W3C home > Mailing lists > Public > www-rdf-interest@w3.org > April 2000

Re: missing rdf:type triples

From: Dan Brickley <danbri@w3.org>
Date: Wed, 12 Apr 2000 12:41:18 -0400 (EDT)
To: jos.deroo.jd@belgium.agfa.com
cc: www-rdf-interest@w3.org, eric@w3.org
Message-ID: <Pine.LNX.4.20.0004121212190.17647-100000@tux.w3.org>

Hi Jos,

The mising rdf:type arcs are I believe a glitch in Eric Prud'humeaux's
Perl parser, which I use in RDFViz to pre-process data before feeding to
the AT&T graph tools. Eric may be able to comment in more detail...
(sorry if I'm mistaken Eric!)

For Perl fans: the RDF Perl tools are snapshotted at 
http://slow1.w3.org/1999/02/26-modules/ but are also available by CVS from
the W3C CVS base, http://dev.w3.org/cvsweb/perl/modules/W3C/Rdf/ 
(see http://dev.w3.org/ for details on how to access files via CVS). 
Eric has quietly been building some nice tools in there -- Perl/Tk GUI
with graph browser and query facilities, SQL-backed RDF interface etc...

Re Euler: this is very interesting. Have just linked it from
http://www.w3.org/RDF/ though I was slightly at a loss as to how to
describe it! Just as in the paper circulated earlier today, the issue of
how to layer rules/logic/inference over RDF is being addressed. 

From your README file:

path(X,Y) :- twoway(X,Z), path(Z,Y). 

Can be written as :

<r:path r:subject="X" r:object="Y">
  <r:twoway r:subject="X" r:object="Z"/>
  <r:path r:subject="Z" r:object="Y"/>
</r:path>

(ie. as XML that can be transformed with XSLT. I'm not clear if you're
treating this as RDF, though I guess it would parse into an RDF model. 

Does this extend to multiple namespaces? Could I ask Euler for 
something like the queries I'm doing in the examples at 
	
	http://www.desire.org/docs/research/demonstrations/uri/

This (our example using Jan Grant's Javascript toy-prolog system)
uses a somewhat messy prolog-like syntax with curly-brackets
around URIs, so for example a query looks like the following (ugh!):

triple({http://rdf.desire.org/vocab/webLoc},AbstractID,{http://sunsite.anu.edu.au/mirrors/dlib/dlib/january98/01kirriemuir.html}),triple({http://www.w3.org/1999/02/22-rdf-syntax-ns#type},Location,{http://rdf.desire.org/vocab/PrimaryManifestation}),triple({http://rdf.desire.org/vocab/webLoc},AbstractID,Location),triple({http://purl.org/dc/elements/1.1/title},AbstractID,Title), triple({http://purl.org/dc/elements/1.1/creator},AbstractID,Creator).

Incidentally, Eric implemented a couple of query syntaxes (no rules stuff
yet) in his Perl modules too. So the Perl code can understand the same
format as that I've given above, and also a lisp-like syntax. Sounds like
we're accumulating enough implementations for an RDF query testbed...


I'm personally not clear on when we should use a query language that
appears as a text blob to XML/RDF versus when we should come up with a
more XML/RDF native representation of queries and rules. 

Anyway, nice demo. What are you planning to do with it next?

Dan

On Wed, 12 Apr 2000 jos.deroo.jd@belgium.agfa.com wrote:

> Oops,  corrected URI's ...
> 
> Hi Dan,
> 
> We were able to fix our broken Euler [1].
> We also played a bit with rule syntax ...
> So we have for example [2]
> <?xml version="1.0" encoding="UTF-8"?>
> <rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
>          xmlns:r="http://euler.org/">
> 
> <r:member r:source="mailto:jos.deroo.jd@belgium.agfa.com"
> r:drain="http://www.agfa.com"/>
> <r:member r:source="http://www.agfa.com" r:drain="http://www.w3.org"/>
> <r:subscribed r:source="http://www.agfa.com"
> r:drain="mailto:w3c-ac-forum@w3.org/"/>
> 
> <r:authenticated r:source="Person" r:drain="Mailinglist">
>   <r:member r:source="Person" r:drain="Institution"/>
>   <r:member r:source="Institution" r:drain="http://www.w3.org"/>
>   <r:subscribed r:source="Institution" r:drain="Mailinglist"/>
> </r:authenticated>
> </rdf:RDF>
> 
> Now if we parse this we don't get rdf:type triples for the nodes
>   <r:member r:source="Person" r:drain="Institution"/>
>   <r:member r:source="Institution" r:drain="http://www.w3.org"/>
>   <r:subscribed r:source="Institution" r:drain="Mailinglist"/>
> 
> You can actually see this much better with your Rudolf: RDFViz [3]
> So we were wondering how this can be explained.
> 
> With kind regards,
> Jos De Roo --- AGFA
> 
> [1] ftp://windsor.agfa.be/outgoing/RCEI/NET/euler/index.html
> [2] ftp://windsor.agfa.be/outgoing/RCEI/NET/euler/authen.rdf
> [3]
> http://snowball.ilrt.bris.ac.uk/cgi-bin/rdf-dot/uri2gif?URI=ftp://windsor.agfa.b
> e/outgoing/RCEI/NET/euler/authen.rdf
> 
Received on Wednesday, 12 April 2000 12:41:24 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:51:43 GMT