[Geojson] coordinate order

Dale Lutz Dale.Lutz at safe.com
Sat Mar 8 22:00:27 PST 2008


For the record, I too think this is a great change.  I really like the way you've kept the "x is x" and still allowed those that really want swapped coordinate systems to eat their cake too -- they just need to be explicit about it.

FME indeed does support GeoJSON, and in fact at our user conference yesterday we were showing FME Server spitting GeoJSON into OpenLayers (Peter mentions this in his blog here
http://geothought.blogspot.com/2008/03/fme-user-conference-report.html).  FME 2008 is finalized now so it won't be able to correctly read any data that has the x/y flipped, but we'll get that into the next FME betas very shortly, and we don't write flipped x/y with FME 2008 so whatever we produce will still work fine as well.

I'll ask around at Safe to see if there's anything that we'd add to the current draft proposal, but our guys had already gone through things carefully and I fully expect that we'd heartily support calling the proposal a 1.0 as well.

Good job.


Dale Lutz              Safe Software Inc.           dale.lutz at safe.com
VP Development         Surrey, BC, CANADA        phone: (604) 501-9985
                       http://www.safe.com         fax: (604) 501-9965

> On Fri, Mar 07, 2008 at 11:34:25AM -0700, Tim Schaub wrote:
> > Makes things a bit more complex for clients who know about CRS and
> > simpler for the rest.
> Agreed. I like the change. I never liked the absolute restriction on
> EPSG identifiers, and this puts the onus on people who want to use them
> to do so such that the meta-information in the CRS object describes what
> clients need to know, and allows for clients who need more to get what
> they want.
> Importantly, No existing data will be made invalid by the changes we've
> made.
> > I think this thing is ready to let out into the wild.
> Agreed. Currently we have GeoJSON in use by:
> Libaries:
>  * OpenLayers
>  * OGR (http://gdal.org/ogr/drv_geojson.html)
>  * MapFish Java implementation?
>    https://trac.mapfish.org/trac/mapfish/ticket/73
>  * GeoJSON Python Library, http://pypi.python.org/pypi/GeoJSON/1.0a3
>  * Shapely (GeoJSON Geometries)
> Servers:
>  * FeatureServer
>  * GeoServer
>  * WPServer
>  * Pleiades?
>  * Knowhere? I Think.
> Desktop Apps:
>  * FME
>  * Cadcorp SIS (I hear? can't find a link)
>  * Arc2Earth, according to http://tinyurl.com/2oo5o8
> Web Services:
>  * geo:truc, http://www.geotruc.net/
>  * FireEagle, I hear?
> Are there more that I'm forgetting?  I googled for a while, but couldn't
> find anything else obvious.
> Is there any reason not to 'go to 1.0' with the current specification?

More information about the GeoJSON mailing list