Re: Support for GeoJSON?

Hey Doug,

Actually, what I had in mind was changing the location representation in 
  the WD so that the geodetic object in the Position object would have 
the structure of a "GeoJSON object", i.e., so that a straightforward 
JSON serialization of Position.coords would result in a GeoJSON object.

More directly to your question, I would imagine that the interop would 
be on the frontend, with a site interacting with multiple sources.  The 
simple use case would be for a site to read location information from 
the API and do an XHR to a service that accepts GeoJSON.  If we think 
this is the only use case, then something like a toGeoJSON() method on 
Position.coords might be sufficient (actually, it might be helpful in 
either case), but it seems a little wasteful to have two representations.

--Richard



Doug Turner wrote:
> Hi Richard,
> 
> Interesting find.  With your suggestion, does one still require a 
> translation between a GeoJSON object and whats in the FPWD?  Also how do 
> you imagine that this interop working?  Is it on the backend side where 
> the UA is reading GeoJSON then exposing it to web content, or a site 
> reading from different sources, or something else?
> 
> Thanks!
> Doug
> 
> 
> 
> On May 11, 2009, at 6:00 PM, Richard Barnes wrote:
> 
>> Hey all,
>>
>> I just became aware of GeoJSON, a JSON format for location information 
>> that's compatible with GML:
>> <http://geojson.org/>
>> It's already supported by several libs and web apps, including 
>> FireEagle and, more recently, Google Latitude:
>> <http://wiki.geojson.org/Users>
>> <http://googlegeodevelopers.blogspot.com/2009/05/build-on-top-of-your-public-latitude.html> 
>>
>>
>> It seems like re-using this object format would help the W3C API to be 
>> compatible with the larger space of geolocation applications, ranging 
>> from the web apps above to full-strength GIS apps.
>>
>> Now, on the other hand, the current object format is limited to points 
>> and complexes thereof (lines, polygons, etc.), but I've hacked 
>> together an extension that supports a better set of shapes that still 
>> maps over to GML:
>> <http://geopriv.dreamhosters.com/geojson/geojson-spec.html>
>> <http://lists.geojson.org/pipermail/geojson-geojson.org/2009-May/000484.html> 
>>
>>
>> I'd like to propose that the Position.coords element be split into two 
>> elements, one of which is a GeoJSON Geometry object (call it 
>> "geometry") for the static parts of Position.coords, and the other of 
>> which (call it "dynamic") which holds the parts related to motion.
>>
>> interface Position {
>>  readonly attribute Geometry geometry;
>>  readonly attribute Dynamic dynamic;
>>  readonly attribute DOMTimeStamp timestamp;
>> };
>>
>> interface Dynamic {
>>  readonly attribute double heading;
>>  readonly attribute double speed;
>> };
>>
>> (The current static attributies (latitude, longitude, altitude, 
>> accuracy, altitudeAccuracy) can be mapped to a circle or ellipsoid.)
>>
>> --Richard
>>
> 
> 
>

Received on Tuesday, 12 May 2009 03:07:36 UTC