[Proj] "Double Ellipsoid" error, reproduction

[Gerald I. Evenden]

On Wednesday 10 December 2008 1:00:16 pm Duncan Agnew wrote:
> Since proj doesn't know where its coordinates come from, +proj=merc
> will do the spherical Mercator as it
> stands (I assume)--no need for a new flag. I suppose this illustrates
> Gerald's concerns about confusion resulting from the mixing of datums
> and projections in one routine; I'll say, as a user, that it is
> convenient to be able to combine datum shifts and grid conversions in
> one routine, though it would be
> equally convenient to separate them and use pipes. The usual problem of
> path-dependence: see the QWERTY
> keyboard and the 4'8.5" railway gauge.
>
> I'll disagree with Gerald about Google Maps--their Mercator is an
> excellent choice for their purpose. It is a terrible choice for a
> global projection of a static map, but an excellent one at large
> scales, which is where all the activity on Google Maps ends up.

I think you have your "large scale" definition mixed up---
	"large scale"= 1:10,000
	"small scale"= 1:1,000,000
1/10,000 > 1/1,000,000
The size of the denominator is inverse the verbal approximation.  Common 
error.

As for large scale (1:10,000) why not use equirectangular---cheap 
computationally and who's to know on a computer screen?

> And for 
> doing preliminary locations, a familiar cylindrical projection has
> merit even at a global scale. I do agree with Gerald that for a paper
> global map, equal-area should be the default choice, absent some other
> need.

For thematic mapping---equal area is always the choice.

> Duncan Agnew
>

-- 
The whole religious complexion of the modern world is due
to the absence from Jerusalem of a lunatic asylum.
-- Havelock Ellis (1859-1939) British psychologist