WanMil wrote
I wonder if it is less important to find big enough areas with this algorithm. The old problem are white stripes which are ugly. Using the described new algorithm there is also a chance that some cuttings are too small. But the probability of a small aspect ratio is much higher than in the old algorithm. Having a smaller aspect ratio the little parts that are missing are also smaller. So I think a small aspect ratio might be a kind of indicator how visible the missing parts are. Maybe it is worthy to try that out. (I am sure at least this algorithm is better than the existing :-)
I agree that it would be better. I am still fighting with the precision (= performance) problems. When that is solved I try to find a algorithm that really cuts. Ciao, Gerd -- View this message in context: http://gis.19327.n5.nabble.com/shapes-with-holes-tp5745424p5747099.html Sent from the Mkgmap Development mailing list archive at Nabble.com.