[josm-dev] WMTS scales - proper definition of 1m along meridian

[Paul Hartmann]

On 05.07.2015 23:13, Wiktor Niesiobedzki wrote:
> After more twiddling with WTMS I found some other challenge regarding
> EPGS - natural coordinate order, examples:
> 4326  - Axes: latitude, longitude. Orientations: north, east
> 2180  - Axes: northing, easting (x,y). Orientations: north, east
> 3857  - Axes: easting, northing (X,Y). Orientations: east, north
> 31370 - Axes: easting, northing (X,Y). Orientations: east, north.
>
> Those above are taken from:
> http://www.epsg-registry.org/
>
> As you can see, there are some projections, where coordinates are
> switched, and some, when they are not.

This is the same for WMS version 1.3 and later.

As JOSM supports both WMS 1.1 and WMS 1.3, this means our implementation 
of WMS 1.3 is as incomplete / broken as WMTS in this regard.

I just assumed that EPSG:4326 is the only relevant case where the axes 
are switched. (You can use the code "CRS:84" to get the same 
"projection", but with longitude first and latitude second.)

Apparently I was wrong and EPSG:2180 is at least one more exception.

> My guess is maybe this could be
> defined by setting axes=neu or axes=enu, as described in proj.4

Sounds good.

> But I do not see this extensively used in proj.4/nad/epsg
> (https://github.com/OSGeo/proj.4/blob/master/nad/epsg) - only 28
> entries with wsu value.

Strange... It would be interesting to see how other projects are 
handling this, e.g. QGIS and GDAL.

> Do you think that we can induce this value from some other information
> from our epsg file, or shall I add this information to our projection
> definition?

I don't think this information can be derived from anything that is 
already known, but we need to add it explicitly.

Best, Paul