[GRASSLIST:4614] RE:Two questions: 1. r.param.scale; usage and etc... an error got into my last post!!

[Markus Neteler]

On Sat, Sep 14, 2002 at 12:12:33AM +0200, Lars Forseth wrote:
> On Tuesday 10 September 2002 08:58, you wrote:
> > Lars,
> >
> > I have been messing around with r.param.scale just recently and found the
> > following. You can obtain more spatially extended classes if you play with
> > the curvature tolerance. In my case, lowering the default c_tol value to
> > 0.001 improved the output image nicely. Is it meangingfull to do this ? I
> > don't know, I have not used those results yet hoping that somebody would
> > comment on them first. Perhaps Jo Wood would like to answer this.
> >
> 
> Hi!
> 
> Indeed, you were right, when I changed the c_tol parameter to 0.001 it gave 
> much more meaningfull output!

We have changed the default c_tol value now to 0.0001.

Thanks for pointing out the problem,

 Markus