[GeoJSON] spec clarification of CRS

[Matt McClure]

On Wed, Dec 11, 2013 at 3:42 PM, Martin Daly <Martin.Daly at cadcorp.com>wrote:

> On Wed, Dec 11, 2013 at 3:38 PM, Raj Singh <raj at rajsingh.org> wrote:

> I'd like to suggest that in section 3.1 Named CRS [1], a sentence be added
> > that says, "EPSG:4326 identifiers such as EPSG:4326 and
> > urn:ogc:def:crs:EPSG:6.6:4326 are prohibited, as they define coordinate
> > reference systems with Y-X axis order, which is illegal according to
> section
> > 2.1.1."
> > (a good reference on the issue is [2]).
>
>

> As we move into 2014, confusion still reigns on this issue. And the
> current
> > wording of the GeoJSON spec implies that EPSG:4326 is OK, which is an
> > internally contradictory statement. Don't make the same mistake OGC did a
> > decade ago!
>
> I, for one, completely disagree. GeoJSON is explicit and consistent about
> axis order of the coordinates. All that it says is " use any CRS you like,
> but *always transfer the coordinates in this pre-defined order*".


> No mistakes, no confusion. At least on the part of GeoJSON.
>

To show one data point that confusion exists, I'll relate my own experience:

As a relative newcomer to GeoJSON, one of the first things I wanted to do
was see if I could use it to model time series data. I was thinking that
I'd use a custom CRS where each coordinate would contain a time stamp and
some number of additional data samples.[1] Examples:

time, longitude, latitude
time, longitude, latitude, elevation
time, heart rate
...

Granted, there is a bullet in section 3, Coordinate Reference System
Objects, that says, "CRS shall not change coordinate ordering". But it
would help to have a clarifying comment that indicates, for example that
(x, y) CRSs and CRSs where the first three dimensions of a coordinate are
not (longitude, latitude, altitude (optional)), in that order, are
non-conformant.

On the other hand, what I'd really like is a similar specification and
implementations that allow arbitrary CRSs without qualification so that I
could implement the examples I described.

Best,
Matt

[1]:
http://lists.geojson.org/pipermail/geojson-geojson.org/2013-September/000832.html

-- 
Matt McClure
http://matthewlmcclure.com
http://www.mapmyfitness.com/profile/matthewlmcclure
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.geojson.org/pipermail/geojson-geojson.org/attachments/20131211/96ed3ed8/attachment.htm>