Thanks for this implementation! If we implement such distance calculation in a 
search engine, and sort the
result according to distance, this would be local search (search within a special region). In this case the search vector has two dimensions (GPS coordinates). This is a special case of general similarity search
http://www.orthuber.com/wpa.htm
In general similarity search the search result is ordered according to a distance functions which can be defined by all http URI owners (using the convention shown in Figure 2 of
http://www.orthuber.com/wp1.pdf ).
It would be possible to search generally for objects which have similar quantitative (numeric) description. So the general application would be very attractive.

Wolfgang

----- Original Message ----- From: "Toby A Inkster" <[email protected]>
To: "Linked Data community" <[email protected]>
Sent: Sunday, June 21, 2009 10:45 AM
Subject: URIs for great circle arcs


Great circle arcs are the shortest surface paths between two points  on a 
spherical body. I've minted some
URIs (and am serving up RDF/ XML) for great circle arcs on the surface of the 
Earth. Amongst other  things,
the RDF/XML returned will tell you the distance between the  points, the 
compass bearing and the midpoint.

For example, the arc between London and Tokyo is represented as:

<http://ontologi.es/place/arc/51.507778;-0.128056/35.683333;139.766667>

which is a geo:SpatialThing. There is also:

<http://ontologi.es/place/arc/ 51.507778;-0.128056/35.683333;139.766667#points>

which is an rdfs:Container representing the (infinite) set of all  points 
between the two endpoints of the
line.

An additional feature is the ability to link to URIs representing the  
endpoints. e.g.:

<http://ontologi.es/place/arc/ 
51.507778;-0.128056/35.683333;139.766667?uri1=http://dbpedia.org/
resource/London&uri2=http://dbpedia.org/resource/Tokyo>

The document you get returned still has a primaryTopic of

<http://ontologi.es/place/arc/51.507778;-0.128056/35.683333;139.766667>

(i.e. it doesn't alter the linked data URI) but now contains explicit  dbpedia 
references for the end
points. Those URIs don't have to be  from dbpedia - they could be anything - 
they're not used in
calculations of distances, etc - just included in the output.

Possible uses: flight and travel linked data.

Any ideas for improvements?

--
Toby A Inkster
<mailto:[email protected]>
<http://tobyinkster.co.uk>







Reply via email to