Re: coordinatesystem (was DOQ)

[Lothar Koppers <lothar.koppers@unibw-muenchen.de>]

hi geovrmlers,

the problem to solve is something very old:

The first question is which accuracy is needed to display geodata. There is no
common solution, because it depend on the problem. The answer in reality is
however, that the real used accuracy is mostly higher than realy needed. The second
question is therefore:  what accuracy is useful. (No one will  display a Sattelite
in a distance of 12000 km).

If we get a 'double' coordinatesystem (dc) and I implizid a large data volume (that
geo-like) the display performance will slow down very soon. I'd like to know, in
which apps such a high accuracy is needed. Therefore is a model to develop, in
which accuray geodata in vrml has to be diplayed. There are some difficulties to be
thougth about, for example to colors have an influence to this point and also the
detail structure of an object is relevant.

But what is when the aim dc is reached? Which dc should be used? gauss-krueger,
UTM, geocentric, geographic and what point of origin will be choosen. To generate
common geo-worlds, it is necessary to use a common coosys. But not all sys are
useful by different ways.

I for myself think it is useful to define an common extern proto for the use of
geoVRMLists to get a common entry to the VRML coordinate system.

In some days, my java package of geo-transformation will be ready (gk/utm ->
geographic -> geocentric 7 parameter Helmert transformation in 3d-coordinate and
the ways back). Maybe this will be the first step in this direction.


cu


Lothar