[Gdal-dev] gdal/ogr/proj and NAD83(CSRS)

[Robert Coup]

On Wed, Jul 29, 2009 at 9:48 AM, Matt Wilkie <maphew at gmail.com> wrote:

> Hi Robert, thank you for contributing your insight, it helps!
>
> > cs2cs -v is quite clear about what it's doing, OGR not so much.
> >
> > Using the "straight" Proj EPSG defs on my system:
> > $ cs2cs -v +init=epsg:3579 +to +init=epsg:4617
> ...
> > No datum transform will happen  (ellipses match, and +towgs84 is ommitted
> (ie ==0)).
>
> Then I would say it is improper for ogr2ogr to write NAD83(CSRS) in
> the output .prj. Proj4 and dependant projects may think WGS84, NAD83
> and NAD83(CSRS) are synonymous but from my reading over the last
> couple of weeks it is clear that while the first two were
> interchangeable at one time, they are not now.
>
> In future a person or program using data defined as NAD83(CSRS) will
> reasonably expect sub-meter datum accuracy, for that is the purpose of
> the datum. if this is not the case, it should not call itself such.


The accuracy of any data is irrelevant to the projection it's in though. I
can digitize data from satellite images that are up to 100m off
horizontally, or from paper maps at 1:500K with cartographic license taken.
Whether I do it in NAD83(CSRS) or any other projection/datum - that doesn't
impact the data quality - that's based off where the information comes from
and what I've done to it.

Whenever i convert datums I will lose information, that's reality. People
tend to convert datums via magic commands and assume it's all perfect but it
needs to be pretty carefully understood what it means (which i think you now
do).

Proj4 just converts numbers. Where the numbers are defined from (EPSG
definitions or your own) and what they mean is a whole other kettle of fish.

Rob :)
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.osgeo.org/pipermail/gdal-dev/attachments/20090729/f536dd5a/attachment.html