[GeoJSON] RFC-001 vote
noreply at geocartic.com
Wed May 23 17:04:46 PDT 2007
Allan Doyle wrote:
> VOTE HERE: http://www.yourfreepoll.com/wzdbtaousp.html
Thanks for keeping this going. I like GeoJSON. However, I voted
against RFC-001. As was suggested, I'm posting my arguments to the list.
The changes I'd like to see are reflected here:
This proposal describes the structure for four classes of GeoJSON
objects: Geometry, GeometryCollection, Feature, and FeatureCollection.
The Geometry types described are Point, LineString, Polygon, MultiPoint,
MultiLineString, MultiPolygon, and Box.
In my mind, this proposal keeps GeoJSON simpler. There are a total of 4
required names in the GeoJSON vocabulary: type, coordinates, members,
and properties. With these four names, the above objects can be
described concisely - balancing human and machine readability. (I like
"data" or "coords" better than "coordinates" but that's just because
they have fewer letters.)
I've left the crs definition vague, since I care less about how that is
The changes between RFC-001 and RFC-2 are summarized as: define what
types of objects may be serializes as GeoJSON, get rid of the point as
array of a single array business, get rid of the box as an array of two
points business, get rid of the holes name, make all multi-part
geometries more similar, and make all GeoJSON objects have a type.
I've updated the OpenLayers GeoJSON parser to read and write according
You can paste the examples from the wiki into the parser to see what
they look like rendered.
Interested to hear any feedback.
A few notes:
- I'm sure there are typos.
- Though the OL example spits out only Feature type GeoJSON, it can
read/write all types.
- A Box gets read in as a polygon in OL - so you don't get the same
GeoJSON on the way out.
- If you get the "Bad GeoJSON" response, hit F12 or Ctrl-Shift-L to see
a (potentially) more useful error message.
More information about the geojson