RE: We should turn facets into URIs in the spec


Since nobody objected, I have just changed the spec and have made constraining facets URIs. Here are the diffs:


Direct Semantics:

RDF Mapping:

XML Syntax:

The profiles didn't need updating because they don't support facets.

Michael Schneider might need to update the RDF-Based Semantics, I'm not sure.



> -----Original Message-----
> From: [] On Behalf Of Boris Motik
> Sent: 22 September 2008 16:32
> To:
> Subject: We should turn facets into URIs in the spec
> Hello,
> Peter has just detected an inconsistency in my resolution of ISSUE-71. The problem is that all facets
> are turned into xsd:<facet
> name>, which puts the langPattern facet into the xsd: namespace -- probably not what we want.
> Luckily, there is an easy fix to this: we should turn facets into URIs. The new XML Schema Datatypes
> document
> ( explicitly says that each facet is
> identified by a URI. Thus, we should
> obey this convention and adapt our spec accordingly; this will immediately resolve the mapping
> issues.
> Concretely, this means that instead of writing
>     DatatypeRestriction( xsd:integer minIncusive "10"^^xsd:integer)
> we should write
>     DatatypeRestriction( xsd:integer xsd:minIncusive "10"^^xsd:integer)
> This will require a small change to the syntax (and the XML Syntax as well). The URIs of these facets
> would be placed into the
> reserved vocabulary, which will prevent any problems with the declarations. Finally, the RDF Mapping
> would be modified to use the
> URI of the facet as is, without putting it explicitly into the xsd: namespace.
> I believe this to be a relatively low-level problem that can be treated as being editorial. Hence,
> unless anyone objects, I'd go
> ahead and make this change in the following days. Please let me know if you disagree with this
> proposal.
> Regards,
> 	Boris

Received on Wednesday, 24 September 2008 18:40:32 UTC