Re: Namespaces spec is incomplete

At 09:28 AM 9/15/98 -0400, Murray Maloney wrote:
>At 09:19 AM 9/15/98 -0400, Ralph R. Swick wrote:
...
>>RDF specifically wishes to leave flexibility for schema implementors
>>to use the URI hierarchy and fragment mechanisms to partition their
>>names in the manner they deem most appropriate for the application.
>>For example, a short metadata vocabulary might be defined within
>>a single HTTP resource using fragments to name the individual objects
>>while a long vocabulary (some are expected to grow to order of tens
>>of thousands of objects) would likely use a different partition design
>>in order to have efficient retrieval of a subset of the definitions.
>>
>Does that apply to the the description of the namespace partition 
>that currently appears in the WD? That is, does RDF wish to leave
>out the current namespace partition language to allow flexibility?

The namespace partition language in WD-xml-names does not seem to
interfere with use of namespaces in the manner RDF wishes.

I should have chosen different terminology in my first message on
this thread; I meant to refer to the choices an individual schema
designer has in deciding where to store the definitions of the
named objects according to the facilities of the particular schema
language chosen to author his schema.  Some schema languages will
permit the schema designer to store the definitions in individual
HTTP resources, while others may encourage (perhaps only as an
optimization) the definitions to be stored together within a single
resource.  We want flexibility in defining the fully qualified name
so that this full range of options can be explored.

-Ralph

Received on Tuesday, 15 September 1998 14:43:35 UTC