Re: RFC on 'Semantic Web Locationary'

On Feb 6, 2007, at 8:12 PM, Joshua Shinavier wrote:

> Hi Ian,
>
> IMH, newbie O, this is good and useful stuff.  I really appreciate the
> human-friendly URIs.  You asked for feedback, so here are some issues
> I ran into:
>

Thanks for the feedback

> 1) The web service response does not have a URI (e.g. the
> CC:SuccessfulReply in http://www.pipian.com/rdf/places/country/Asia),
> so it's not straightforward to find statements about it.
>
>  <CC:SuccessfulReply>
>    <CC:inReplyTo>
>      <CC:Request>
>        <dc:subject>
>          <wail:Place>
>            <wail:name>Asia</wail:name>
>            <dc:type>continent</dc:type>
>          </wail:Place>
>        </dc:subject>
>      </CC:Request>
>    </CC:inReplyTo>
>    <rdfs:seeAlso rdf:resource="#Asia" />
>  </CC:SuccessfulReply>

I had a reply about this a little earlier today and this should be  
fixed.

>
> 2) When I dereference a continent, region or planet URI in one of the
> pipian namespaces, I get a number of statements about that URI, and my
> (forward-traversing) app is happy.  However, when I dereference a
> country or subdiv (for instance,
> <http://www.pipian.com/rdf/places/country/AZ#AZ>) I get statements
> about a different URI with an owl:sameAs edge to the original URI:
>
>    <urn:iso:3166-1:AZ> owl:sameAs
> <http://www.pipian.com/rdf/places/country/AZ#AZ>
>    <http://www.pipian.com/rdf/places/country/AZ#AZ> rdf:type  
> rdfs:Resource
>
> The point is that most of the information I get back is about
> <urn:iso:3166-1:AZ> instead of the handy HTTP URI.  Why not turn the
> owl:sameAs around and make statements about
> <http://www.pipian.com/rdf/places/country/AZ#AZ> instead?

This SHOULD be fixed.

>
> 3) The graphs you return contain a lot of what seem to me to be
> unnecessary, redundant statements (for instance, statements about Asia
> in the graph about Azerbaijan).  If I want information about Asia, I
> can just dereference Asia's URI.

Not sure I want to completely get rid of this, in part because this  
allows the pages to be cached to get basic information about the  
hierarchy and children of the place without doing multiple URL accesses.

>
> 4) I noticed a circular owl:sameAs statement in each country and
> region graph, dunno about other types.  I also don't know whether it
> matters.
>
>    <http://www.pipian.com/rdf/places/continent/Asia#Asia> owl:sameAs
> <http://www.pipian.com/rdf/places/continent/Asia#Asia>

See my argument about #3 (offline caching of data, so that, if this  
is saved to file:///Asia.rdf it will map <file:///Asia.rdf#Asia>  
owl:sameAs <http://www.pipian.com/rdf/places/continent/Asia#Asia>.   
It also allows for altered data later effecting interpretation of  
cached data, so that an updated population data could be included  
from the URL into a local datastore.

>
> 5) The namespace declaration for the SWEET Space ontology is missing a
> '#', which breaks a few URIs:
>
>         xmlns:sweet = "http://sweet.jpl.nasa.gov/ontology/space.owl"

Thanks.  I was worried about that.

Ian

Received on Wednesday, 7 February 2007 03:35:29 UTC