[Proj] WGS84 definition

[Martin Vermeer]

On Mon, Dec 05, 2005 at 06:17:24PM -0500, Frank Warmerdam wrote:
> On 12/5/05, Martin Vermeer <martin.vermeer at hut.fi> wrote:
> > I suppose you are right -- where the current user base of PROJ4 is
> > concerned. But I have been thinking if it would be possible to extend it
> > by simple means to also serve geodetic uses, as that is what interests
> > me. And the package has been moving from mere projections to geodetic
> > datums and transformations anyway.
> 
> Martin,
> 
> As Cliff suggests, I think it is wrong to think there is some deep
> attachment to geodetic concepts of what the different datums are.
> PROJ.4 basically knows about ellipsoids, projections, and
> "transformation methods" between datums.   When you say +datum=WGS84
> in PROJ.4 it translates this into an ellipsoid definition and recognises that
> to transform to WGS84 datum it can apply a 0,0,0 offset.  It doesn't really
> know anything about geoid, or anything else.
> 
> We can try and other kinds of datum transformations but ultimately
> a rigerous definition of datum does not existing in PROJ.4.  Only the
> implicit definition of what it takes to convert to WGS84 (whatever
> that is).

OK. If I understand this correctly, this means that if you try to
convert point set coordinates from datumA to datumB over "WGS84", is
doesn't matter what "WGS84" precisely is. Yes, I understand that notion.
But shouldn't then the PROJ4 definitions on datumA rep. datumB
implicitly agree to use _the same_ WGS84?

E.g. it appears (as I verified on the several-dm level) that the Finnish
KKJ definitions (epsg:2393 and around) transfer to EUREF-FIN, which is
in practice (?) the same as ETRF-89. However, Sweden for some time
used a national EUREF realization which differed by several cm from
ETRF-89. This gave rise to noticable (on geodetic accuracy :)
cross-border discrepancies.

BTW on an unrelated note, I believe geoid maps could be added using
something like the datum grid mechanism, if there is an interest.

And there is a difference between what _is_, and what _could be_ ;-)

Kind regards

Martin

-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
URL: <http://lists.osgeo.org/pipermail/proj/attachments/20051206/f93b5dd6/attachment.sig>