Re: requesting XML records

Andy Powell wrote:

> Ray,
> can you confirm that my understanding of what you are saying is correct...
>
> 1) where the record syntax is XML, the "schema" being used is named in the
>    ESN.

I'm trying to be careful here. If you mean the actual XML schema name, no, not
necessarily.

But let's say we  have the following definition (Theo's):

     "the XML records are encoding according to the DCMI guidelines and
     contain terms from the dc and dcterms namespaces. The records may
     contain terms from other namespaces when they could not - within
     reason - be expressed by terms from the dc and dcterms namespaces.
     It is recommended that as much as possible terms from DCMI
     registered elementsets are being used. "

And, let's say we call this definition a schema (not an XML schema, just a
schema).  And let's say we name it:
 http://www.kb.nl/persons/theo/dcx/

Then, yes.


> 2) the name *may* be a URI.  [I presume that this is *may* rather than
>    *must*]

Yes ("may" not "must").


> 3) if the name is a URI, the URI need not resolve to anything.

Not only need it not resolvel to anything, if it does resolve to anything,
that would be completely orthogonal to any rules we may develop.


> 4) if the name is a URI and the URI does resolve to something, that
>    something may be an XML schema, an RDF schema, a text description of
>    the schema or something else.

Yes.

>
> 5) URIs in the ESN should be used to uniquely name one schema.  I.e. URIs
>    should not be used to name more than one schema and each schema should,
>    as far as possible, only be assigned one URI.

Well this addresses unambiguity as well as uniqueness.  Yes, on unambiguity
(URIs should not be used to name more than one schema definition -- "schema"
in the broad sense as used above).  However, on  uniqueness (each schema
assigned a unique URI), I don't know.

--Ray