I've just finished reading a past thread on r.cost execution time (see GrassList #7781) and I was wondering if r.cost suports negative DEM values now/yet? This would be very helpful for cases where bathymetry DEM's are used, for instance - all z values are always negative. ~ Eric.