[Proj] Troubles with Newton-Raphson inverse projections

[support.mn at elisanet.fi]

"Gerald I. Evenden" <geraldi.evenden at gmail.com> kirjoitti: 
> I fear those in this group that will expect type double machine
> precision. :-(

Something is always better than nothing. Why not implement
the "accepted level of error" parameter? And give some message
if that can not be achieved. At least they knew that they were not
getting double precision. Then it is their problem and no more
our problem. Hiding it is not a good idea, better to tell
somehow. They can tell the end users and it is then their problem
to know.

Or if that is difficult to implement, maybe some kind of zone system?
Near equator it would be the zone 1 and near poles maybe zone n
or something similar? Where the zone is a predefined area and the
numbering indicates precision achieved with that projection. There
could be some pictures, maybe with colors, that would indicate
the problem zones? Users should then know what are the limits.

Janne.