Indeed rasters are key - but rasters are one kind of data structure to
represent a coverage - nothing that I said had anything to do with
eliminating rasters - just one should not equate coverage with raster -
it is one kind of coverage representation (and a very important one).
Note that one could also represent the geometry of so called
conventional features like roads by creating a coverage which is the
characteristic function of the road - e.g. set all points on the road =1
and set all points off the road = 0 - this is a "where the road is"
coverage.
R
-----Original Message-----
From: Peter Baumann [mailto:p.baumann@xxxxxxxxxxxxxxxxxxxx]
Sent: October 7, 2008 3:01 PM
To: Ron Lake
Cc: Wenli Yang; Unidata GALEON
Subject: Re: [galeon] Features and Coverages
weird roads you have over there ;-)
Probably one operative word in Wenli's explanation was "...the *main*
reason...".
Discussion over the last days IMHO nicely illustrates that we are not
really done with rasters, while we already are contemplating
generalizations.
Personally, I enjoy this discussion a lot.
nite,
Peter