Geo in RDF (was Re: Censorship?)

Hello William,

On 08/11/2010 23:48, William Waites wrote:
> On Mon, Nov 08, 2010 at 03:28:12PM -0800, Gannon Dick wrote:
>>
>> Latitude and Longitude are a complete coordinate system - the
>> ordering is a continuous function.  Entity Names and
>> Vocabulary Encoding Schemes form a complete set, something a
>> bit different.
>
> If I understand correctly, you see a problem with RDF where
> there is no standard way to express things other than points?
> Such as lines, polygons, polygons with holes, multipolygons,
> geometry collections, the whole suite of shapes that GIS
> systems normally deal with?
>
> If so I think you are partially right. As far as I know there
> has been little work done in modelling these sorts of things
> in RDF, and I think the triplestores that have even very basic
> support for geodata (e.g. points) only support the simplest
> of operations with them (e.g. bounding box or radius search).
>
> That said there's no reason you couldn't express more complex
> shapes in RDF. The process would be fairly mechanical (e.g.
> straightforward translation of WKT, KML or whatever) this
> is already a very well understood area.
Though they may fall short of being 'standards' at the moment, you might be 
interested in the vocabulary at [1,2] if you haven't already come across it. It 
picks up LineString, LinearRing, Envelope (i.e. bounding box) and Polygon from 
GeoRSS[3] and AFAICT with care gives an RDF/XML reading of a a piece foe 
GeoRSS/XML syntax. It also provides linkage to non-WGS84 coordinate reference 
systems.

I'm also aware of an approach that adds a GML/XML expression as an XMLLiteral to 
an RDF node representing the geometric extent of some feature  an :asGML 
property e.g.:

@prefix geometry: <http://data.ordnancesurvey.co.uk/ontology/geometry/> .
@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#> .

<http://data.ordnancesurvey.co.uk/id/7000000000025559>
         skos:prefLabel    "The District of South Gloucestershire";
         geometry:extent <http://data.ordnancesurvey.co.uk/doc/geometry/71548>;
         ... ;
         .

<http://data.ordnancesurvey.co.uk/id/geometry/71548>
         rdf:type geometry:AbstractGeometry ;
         geometry:hectares "53664.694"
         geometry:asGML  "<gml:Polygon xmlns:gml=\"http://www.opengis.net/gml\ 
"srsName=\"os:BNG\"><gml:exterior><gml:LinearRing><gml:posList 
srsDimension=\"2\">356578.2 193831.7 356573.4 193822.9 356529.9 193733.7 356510 
..3 193690.2 356495.1 193661.9...  356578.2 193831.7 
</gml:posList></gml:LinearRing></gml:exterior></gml:Polygon>"^^rdf:XMLLiteral .

This is work from John Goodwin at Ordnance Survey. At present I don't think the 
:asGML triples have been published, but I believe that they are coming (soon).

> By far the easiest way to deal with it is just to put WKT
> into, e.g. dc:spatial (maybe we need a WKT datatype) and
> use any GIS system you like to do the actual indexing. Maybe
> add some built-in functions to a SPARQL engine to help with
> querying...

A widely adopted, ideally single, common practice would be nice!

> Or have I misunderstood completely?
>
> And what does this have to do with censorship?
>
> Cheers,
> -w
>
BR

Stuart
-- 

Epimorphics Ltd                        www.epimorphics.com
Court Lodge, 105 High Street, Portishead, Bristol BS20 6PT
Tel: 01275 399069

Epimorphics Ltd. is a limited company registered in England (number 7016688)
Registered address: Court Lodge, 105 High Street, Portishead, Bristol BS20 6PT, UK

Received on Tuesday, 9 November 2010 12:11:36 UTC