[Proj] "Double Ellipsoid" error, reproduction

[Richard Greenwood]

On Sun, Dec 7, 2008 at 8:45 PM, Gerald I. Evenden
<geraldi.evenden at gmail.com> wrote:
> On Sunday 07 December 2008 9:59:11 pm Frank Warmerdam wrote:
>> Gerald I. Evenden wrote:
>> > I find this thread so incredible confusing I have tried to stay out of it
>        ...
>> > If the issue is simply two completely separate operations--- datum
>> > conversion and ellipsoid-projection--- then what is all the discussion
>> > about???
>>
>> This is why I believe it was suggested we should be able (potentially)
>> to use different ellipsoid values for the projection calculations and
>> for the datum shift.
>
> I do not think we are in any disagreement.  Diagramatically I look at the
> whole problem simply as:
>
> original map data
> with known projection
> and projection ellipsoid
> (it might even be a sphere)
>              |
>              V
> inv. project with above
> ellipsoid and proj param.
>               |
>               V
> geographic data
>               |
>               V
> datum shift with whatever
> info it takes.
>               |
>               V
> geographic data in terms of new datum
>               |
>               V
> project to new map with
> parameters and proj ellipsoid
> of choice (maybe a sphere)
>
> The nature of the projection at either end of the above operation is
> completely arbitrary as is the ellipsoid and/or sphere to which it is
> projected.  The user *only* has to know how the original data was projected
> and its parameters and record for posterity how he/she created the new map.
>
> This is why I looked at cs2cs as simply the following procedure:
>
> proj4 -I <params> | datum_convert <params> | proj4 <params>

That is essentially what cs2cs does. I believe the problem is that
there is no mechanism to provide different parameters for the
projection and the datum shift. The coordinate system is most often
defined by EPSG code, and I have not seen EPSG codes that distinguish
between projection ellipsoid and datum ellipsoid.

-- 
Richard Greenwood
richard.greenwood at gmail.com
www.greenwoodmap.com