wrote:
> Below is a schematic of the necessary changes. It should be adequate for
> understanding the scope and effect of the changes.
> 
> The only deviator will be RDF/XML which, for historical reasons, we
> cannot change.
> 
> I tried to align it a bit with Turtle's @prefix syntax. That's a little
> nasty for OWL/XML e.g., <prefix prefix=""..> but, whatever :)
> 
> If we can make a decision about this on Wed, that'd rock.
> 
> I'll try to have a draft of a comment note to the XHTML wg. At the worst
> (they disband) at least future folks would have something to work from.
> 
> =========================
>                 ABSTRACTLY
> CURIE DEFINITION:
>         Declaration of a curie prefix is called PREFIX (or PREFIXDECL or
> whatever).
>         The prefix itself is an NCName (as is the case now).
>         The expansion is an anyURI
> 
> CURIE REFERENCE:
>         The prefix itself is an NCName (as is the case now).
>         You can have
>         The local name follows SPARQL production 100:
>             [100]  PN_LOCAL  ::=  ( PN_CHARS_U | [0-9] )
> ((PN_CHARS|'.')* PN_CHARS)?
>             I'll call this LOCAL
>         and non-normatively references it.
> 
> PROCESSING MODEL:
>         Last PREFIX declaration wins for a given prefix (this follows N3
> & Turtle)
>         I believe the processing model needs to be added
> everywhere...Boris or Peter, can you confirm?
> 
>         Text could be:
>     """"When there are more than one prefix declaration for the same
> prefix, the last declaration overrides all the earlier ones."""
> 
> =================================
>                 MANCHESTER SYNTAX
> 
> Substantive changes occur in:
>     http://www.w3.org/2007/OWL/wiki/ManchesterSyntax#IRIs.2C_Integers.2C_Literals.2C_and_Entities
> 
> 
> CURIE DEFINITION:
>     The productions
>         ontologyDocument ::=  { namespace } ontology
>         namespace ::= 'Namespace:' [ prefix ] full-IRI
>     are replaced with:
>         ontologyDocument ::=  { prefix } ontology
>         prefix ::= 'Prefix:' [ prefix ] full-IRI
> 
> 
> CURIE REFERENCE:
>     The production rule
>         curie := [ [ prefix ] ':' ] reference
>     are replaced with:
>         curie := [ [ prefix ] ':' ] LOCAL
> 
> =================================
>             FUNCTIONAL SYNTAX
> 
> CURIE DEFINITION:
> Substantive changes occur in:
>     http://www.w3.org/2007/OWL/wiki/Syntax#Functional-Style_Syntax
>     The production
>         prefixDefinition := 'Namespace' '(' [ prefix ] '=' namespace ')'
>     is replaced with:
>         prefix ::= 'Prefix' '(' [ prefix ] '=' full-IRI ')'
> 
> 
> CURIE REFERENCE:
> Substantive changes occur in:
>     http://www.w3.org/2007/OWL/wiki/Syntax#IRIs_and_Namespaces
>     The production rule
>         curie := [ [ prefix ] ':' ] reference
>     is replaced with:
>         curie := [ [ prefix ] ':' ] LOCAL
> 
> Then need to add processing instructions and adjust examples and
> references to "namespace". It'd be good to have a bit which says, "We
> deviate from the CURIE draft in such and such a way and follow SPARQL."
> 
> =================================
>     XML SYNTAX
> 
> Pretty  much the same as in:
>     <http://www.w3.org/mid/782D37BF-C450-47A5-A922-D9EDC2252353@cs.man.ac.uk>
> 
> except that the element name is PREFIX or PREFIXDECL (appropriately
> capitalized). I like using the leading @ to mark curies and thus keep
> IRI as the only such attribute.
> 
> The text should include the processing model sentence.
> 
> I'm happy to make these and related changes across these documents if
> the working group so approves.
> 
> Cheers,
> Bijan.
> 

-- 

Ivan Herman, W3C Semantic Web Activity Lead
Home: http://www.w3.org/People/Ivan/
mobile: +31-641044153
PGP Key: http://www.ivan-herman.net/pgpkey.html
FOAF: http://www.ivan-herman.net/foaf.rdf