Thanx, I am just now working on the viewpoint problem which is really severe as they have a tendency to overlap :-( So I try to develop something like a visual patch on which i can apply a kind of FRAGSTATS or patch analyst method. Alasdair Turner wrote: >> is there anybody working with space syntax in the open land ie. trying >>to describe spatial composition and configuration ? >> >> > >David O'Sullivan and I had a go at using VGA on landscapes. We didn't >really manage to conclude that much from it: VGA integration is all but >meaningless (extending any area outwards may find another peak, and any >peak drastically increases VGA integration). VGA clustering coefficient is >quite interesting as it picks out troughs and ridges -- but: a) this has to >be taken in the light of a huge amount of research into viewshed analysis, >and b) is it space syntax? > >See: http://www.geog.psu.edu/faculty/osullivan/index.html > > > >