[Geojson] EPSG + Coordinate Ordering (Was Re: GeoJSON '1.0'?)

[Christopher Schmidt]

Members of the OSGeo standards list:

In discussion of the creation of the GeoJSON specification, we're
running into a question of whether you can refer to something as
EPSG:4326 and 'override' the coordinate order, insisting that it is
x,y[,z] regardless of what EPSG says. Apparently this came up before in
WMS 1.3.0 discussions: if anyone has more context, please respond to
this message on the osgeo-standards list  (I will summarize the contents
back to the GeoJSON list).

More context is available at
http://lists.geojson.org/pipermail/geojson-geojson.org/2008-March/000354.html

On Thu, Mar 13, 2008 at 09:36:55AM -0400, Daniel Morissette wrote:
> Christopher Schmidt wrote:
> > On Thu, Mar 13, 2008 at 07:43:59AM -0400, Panagiotis (Peter) A. Vretanos wrote:
> >> So if you use EPSG:4326 and are encoding the coordinates in x,y (or 
> >> long,lat) ... well, Raj said it best "You can't do that."!
> > 
> > *Why?* The answer is obviously not the same for GeoJSON as it is for
> > WMS: In WMS, there was no explicit statement that said "Ignore the CRS:
> > It's always in x, y." If the reason that there is a problem is because
> > the spec is poorly written, then that problem is *solved* with GeoJSON. 
> > 
> > Assuming that you override the ordering of the CRS in your spec: why
> > can't you use EPSG:4326?
> > 
> 
> I don't know if you're allowed by the EPSG license to refer to EPSG and 
> explicitly override the coordinate order... others such as Frank 
> Warmerdam seem to know that license fairly well and could comment on 
> that (see for instance 
> http://www.mail-archive.com/geotools-devel@lists.sourceforge.net/msg03279.html)

I know that most people participating in this particular spec writing
process are of the opinion that using the letters "EPSG" does not
somehow make us beholden to EPSG in any way. (Obviously, distributing
their data would change this.) 

> Anyway, back to the point, I 
> believe (but could be wrong since I was not there at the time) that an 
> attempt had been made by the WMS 1.3.0 RWG before I joined to simply 
> state that all CRS use x,y ordering as you suggest here, but later on, 
> as part of the ISO review process, it was pointed out that OGC could not 
> do that. What I remember of it is essentially that the RWG was told by 
> ISO that "if you refer to the EPSG database then you cannot pick and use 
> only portions of it, you must adhere to all of it, including coordinate 
> order. Period."

It seems like a weird requirement, and I hope that someone more
intimately familiar with that part of the process can comment on why
that might be. Certainly, to a layman, there's no material on the web
that indicates something like this, and as I pointed out above, EPSG is
just a set of letters, so trying to claim that it forces something
special on users of those letters seems... a bit odd, at best.

Hm, maybe someone from the OSGeo standards list can point me where to go
with this issue. I've added that list as a CC.

> I don't know if coordinate order will be an issue if you stay away from 
> ISO. Personally I still consider that this coordinate order issue ruined 
> the simplicity of WMS... if Geojson can avoid it that would be for the best.

There is no doubt in my mind that GeoJSON will be x,y[,z], always. The
only question is "How do we get there in a way that isn't shooting
ourselves in the foot?" Since it seems like part of the problem in this
case was ISO, maybe the answer is just "avoid ISO": since we're working
with a spec that's designed for content on the web, that doesn't seem
like a terribly painful decision to make to me, but maybe others have
problems with it. (IETF is a more natural home for GeoJSON than ISO, in
other words.)

Regards,
-- 
Christopher Schmidt
MetaCarta