[Geojson] Need to know the type of a Feature

[Tim Schaub]

Hello Antoine,

In case it's not already clear, the spec allows you to add arbitrary
members to features or other GeoJSON objects.

http://geojson.org/geojson-spec.html#geojson-objects

So adding a "name" member to Feature objects would still be valid GeoJSON.

Are you interested in having the spec edited so that parsers would
preserve the "name" member or in hopes that people would start sharing
known feature types?

Tim

On Mon, Nov 5, 2012 at 8:11 AM, Riche, Antoine (GE Energy)
<antoine.riche at ge.com> wrote:
> The GeoJSON format does not provide a way to describe the type of a
> Feature. The "type" member of a Feature object just says it is a
> Feature, so parsing GeoJSON data does not enable a web client to find
> out whether a feature is a motorway or a footpath, whether it is a gas
> pipe or a cable. This information however might be needed for the
> response to a WMS GetFeatureInfo request, as there might be several
> features of different types in the same position (where the user clicks
> on a map). Generally speaking a FeatureCollection can be heterogeneous,
> so providing a mean to distinguish the features it contains would be
> useful.
>
> I can think if a couple of ways to add this information to the GeoJSON
> format :
>
> 1 - Add a member called "name" to Feature objects, e.g.
>     { "type": "Feature",
>       "name": "Cable",
>       "geometry": {"type": "LineString", "coordinates": [[102.0,
> 0.5],[108.5,12.0]]},
>       "properties": {"depth": "1.5 metres"}
>     }
>
> 2 - Allow the type value of a Feature to be any string, e.g.
>     { "type": "Cable",
>       "geometry": {"type": "LineString", "coordinates": [[102.0,
> 0.5],[108.5,12.0]]},
>       "properties": {"depth": "1.5 metres"}
>     }
> Although this loosens the third rule for GeoJSON Objects
> <http://www.geojson.org/geojson-spec.html> , this could be handled by
> parsers as long as the value of the type member is relaxed for Feature
> objects only - note the type "Feature" remains valid for anonymous
> features.
>
> The benefit of the first solution is backward compatibility (as long as
> parsers ignore members whose name is unknown). The benefit of the second
> solution is that it is less verbose - which is one of the goals of JSON.
>
> Regards,
> Antoine Riche.
>
> _______________________________________________
> Geojson mailing list
> Geojson at lists.geojson.org
> http://lists.geojson.org/listinfo.cgi/geojson-geojson.org



-- 
Tim Schaub
OpenGeo http://opengeo.org/
Expert service straight from the developers.