URIs, names and well known RDDL names, was: Re: Quick edit

[Jonathan Borden <jborden@mediaone.net>]

Henry S. Thompson wrote:

>... so far so good
>
>         "When the related resource is an XML namespace compliant
>         document and when the resource can be distinguished by the
>         namespace of the root element, this namespace URI can be used
>         as the value of the xlink:arcrole attribute."
>
> right, except replace 'can' by 'must' (see below)
>
>         "It is anticipated that many related-resource types will be
>         well known. A list of well-known resource types may be found
>         at http://www.rddl.org/arcrole.htm (which itself is an RDDL
>         directory)."
>
> OK, but why doesn't this file include the relevant namespace URIs, for
> use as the value of 'arcrole' as the first two paragraphs suggest?

	I've edited arcrole.htm to reflect that.
>
>         "The well known names defined in arcrole.htm are specifications
>         which they name."

	and remove this (written late night??)
>
> Hunh?  I can't make _any_ sense of this sentence.  When I go to
> arcroles.htm [2], I find statements such as
>
>   "The URI http://www.rddl.org/arcrole.htm#RELAX can be used as a
>   well-known URI for RELAX."
>
> Why are you proliferating names which are _not_ well known, when
> perfectly good ones exist, e.g. "http://www.xml.gr.jp/xmlns/relaxCore"?

	The point is very well taken that often the namespace URI of the root
element of a document *may* be the correct value for xlink:arcrole, but it
need not be and hence I wouldn't go so far as to make this a MUST rather
than a MAY or even generally SHOULD.

	For example:

1) RDDL itself, root namespace = "http://www.w3.org/1999/xlink" ... in this
case the root namespace is not useful to  distinguish between an XHTML 1.0
strict, transitional, frameset, XHTML Basic 1.0, or RDDL document. So we
very well may need to pick another URI to refer to each of these formats.


> The whole point of the RDDL exercise is lost if software has to keep
> up with an arbitrary set of made-up URIs to identify resource types:
> please identify _existing_ namespace URIs whenever possible,


	Right, emphasis on "whenever possible" and realize that
http://www.w3.org/2000/10/XMLSchema may work for XSD under most conditions
but we need to retain the flexibility to support, for example, multiple
schemas on the same namespace URI. For example, suppose we have one schema
for  editing and another for runtime validation. Each may need a different
arcrole URI (or given a constant URI some other mechanism to distinguish
between related resources).

	The idea is that arcrole.htm is where a developer can go to find a
convenient and documented name (URI) for something.


> and open
> discussion on this list for alternatives when no obvious candidate
> exists (e.g. for CSS).


	There are lots of ways to generate URIs for stuff like MIME types e.g.

	http://foo.org/MIME/text/css

	the advantage of arcrole.htm is that these names would be documented in one
location -- but the values of the attribute xlink:arcrole are not limited to
what's defined in arcrole.htm

-Jonathan
>