[mapserver-dev] Motion: Adopt RFC 41: Support of WCS 1.1.x Protocol

[Frank Warmerdam]

Daniel Morissette wrote:
> BTW, I like your solution for the axis order (that will do the trick for 
> WMS 1.3.0 as well), but are you planning to add +epsgaxis=... params to 
> all projections in the epsg file that comes with PROJ4 in a future 
> release,

Daniel,

No. In fact, my intention is that this would be essentially a mapserver
specific thing.  At some point in the future there may be an "epsg" file
like the currently one, and an "epsga" (EPSG Authentic or any other "a"
word implying it sticks stickly to the rules) file that includes keywords
describing the intended EPSG axis ordering.  But even then I won't use
the keyword epsgaxis.

So +epsgaxis will be a MapServer specific keyword which is intended to
mean "this is the axis order that EPSG things should apply".

 >  or will you hardcode the list of geographic coord. systems ids
> and add the +epsgaxis=... param in msLoadProjectionString()?

This RFC will modify msLoadProjectionString() to insert this keyword
*only* when expanding URNs in the EPSG space.  When we get to WMS we
will need to slightly extend things, perhaps with a mechanism for
the WMS 1.3 code to request the same thing as appropriate.

Note also that msAxisNormalizePoints() and msAxisDenormalizePoints()
need to be called in appropriate places.  For WCS this is done in a
few WCS and OWS specific places.

 >  If it's the
> latter then how reliable will that be? Are there chances of more 
> geographic coord. systems being added in the future and breaking WCS 
> services?

Currently the code in msLoadProjectionString() assumes that codes
in the 4000 to 5000 range will need this +epsgaxis keyword.  This
logic may need to be more sophisticated in the future if GCSes
are added outside that range or if a mixture of axes end up being
used.  It might even have to start looking stuff up in the EPSG
dataset itself.

But I'm willing to put that issue off till it arises, and at least
there is a fairly centralized place to add the logic.  We should
pretty much be able to grep for epsgaxis, and the normalize/denormalize
functions to get a sense of where axis stuff happens.

Best regards,
-- 
---------------------------------------+--------------------------------------
I set the clouds in motion - turn up   | Frank Warmerdam, warmerdam at pobox.com
light and sound - activate the windows | http://pobox.com/~warmerdam
and watch the world go round - Rush    | President OSGeo, http://osgeo.org