W3C home > Mailing lists > Public > www-rdf-interest@w3.org > June 2002

Re: Innovation, community and queries

From: Arnold deVos <adv@langdale.com.au>
Date: Tue, 4 Jun 2002 09:25:57 +1000
Message-ID: <004501c20b56$08dc6630$ea01a8c0@cockatoo>
To: "Patrick Stickler" <patrick.stickler@nokia.com>, "Danny Ayers" <danny666@virgilio.it>, "Uche Ogbuji" <uche.ogbuji@fourthought.com>
Cc: "Seaborne, Andy" <Andy_Seaborne@hplb.hpl.hp.com>, "ext Thomas B. Passin" <tpassin@comcast.net>, "RDF Interest" <www-rdf-interest@w3.org>


Thanks for the comments.


----- Original Message -----
From: "Patrick Stickler" <patrick.stickler@nokia.com>
>
> > [1]  http://www.langdale.com.au/RDF/NexusQueryLanguage.pdf
>
> This seems more like query in XML than in RDF. Perhaps I missed something.

Oops, the presentation is intended for a different audience and is given in
terms of a non-striped syntax.  Sorry for the confusion - I should have
prefaced this reference with an explanation.  The template and call elements
translate to properties in a query graph.  Internally, the implementation
accepts two graphs (the query and the context) and produces XML.

> The template priority looks interesting. Examples of its use in practice
> would be nice to see.
>

Template priority is a feature that occasionally gets used.  It is analogous
to the same feature in XSLT and disambiguates when more than one template
would otherwise match.  Consider a general template (say for matchning
serializing rdfs:label's) and a special case template (say for handling
rdf:label's of rdfs:Class's).  The latter would be given higher priority.

> > [2] http://www.langdale.com.au/RDF/DAML-Query.html
>
> This is clearly RDF query in RDF (well DAML ;-)
>
> I find the partitioning of the query into the select and from portions
> a bit cumbersome -- in the same way that trying to view an XML instance
> stored in an RDBMS is cumbersome.

Yes it looks like a throwback to RDBMS :-)

>  You have to kinda keep track of
> the target description on several levels -- first specify the properties
> of relevance in the select statements and then describe, be means of
> a class no less, the actual value constraints for those properties.
>

No, the from section constrains the subjects of the result statements and is
independent of of the select section which constrains the predicates of the
result statements.


> And don't the property restrictions make the property select statements
> redundant?
>

Well, the property restictions are part of the definition of a set of
subjects.  The select properties determine what statements about those
subjects will appear in the results.

> I think that the average RDF user (or even the average advanced RDF
> but not DAML user) will not warm too much to such a representation.
>
> Still, DAML die-hards may feel right at home with it ;-)
>

Probably needs work.  I am still looking for something to use in my
application (alluded to in the introduction)....

 - Arnold
Received on Monday, 3 June 2002 19:24:20 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:51:54 GMT