Moritz wrote: > Martin Album Ytre-Eide wrote: > > I think Arc& copycats use class breaks so much in their > > legends simply as a matter of ...
(fwiw, those were my words, not Martin's) > Definitely an interesting debate. But even the author of > "How to Lie with Maps" does not put into question the use of > categories in maps... Perhaps not, but I certainly do! The sensitivity in selection of number of histogram breaks when your "n" is small or not purely a normal distribution is hardly a fringe theory or novel insight though, it's basic statistics. I'm not saying don't use category breaks if you want to, just that if you do, be aware of what you're doing and perhaps do some sensitivity trials with breaks set at different places and see if your interpretation is the same. It's only since I started doing those that I've really been aware how important it is. As earlier mentioned my current favorite method is to have contour lines (ie category breaks) drawn over the top of a continuous gradient background. Then you get the benefits of both ways. regards, Hamish _______________________________________________ grass-user mailing list grass-user@lists.osgeo.org http://lists.osgeo.org/mailman/listinfo/grass-user
