[Geojson] Editing the 1.0 text
Martin.Daly at cadcorp.com
Thu May 15 07:28:53 PDT 2008
> I've taken it upon myself to edit the official 1.0 text. No changes
> being made to the spec, I'm only editing for clarity and consistency.
> Please reply with feedback ASAP, because the overwhelming consensus is
> that it's time to wrap this up.
> You can always see the latest at
> with HTML version at
> and all changes logged at
I've attached an edited version of the draft text, with changes
1. Use keyword text from RFC2219
2. Clarify that x,y,z means easting,northing,altitude for a projected
3. Move LinearRing description to its own paragraph
4. Allow null as the value for a feature object properties member
5. Reworded 3 to avoid (my) confusion between CRS, crs member, and crs
6. Added a clause to 3 to suggest that the top-level GeoJSON object
should be the crs owner
7. Changed the example file extension to be .crs, instead of .proj, and
the type to "ogcwkt" for variety
8. Changed a couple of mentions of "spatial reference" to "coordinate
9. Changed a should to a must in 4. Bounding boxes
10. Fixed some typos
11. Moved myself to the top of the list of contributors.
12. I was joking about 11.
That means I did sneak one hard change to the spec in there, which is to
allowing as the value for the properties member of a feature object.
That was just for consistency with geometry. If you can have a feature
with properties and no geometry, then why not a feature with geometry
and no properties?
There is also a soft change related to crs objects being only on the
parent or grandparent of an object. This reflects my prejudice that the
feature collection should be the holder, if present, and that the crs
should (but not must, hence the soft change) be homogeneous within the
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
More information about the GeoJSON