[Geojson] -1 actually (Re: GeoJSON '1.0'?)

Sean Gillies sgillies at frii.com
Fri Mar 14 07:28:20 PDT 2008


Christopher Schmidt wrote:
> On Thu, Mar 13, 2008 at 04:25:50PM -0600, Sean Gillies wrote:
>> As I said in another thread: at the very least I believe we need to 1)
>> explain how most users can use the default CRS instead of specifying
>> "EPSG:4326", and 2) recommend the upcoming OGC URNs over legacy EPSG:n
>> identifiers. So I'm -1 on draft version 5 becoming 1.0. I apologize for
>> not paying closer attention to the CRS section until this week.
> 
> I don't believe OGC URNs offer anything over EPSG identifiers, since
> there's no way to resolve them? Either way, it's just an opaque string
> that a client has to understand -- except in the case of URNs, fewer 
> clients understand them, and there's currently a dichotomy because
> the 'current' URN usage is all via x-ogc.  Am I wrong in some of these
> points somehow? 
> 
> Regards,

The EPSG database is proprietary and there is a legacy of confusion
around EPSG geographic coordinate systems. The OGC URN scheme fits Web
architecture, is non-proprietary, and designed to eliminate that
confusion. Carl Reed's IETF memo appears to be all but approved. Using
OGC CRS URNs is going to be the best practice in a few months. I believe
The GeoJSON spec should be forward-looking and recommend the best
identifiers for 2008-9-10, not the best identifiers for 2001.

Cheers,
Sean



More information about the Geojson mailing list