[Proj] Programmatically obtaining the domain of a projection

[Hamish]

Phil wrote:
> I've got a case where I would like to identify the domain of
> a projection programmatically. Take the geos projection for
> example:

others here will have better insights than me, but I'd consider
that many of the mathematical functions involved are continuous,
and without limit. Others will of course blow up once you get too
far away/past 90 deg lat. But just because others don't
immediately blow up, doesn't mean they are still useful far away.
The best you could I would think would be to define some
arbitrary limit of distortion and then solve for the boundary
value that gives you that, or decide on some arbitrary distance
cut off from a regional named CRS.

As different people in different fields of study will have
different requirements for maximum allowable distortion, setting
a fixed geographical limit on it will always be a compromise.


aka, how long is a piece of string?


best,
Hamish