[Geojson] GeoJSON '1.0'?

Tim Schaub tschaub at openplans.org
Wed Mar 12 23:07:58 PDT 2008

Christopher Schmidt wrote:
> It is my understanding that using the "EPSG:4326" forces the assumption
> that the ordering is y, x. This understanding is based on feedback from
> conversations at FOSS4G2007, the most memorable of which was with Raj
> (from OGC), who essentially told me (as far as I remember) "You can't do
> that."

Ok, I think we have all agreed that the order of our coordinates array 
is x, y [, z[, whatever]].

Given that (please stop me here if that is not something we all agree 
on), do we really think we can either

1) allow things like

     "crs": {
         "type": "EPSG",
         "properties": {
             "code": 2805

while disallowing the same structure with "code": 4326, or

2) forbid the EPSG codes from being used at all?

I think both are a bit foolish to believe.

I know that there are plenty of people who object to x, y ordering 
paired with a reference to EPSG:4326.  Has anyone spoken to any of these 
people and been given the impression that having a "coordinate_order" 
array makes them happy?

> I have no idea what "You can't do that" means. The letters EPSG mean
> nothing special to me, and to claim that they do seems weird to me. 
> Given that, I have no particular opinion of my own basd on my own
> knowledge. However, it appears to be a not-uncommon opinion that using
> the EPSG code in this way would be 'bad form'. 
> Or maybe I really just misunderstood, and this should have all been
> resolved 6 months ago.
> I don't know enough to have any personal prference: my understanding is
> based entirely on what I have been told, and there is no significant
> support for what I have been told that I am able to point to.
> Regards,

More information about the GeoJSON mailing list