[geos-devel] GEOS Report (October 29, 2002)

Yury A. Bychkov ybychkov at direct.ca
Tue Oct 29 12:43:04 EST 2002

Hi Everyone,

Current status:
C++ version still runs slower then Java, though not by much now (about 2.5 vs
3.5 seconds for 100 tests). It seems that this happens due to a number of small
inefficiences and not a bug somewhere in the computeIntersections() algorithm.

Thisngs to do:
1. Test the code on a VERY BIG starfish-shaped geometry (10000 nodes or so) to
check the assertion above.
2. Streamline GEOS XML tester a bit to make it more command-line friendly.
3. Do some performance tuning and try to det rid of some of the inefficiences.


More information about the geos-devel mailing list