<html>
<head>
<meta content="text/html; charset=ISO-8859-1"
http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#000000">
On 09/01/2012 15:52, Jean-Claude Repetto wrote:
<blockquote cite="mid:4F0AFF3A.5040601@free.fr" type="cite">Le
09/01/2012 14:24, Pepijn Van Eeckhoudt a écrit :
<br>
<blockquote type="cite">
different approach and defaults to +towgs84=0,0,0 instead.
<br>
</blockquote>
Blindly adding "+towgs84=0,0,0" will unlikely give correct
results. You'll have to provide the correct datum transformation
parameters.
<br>
</blockquote>
Thanks for the answers. You're absolutely right that this is highly
unlikely to be correct. I inherited the code in question and I'm
inclined this behavior now.<br>
<br>
I can understand why <a
href="http://trac.osgeo.org/proj/wiki/FAQ#WhydoIgetdifferentresultswith4.5.0and4.6.0">the
datum shift change</a> was made; the old behavior was basically
the same as adding towgs84=0,0,0. It's rather surprising then though
that gdalwarp (and cs2cs) don't complain in these cases. I would
expect the tools to refuse this type of input since you're asking
them to do transformations that they cannot perform correctly.<br>
<br>
Pepijn<br>
</body>
</html>