Re: RDDL: Namespace URIs as document types. was Re: URIs,names and well known RDDL names

[Jonathan Borden <jborden@mediaone.net>]

Jason Diamond wrote:

>
> Look closely at the grammar for RDF. The rdf:RDF element is optional [1].
> That implies that any element from any namespace (using the typed node
> production) could be the document element for an RDF document.

Uggh. right --- this just reinforces the problems associated with making any
sort of assumptions of what the intended purpose of something is depending
on what it is. (XML Schema is looking more and more unique in our ability to
determine what it is from the root namespace URI).

This isn't a problem for RDDL documents because we can say what it is with
the xl:role, and say what we are using it for with xl:arcrole.

>
> > This spec is intended to be simple and it should be simple for people to
> > understand. Since I'm confused about what the role of xlink:role vs.
> > xlink:arcrole ought be, we need work on clearly defining these issues
and
> > recommended practices.
>
> I think that we basically agree here. The simplest, workable solution that
I
> can imagine is using simple URIs.
>
> xlink:role can be used to identify the "semantic type" of the resource.
This
> should be a URI and, in fact, has to be. XLink requires that both role and
> arcrole be absoulute URIs [4].

No, they must be URI references, but the URI portion of the reference must
be absolute. This is the same as the updated recommendation for namespace
URIs.

...
>
> To me this is as simple as it gets and doesn't require any inferring of
type
> based on document content which you already pointed out was unreliable.
>

Unreliable but perhaps we should give some guidance for software when it
encounters a non-RDDL resource at the end of a namespace URI -- this
discussion points out the problems that software will encounter trying to
make sense out of dereferencing an arbitrary namespace URI in the absence of
RDDL, but there are going to be alot of namespaces out there that may be
hardwired to a single fixed document and our software is going to need to
deal with this possibility.

I was thinking that we might say that when a non-RDDL document is
encountered a "virtual" RDDL directory is defined having a single resource:

-- When the document is XML the xl:role would be equated to the root
namespace URI (or nsURI#root-element)
-- When the document is non-XML the xl:role would be a URL created from the
MIME type (as what you and John Cowan have suggested).

In this case should we leave xl:arcrole null or equate it to the value of
xl:role (which has been determined as above)?

-Jonathan