[OSRS-PROJ] Orthographic Projections and MapServer

[Gerald I. Evenden]

On Fri, 2003-08-22 at 04:16, Ron Russell wrote:
	...
> For example it would be good to be able to
> define the limits of a Transverse Mercator projection to
> avoid the sillyness we have had of people trying to display
> New Zealand in the British National Grid! ("The projected
> space was available, so why not?"  was the response I
> got!).

LOL.  I've seen the problem many times.  A frequent problem
for me were people who wanted to extend UTM across 10° of
longitude and expect the portions of the map to overlay
AMS sheets.

At one time I thought about putting a limit on the UTM
projection and reject points that were beyond the 3½° limit
but decided that I didn't want to listen to the bitching
and screaming from the users.  Let'em hang themselves.  ;-)
There's also the N/S limits.

> More seriously, we currently have a project that uses
> world data as a background. There are two huge polygons
> in this data (The American continent and Europe, Asia and
> Africa). When this is displayed in British National Grid parts
> of what I assume to be Siberia/Alaska appear in Europe
> (presumably an artifact of the clipping of coordinates
> approaching infinity).

You certainly do have problems.  What graphic package are you
using?

> I assume that these limits would have to be stored in terms of
> Geographical Coordinates which leads to problems of deciding
> their granularity. (I am assuming that the projection is independent
> of the data; it certainly is in our systems).
> 
> Clipping to the limits would be done before projection using
> spherical/ellipsoidal geometry. This seems
> horrific, particularly for polygons, but then computing power is
> becoming cheaper!

Two levels of clipping is certainly reasonable: first in the
geographic domain and secondly in the Cartesian domain.  I do
exactly that in my graphic applications.  Normally, clipping
in the geographic domain is not that computational intense---
about the same as Cartesian clipping.  But clipping in the
geographic domain is normally only done done on 180° delta
longitude from CM "wrap around."

> Finally, in my experience, clipping of polygons cannot be done
> on the fly. You have to bite the bullet and do it all at once. Our
> polygons contain holes (like the Great Lakes and the Caspian
> Sea).

The hole problem need not be done in the geographic domain but
deferred to the Cartesian level where standard procedures compensate
for data holes.  The linkage between the hole and outling vector sets
does need to be maintained.  See the PostScript manual for comments
about how they handle holes in fills.

> ---------------------------------------------------------------------
> Ron Russell
> Senior Software Engineer, Laser-Scan
> +44(0)1223 420414
> 
> http://www.laser-scan.com
	...
-- 
Gerald I. Evenden <gerald.evenden at verizon.net>

----------------------------------------
PROJ.4 Discussion List
See http://www.remotesensing.org/proj for subscription, unsubscription
and other information.