[Mapserver-users] WMS BBOX vs. Actual Bounds

[Ed McNierney]

Ian -

If you have different scales in X and Y directions, the projection
should reflect that.  If the projection is not defined to have different
scales, then asking for different scales is not consistent.

If I ask for a 1,000-meter by 1,000-meter area in UTM projection in an
image that's 300 pixels by 200 pixels, it's not possible to do all
three.  If I keep the area 1,000x1,000 and stretch it to fit in 300x200
pixels, then the projection is no longer UTM.  If I keep the area
1,000x1,000 and the projection UTM, I have to modify the image size and
return something square - 300x300, for example.  If I keep the
projection UTM and the output image 300x200, then I have to modify the
map area to something 1.5:1 - 1,500x1,000 meters, for example.
MapServer, of course, chooses the last option, but I have a hard time
seeing how any one of the three choices is the "right" answer.

If you're creating output for a device with non-square pixels, it seems
that the way to handle that would be to provide an additional input
parameter to the WMS interface, describing the aspect ratio of the the
pixels.  This would still allow a client to make an inconsistent
request, but it would allow requests intended for non-square pixels to
be made consistent with the specified projection.

However, you are absolutely correct that this is NOT how the WMS
specification handles it.  The specification does require the server to
stretch the image to fit the box, thereby presuming that any
inconsistency in the request is due to a careful client requesting an
image for non-square pixels, rather than an error.

I suppose support for non-rectangular pixels can be left for another
discussion <g>.

	- Ed

Ed McNierney
President and Chief Mapmaker
TopoZone.com
ed at topozone.com


-----Original Message-----
From: Ian Turton [mailto:ian at geography.leeds.ac.uk] 
Sent: Thursday, January 08, 2004 10:56 AM
To: Ed McNierney; mapserver-users at lists.gis.umn.edu
Subject: RE: [Mapserver-users] WMS BBOX vs. Actual Bounds

At 09:20 08/01/2004 -0500, Ed McNierney wrote:
>Ian -
>
>How is it possible to honor all parameters when they are not
consistent?

You have different scales in the X and Y directions.

>I certainly understand the need to deal with unusual output devices as 
>you describe, but it is not very difficult for the client to address 
>that situation.

It is difficult for the client to deal with if it asks for the image to
be distorted but the one it gets back isn't. Mapserver should respect
the WMS spec on this point (at least in wms mode).

Ian

Ian Turton, Director, Centre for Computational Geography, University of
Leeds, Leeds, LS2 9JT http://www.geog.leeds.ac.uk/people/i.turton
http://www.ccg.leeds.ac.uk http://www.geotools.org
+44 (0) 113 343 3392 fax: +44 (0) 113 343 3308