Re: [docbook-publishers] Publishers schema changed

[Scott Hudson <scott.hudson@flatironssolutions.com>]

No offense taken, Norm! Since Dublin Core was only available in XSD, we 
made a "best guess" at how to reproduce the patterns to conform to the 
Dublin Core specification. I think you are correct that the simple 
literal was meant to be restricted to containing a string.

 From the Dublin Core XML Guidelines 
(http://dublincore.org/documents/2003/04/02/dc-xml-guidelines/):

    * A /qualified DC record/ is made up of one or more /properties/ and
      their associated /values/.
    * Each /property/ is an attribute of the /resource/ being described.
    * Each /property/ must be either:
          o one of the 15 DC elements,
          o one of the other elements recommended by the DCMI (e.g.
            audience) [DCTERMS
            <http://dublincore.org/documents/2003/04/02/dc-xml-guidelines/#DCTERMS>],
          o one of the /element refinements/ listed in the DCMI Metadata
            Terms recommendation [DCTERMS
            <http://dublincore.org/documents/2003/04/02/dc-xml-guidelines/#DCTERMS>].
    * /Properties/ may be repeated.
    * Each /value/ is a literal string.
    * Each /value/ may have an associated /encoding scheme/.
    * Each /encoding scheme/ has a /name/.
    * Each literal string /value/ may have an associated language (e.g.
      en-GB).

Note, that the above guidelines are a bit dated, and there is a new 
version of the schema available in RDF / GRDDL. I do not think we want 
to add this level of complexity. We may still need to take a closer look 
at the Dublin Core portion of the schema, especially the "any" pattern.

Thank you Norm for troubleshooting this!

Best regards,

--Scott

Scott Hudson
Scott.Hudson@flatironssolutions.com
http://www.flatironssolutions.com


On 01-Apr-10 5:11 PM, Norman Walsh wrote:
> Folks,
>
> Apologies, in advance, to anyone offended by my cavalier approach...
>
> The patterns in the publishers schema around SimpleLiteral seemed wrong
> to me. They looked like this:
>
> LCSH = SimpleLiteral, xsd:string
>
> Meaning the LCSH pattern required a SimpleLiteral followed by a string.
>
> Looking through the DC XSDs, I think the intent was that an LCSH
> should *be* a SimpleLiteral restricted to containing a string.
>
> I implemented that like this:
>
> SimpleStringLiteral = element * { text }?
> SimpleURILiteral = element * { xsd:anyURI }?
> SimpleDCMITypeLiteral = element * { DCMIType.enumeration }?
> SimpleLanguageLiteral = element * { xsd:language }?
> SimpleLiteral =
>     (SimpleStringLiteral | SimpleURILiteral
>      | SimpleDCMITypeLiteral | SimpleLanguageLiteral)
>
> <aside>Are those supposed to be optional, I wonder?</aside>
>
> ...
>
> LCSH = SimpleStringLiteral
>
> I also tweaked DCMIType so that the enumeration was split out.
>
> I suspect that this:
>
> dcterms.any = element any { SimpleLiteral }
>
> is a mistake too, that the "any" there is also a mistranslation from
> XSD where it was "any element". But I could be wrong and I'm trying
> not to make too many changes.
>
> BTW, all of this is sort of moot since these patterns are unused and
> therefore deleted by the "compile" process.
>
>                                          Be seeing you,
>                                            norm
>
>