OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: Quick edit

> The sentence
> "The rddl:resource element represents a simple xlink (using the
> attributes defined in the XLink namespace), and an additional
> attribute content-type which provides for an optional content
> type specifier."
> Is grammatically awkward; I think the parentheses need to become
> commas like
> so:
> "The rddl:resource element represents a simple xlink, using the
> attributes defined in the XLink namespace.  It has an additional
> attribute content-type which provides for an optional content
> type specifier."
> (mind you, with every day that goes by, I'm becoming more in favor of
> nuking content-type, but let's see what other people say).

Yeah I've been reconsidering this also. Architecturally its the right thing
to do. The reason I thought content-type might be useful is that it would
provide a way to map non-RDDL namespaces i.e. namespace URIs which resolve
to a single static resource, into the RDDL framework, but it sort of turns
out that everything we are interested in might just be called "text/xml" --
i.e. this could be XSD, RDFS, RELAX etc. so content-type doesn't buy us
much. I can think of some situations when it might be useful, but for those
we could to a http://www.rddl.org/content-types.htm RDDL which stuffs the
content-type in the xlink:title and maps based on that.

bottom line: the API will be enabled to map xlink:title -> arcrole so we can
probably take content-type out (I was going to wait until the API gets more
developed but will defer to the group's wishes in the meantime).

> Just before the "Attributes" section, we need an example of an
> rddl:resource; I'd just grab one of ours from below and paste it in,
> with the lable "The following is an example of an
> <code>rddl:resource</code>
> element, taken from this document.
> I'm wondering if we should say, in the arcrole= section, anything
> along the lines of "if the resource type is an XML language and
> its definition specifies a namespace, the <code>arcrole</code> value
> should be the same as the namespace name."

That's why I originally used arcrole="http://www.w3.org/2000/10/XMLSchema"
to refer to XSD and arcrole="http://www.w3.org/1999/02/22-rdf-syntax-ns#" to
refer to RDF Schema (those are the root namespace URIs).

The API is shaping out to allow either the root namespace URI *or* the well
known name in arcroles.htm. We can change to spec to explicitly reflect

> Why the "#resource" on the end of the role= value.  Not disagreeing,
> just not sure what it buys us.

http://www.rddl.org/#resource is the expansion of the qname "rddl:resource"
(see http://www.openhealth.org/RDF/QNameToURI.htm )

basically adding the "#resource" to http://www.rddl.org/ makes the name
compatible with RDF without causing any harm.

> The new resources, including the ZIP, are cool.  Once you get your
> java interface worked out (I've seen nothing to disagree with
> yet) that can go in there too.

yep. People are asking for all kinds of great stuff, and what makes it
really fun is that most of these are fairly easy to do.