RFC 14 out for comment...

[Steve Lime]

I don't like the idea of treating this as a projection issue. I think it over complicates
things and will ultimately will lead to confusion. I'm kind of partial to extending the
use of TRANSFORM since it has always meant toggling between map and image
coordinates anyway. I think that would lead to little extra confusion, would mean
we wouldn't have to define yet another keyword and would be easy to document
and test.

Steve

>>> Frank Warmerdam <warmerdam at pobox.com> 4/11/2006 11:41:14 AM >>>
Steve Lime wrote:
> Sean: Good point. I thought about that too and  would agree except 
> that for the types of things folks want to do it might be more convenient 
> to define this at the feature level. I think it will be quite common to place 
> text in multiple corners of an output image. Granted, that could be handled 
> with multiple layer definitions. I do like the ORIGIN parameter regardless, 
> sounds better than RELATIVETO...
> 
> What do folks think about having to define multiple layers to define
> multiple origins?

Steve,

The other concern I have with treating this as a PROJECTION SRS type of
issue, which is what I assume Sean is getting at, is that currently the
projectionObj doesn't know anything about the raster coordinate system and
I am nervous about introducing this into the projectionObj in a way that
only really applies for layers.

However, this is just a general uneasiness, I haven't really thought it
through carefully.

Best regards,
-- 
---------------------------------------+--------------------------------------
I set the clouds in motion - turn up   | Frank Warmerdam, warmerdam at pobox.com 
light and sound - activate the windows | http://pobox.com/~warmerdam 
and watch the world go round - Rush    | President OSGF, http://osgeo.org