[Proj] Re: how to compare 2 proj4 definitions? or: how to normalizeproj4 definition?

[Ed McNierney]

Sergey -

I never generated any code - I simply worked on trying to define the
problem.  The problem with a user-configurable "epsilon" value is
whether that is what the creator of the definition intended.  We can
provide tools for the user to adjust, but it's hard for the user to
determine whether they did the right thing or simply succeeded in
forcing two DIFFERENT definitions together.  But having some kind of
tolerance value will certainly be needed if you want to consider 3.28
and 3.281 as both being the "same" multiplier when converting meters to
feet.

If possible, I think using the EPSG database is helpful.  If you can
determine that a given PROJ definition is congruent to some EPSG code,
and that another PROJ definition is also congruent to that EPSG code,
then you can conclude that they should be treated as the same
definition.

	- Ed

Ed McNierney
President and Chief Mapmaker
TopoZone.com / Maps a la carte, Inc.
73 Princeton Street, Suite 305
North Chelmsford, MA  01863
Phone: +1 (978) 251-4242
Fax: +1 (978) 251-1396
ed at topozone.com 

-----Original Message-----
From: proj-bounces at lists.maptools.org
[mailto:proj-bounces at lists.maptools.org] On Behalf Of Sergey Spiridonov
Sent: Thursday, March 01, 2007 7:36 AM
To: proj at lists.maptools.org
Subject: [Proj] Re: how to compare 2 proj4 definitions? or: how to
normalizeproj4 definition?

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi, Ed

Ed McNierney wrote:
> I made a mild effort at this task a few years ago and gave up.  I had

May be you still have some sources which I can use to start from? Even
if it is not perfect, it will be easier for me and can save some
resources.

> The chief difficulty is in determining whether two very similar 
> parameter values are supposed to be identical.  Is 0.9997 the same as 
> 0.999721, just rounded, or is it supposed to be a different number?

May be there it makes sense to have possibility to specify the level of
exactness of comparision for every group of parameters?

Like user-configurable "epsilon" 0.001 degree for lat/lon, 0.01 meter
for x_0/y_0 with some good default values?

Any other ideas?

PS Sorry if my English cause misundersanding...
- --
Best regards, Sergey Spiridonov
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFF5si0J3NEeIx4r3YRAqj2AJ9O6UzFCvHqYUVKNnr2xbIUTQjsFwCePql0
9lT+zTKVMJdDbqkdOSj8aVc=
=8bMq
-----END PGP SIGNATURE-----

_______________________________________________
Proj mailing list
Proj at lists.maptools.org
http://lists.maptools.org/mailman/listinfo/proj