Re: XML names are cool

At 12:12 AM 10/27/97 -0500, Al Gilman wrote:
>The way the XML name facility is discussed in the RDF core draft
>it is what I would call a "qualified name reference"

Yes.  The key point is that the qualifier actually expands to a
full URL so rather than a central registry of metadata we can
create "global names" simply with URLs.

>   For the
>name "bar" that appears in only one of the dictionaries, you
>don't bother with a prefix and just say "bar".

No, you will always want to specify the namespace qualifier.
Most likely an unqualified name will have a very particular
binding (e.g. to the DTD of the entire document if there is one).
Best not to assume any semantics for now.

>PS: We don't want to reserve a prefix.  The prefix should be
>selected relative to the using document, not solely determined by
>the used document.

Right; the prefix (or "NSpart" as the current proposal in the XML
WG reads) is strictly local to the referencing document.  It serves
only to syntactically substitute for a full URL on every element name.
The referencing document must also give the corresponding URL.

What you *will* want to specify at some point is the URL that means
"The W3C WAI Specification, Version 1.0".  The content that lives at
the other end of that URL can be 1) the W3C-REC document that is your
WG's output, 2) a machine-interpretable RDF Schema describing the
precise semantics of each of the WAI elements, 3) an XML DTD for
validating documents (if you wish), 4) all of the above.


Received on Monday, 27 October 1997 08:56:16 UTC