[rdfweb-dev] Re: Suggestions for FOAF model alterations

Jonathan Greensted jonathan.greensted at sentient.co.uk
Thu Aug 7 13:25:13 UTC 2003


Thanks James

We want to be good RDF citizens but we are fairly new to it so all advice is 
appreciated.

I will amend the Outlook2FOAF tool to be RDF compliant.

Has anyone written a RDF validator Web Service so I could knock up a quick drag 
and drop tool rather than having to launch browser, copy&paste, etc.  (I'm lazy 
like that!)

Thanks again.

J.

PS. What do people think about extending the namespace with all the Outlook 
properties?  (Including birthday).  Is this OK or will we get into trouble with 
the FOAF police?

Quoting James Carlyle <james.carlyle at takepart.com>:

> Jonathan
> 
> Your suggestions for changes to foaf:phone property are commendable but are
> invalid from an RDF perspective.  I would suggest that all extensions to
> FOAF are run by the always-available RDF validator at
> http://www.w3.org/RDF/Validator/.  It is also easy to see what the XML is
> describing in terms of statements, and worth checking that this is what you
> want.  I personally _really_ want to keep the RDF model of common FOAF
> usage, and it is easy to loose, and once lost, probably impossible to
> regain.
> 
> > <foaf:Person>
> >    <foaf:name>Jonathan Greensted</foaf:name>
> >
> <foaf:mboxsha1sum>1d353ed24aa4827c34b6c0546e02122d123f1853</foaf:mboxsha1sum
> >
> >    <foaf:phone foaf:type="mobile">tel:070 9211 9075</foaf:phone>
> >    <foaf:phone foaf:type="business">tel:014 8368 5555</foaf:phone>
> >    <foaf:weburi>http://www.sentient.co.uk</foaf:weburi>
> >    <foaf:imageuri>http://www.sentient.co.uk/Images/me.gif</foaf:imageuri>
> > </foaf:Person>
> 
> This is not valid RDF.  You cannot qualify a property such as foaf:phone in
> the way that you are trying to do, because the RDF parser is looking at
> nodes (in this case, Person), and properties (in this case, phone), and the
> attributes for a property element are valid only if the property element is
> empty
> (http://www.w3.org/TR/rdf-syntax-grammar/#section-Syntax-property-attributes
> -on-property-element)
> 
> Your RDF model could be changed, so that the XML looks like:
> 
> <rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
>   xmlns:foaf="http://purl.org/foaf_somewhere/">
> <foaf:Person>
> <foaf:phone foaf:type="mobile" foaf:number="070 9211 9075"/>
> </foaf:Person>
> </rdf:RDF>
> 
> This gives an implied untyped node for the phone property, that has a type
> property and a number property.  But I am generally against untyped nodes in
> FOAF.  Using a Phone node, we would have:
> 
> <rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
>   xmlns="http://purl.org/foaf_somewhere/">
> <Person>
>     <contactableBy>
>         <Phone>
>             <type>mobile</type>
>             <number>070 9211 9075</number>
>         </Phone>
>     </contactableBy>
> </Person>
> </rdf:RDF>
> 
> Although this is longer in XML terms, in RDF terms it is the same as the
> previous example:
> node --> property --> node --> property*2 --> node
> Notice the striping: properties generally start with a lower case name,
> nodes start with an uppercase name.
> 
> I am happy to help bat some ideas around and check RDF compliance if you
> wish.
> 
> James
> weblog: http://takepart.com
> 
> 


_______________________________________________________
Jonathan Greensted ~ www.Sentient.co.uk ~ 078 0221 6262



More information about the foaf-dev mailing list