XMIL in ISIS

Hi Paul,

I've recently come across OpenIsis. I am working on a project called SIMILE,
which is investigating the application of the Semantic Web, particularly
RDF, to libraries. For more background, see http://web.mit.edu/www/simile/

I was interested to see you had considered the use of RDF in OpenIsis on
this page. 
http://openisis.org/Doc/xmlisis

I suspect the example you give has some errors in it. Specifically you don't
want to put the Resource attribute in the datamodel, because this isn't part
of the datamodel, it's an instruction to the RDF parser about how to read
the RDF.

RDF/XML has a striped syntax - see
http://www.w3.org/2001/10/stripes/ 
so really we should write the DC example as

<rdf:Description>
<DC:Creator> 
<rdf:Description>
<vCard:FN>Dr Jacky J Crystal</vCard:FN>
<vCard:TITLE>Director</vCard:TITLE> 
<vCard:EMAIL>jacky@dstc.com.au</vCard:EMAIL> 
<vCard:ROLE>Researcher</vCard:ROLE> 
</rdf:Description>
</DC:Creator>
</rdf:Description> 

where creator is a property, Description is the resource that is the object
of that property (an anon node), and FN, TITLE, EMAIL and ROLE are all
properties of that anon node. To understand this, it may help you to think
of the underlying RDF model - see the enclosed diagram.

Using rdf:parseType="Resource" is just a short hand to get rid of the
rdf:Description element inside DC:Creator, but it generates exactly the same
RDF model. So the data model in ISIS should reflect the RDF model, rather
than the serialisation. One complexity in RDF/XML is there are many possible
serialisations for the same model.

So to summarise this page 
http://openisis.org/Doc/xmlisis
is trying to describe a standard way of mapping XML onto Isis, right? So
although your algorithm may work for XML, your example demonstrates how it
fails for RDF/XML so you probably need a different way of mapping RDF/XML
onto the database?

Now unfortunately I can't yet give you advice about how to fix the
algorithm, as I'm still trying to understand how ISIS works, but we might be
able to have a dialogue about this?  I've :cc'd this to the SIMILE list,
because there are a number of other people familiar with RDF there, so I'm
hoping they may have some suggestions here?

best regards

Dr Mark H. Butler
Research Scientist                HP Labs Bristol
mark-h_butler@hp.com
Internet: http://www-uk.hpl.hp.com/people/marbut/