[OSRS-PROJ] projection precision

Gerald I. Evenden gerald.evenden at verizon.net
Fri Mar 14 11:51:30 EST 2003


Because there was a discussion about projection precision
related to the truncated conformal conic I wanted to know what
the reaction would be to increased projection precision in general.

Case in point: the routine to compute the distance along a
meridian to a specified latitude has always be "truncated"
and limited in precision.  Original GCTP wasn't much better
than a few millimeters and later PROJ has a 1x10^-5 meter
version.  In my closet for a number of years is a 1x10^-9
meter version (limit of Intel 64 bit double).  Now I have
an adaptable precision version good for any size double.

"Get to the point Evenden."

Ok, what is going to be the
reaction if a new release of UTM produced different numbers
in the sub-millimeter range than the previous versions?
Is this going to cause panic?

For the moment, initialization, forward and inverse calls
are the same as the old pj_enfn, pj_mlfn and pj_inv_mlfn
except that the returned double pointer of pj_enfn is now
a void structure pointer.

Note: the dumb names are carry overs from GCTP.


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



More information about the Proj mailing list