`srid` in GeoJSON output about geo

[bryanjos]

@larsbutler Geo puts it in the crs section which is part of the geojson spec as far as I can tell, but we are probably not using it correctly in Geo: http://geojson.org/geojson-spec.html#coordinate-reference-system-objects

I know there is also the properties field. Would it not work there either?

[larsbutler]

Ah, I misinterpreted the output; I thought you were putting the srid in an srid member like so:

{"type": "Point", "coordinates": [30, 90], "srid": "4326"}

But now I see that it's placed in a crs object:

iex(1)> import Geo
Geo
iex(2)> import Poison
Poison
iex(3)> point = Geo.WKB.decode("0101000020E61000009EFB613A637B4240CF2C0950D3735EC0")
%Geo.Point{coordinates: {36.9639657, -121.8097725}, srid: 4326}
iex(4)> Geo.JSON.encode(point) |> Poison.encode!  
"{\"type\":\"Point\",\"crs\":{\"type\":\"name\",\"properties\":{\"name\":\"EPSG4326\"}},\"coordinates\":[36.9639657,-121.8097725]}"

Geo puts it in the crs section which is part of the geojson spec as far as I can tell, but we are probably not using it correctly in Geo.

It does seem that the crs section was removed from the official spec, according to https://tools.ietf.org/html/rfc7946#section-4. I wouldn't say your usage is incorrect, though. Because the spec doesn't say anything specific, we can just do whatever we want. :) It would be nice to have some standardization, though; I'm a bit frustrated with the lack of specificity in the spec, to be honest.

I know there is also the properties field. Would it not work there either?

There is no properties in a Geometry object. In fact, the spec explicitly states that Geometry objects must not contain a properties object. :/ Therefore, I think your approach of including a crs object is the most correct approach, since it follows http://geojson.org/geojson-spec.html#coordinate-reference-system-objects. Even though the CRS spec there is obsolete, it's better than nothing.

[bryanjos]

Thanks for pointing that out. Will probably have to make some changes here, but not sure exactly what just yet

[larsbutler]

Hey @bryanjos, even though your approach uses a deprecated (but documented--in the old geojson spec) approach and there is no new spec to replace it, I think we can close this issue. For consistency I will follow your approach in geomet.

Thanks for the discussion!