First the semantics - Lat/Long isn't a 'projection'. A projection is
when you cast (project) a spherical object onto a plane. A projection
is generally an equation that converts between shperical coordinates
and planar coordinates. The sphere should be referenced to a datum,
which for the sake of discussion, can be thought of as the sphere's
origin in space.
When you define both a projection and a datum, you have a 'coordinate
system'. In other words, you have defined how to get from planar
coordinates to shperical coordinates, and you have defined where the
sphere is in space (the datum). I'm simplifying a bit, but I think
this simplified view is sufficient to answer the question.
If I want to convert a State Plane NAD27 coordinate to a UTM NAD83
coordinate three steps would be required:
1. convert the planar State Plane coordiante to Lat/Long.
2. move the Lat/Long from NAD27 datum to NAD83 datum.
3. convert the NAD83 LAT/Long to UTM.
MapInfo's basic 'Lat/Long' isn't associated with any datum, as you
have noticed. So if you convert a MapInfo Lat/Long to UTM NAD83 it
does not attempt to convert the datum (step 2 above). And it you
convert the same Lat/Long value to State Plane NAD27 it still doesn't
attempt to transform the datum. Which is why you referred to it as a
'chameleon projection'. Unfortunately, it isn't actually changing to
match the target datum, it's not changing at all. So I'd call it
'color blind'.
Your data is most surely associated with some datum. You need to find
out what it is and then use MapInfo's "save as" to save the generic
Lat/Long to Lat/Long xyz, where xyz is the correct datum. UT is making
a possibly incorrect assumption in casting a generic MapInfo Lat/Long
to a WGS84 Lat/Long. It has no way of knowing what datum your data is
on. Additionally, UT does not reproject data, nor do datum
transformations. UT converts between data formats, not coordinate
systems.
And a caveat: MapInfo's support for datums is pretty thin in this day
and age. It's not really adequate for sub-meter mapping. But that's a
topic for another thread.
Rich
--
Richard Greenwood
richard....@gmail.com
www.greenwoodmap.com
Best regards / Med venlig hilsen Lars I. Nielsen GIS & DB Integrator GisPro