Hi
On Sun 01-Jul-2001 at 08:29:44 -0400, Dan Brickley wrote:
> On Sun, 1 Jul 2001, Chris Croome wrote:
>
> > On Sun 01-Jul-2001 at 02:49:59 +0100, Chris Croome wrote:
> > >
> > > There are 3 types of relation I can represent, child
> > > documents, the parent document (all documents apart from the
> > > front page) and all in line hyperlinks. For all of these I can
> > > get a URI, title and description.
> >
> > Thinking about it I also have the language for the parent and
> > child documents.
> >
> > And I also potentially have dcq:replaces from the redirects that
> > get set up if a document moves.
>
> cool :)
>
> > Is dcq:isPartOf good to use for the parent document and is
> > dcq:hasPart any good for listing child documents?
>
> That sounds plausible; I think they're pretty loosly defined.
OK all the above is now set up, all documents apart from the home
page have a dcq:isPartOf for the parent document, like this:
<dcq:isPartOf>
<rdf:Description rdf:about="/">
<dc:title xml:lang="en-gb">MKDoc testers</dc:title>
<dc:description xml:lang="en-gb">
This particular mkdoc site is for mkdoc testers to play with.
</dc:description>
<rdfs:Comment>
MKDoc testers is the parent document of Help.
</rdfs:Comment>
</rdf:Description>
</dcq:isPartOf>
I have added a comment to make it clear what I'm using isPartOf for,
I don't know is this is sensible.
All documents that have children have a dcq:hasPart, like this:
<dcq:hasPart>
<rdf:Description rdf:about="http://testers.mkdoc.com/help/">
<dc:title xml::lang="en-gb">Help</dc:title>
<dc:description xml::lang="en-gb">
MKDoc help information
</dc:description>
<rdfs:Comment>
Help is a child document of MKDoc testers.
</rdfs:Comment>
</rdf:Description>
</dcq:hasPart>
Is it a bit verbose to have the descriptions of child and parent
documents?
I have not listed the child documents as a Seq but I think I will
since I can't think of any other way to represent the order of the
child documents, however if the following was in place that could be
used...
> For purposes of testing, I simply invented a (fictional for now)
> RDF vocabulary for the HTML LINK types. (Maybe a W3C Note for
> those would be a good idea?)
>
> The HTML LINK types need mapping into RDF at least (and into
> XLink; which is inter-convertable with RDF). DC duplicates bits of
> this (isPartOf etc) but I think there's information (next/prev...)
> in HTML LINK that's not in a DC vocab yet. I don't see any great
> need to duplicate that...
>
> I have an old document on this from Jan 1999, based on experiments
> with using RDF, DC, HTML link etc for sitemaps in Mozilla and in
> Java servlet apps. It was a draft of what might've been a spec,
> but never finished. Maybe of interest now in the context of the
> new improved DC/RDF work:
>
> http://www.ilrt.bris.ac.uk/discovery/rdf-dev/purls/papers/sitemap/
It looks good to me, why not make it official :-)
To sum up, as a result of all this the only things that I don't have
represented in the metadata file is:
* the order of the child documents
* the Contributor who last Modified the document
Chris
--
Chris Croome
http://www.webarchitects.co.uk/
|