Hi all.
A new version of r.li (0.2.6) is available at the usual site: http://www.faunalia.it/download/r_li/
A bug preventing floating points calculations has been discovered and fixed.
First test on the comparison with r.le give very encouraging results:
numeric differences between the two programs are *very* small,
presumably due to rounding effects. A map of the differences is available: http://www.faunalia.it/download/r_li/shannon.png is the resulting map
(Shannon index) http://www.faunalia.it/download/r_li/diff.png is the map of the differences.
All the best.
pc
- --
Paolo Cavallini
email+jabber: cavallini@faunalia.it
www.faunalia.it
Piazza Garibaldi 5 - 56025 Pontedera (PI), Italy Tel: (+39)348-3801953
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

A new version of r.li (0.2.6) is available at the usual site: http://www.faunalia.it/download/r_li/
A bug preventing floating points calculations has been discovered and
fixed. First test on the comparison with r.le give very encouraging
results: numeric differences between the two programs are *very*
small, presumably due to rounding effects. A map of the differences is
available: http://www.faunalia.it/download/r_li/shannon.png is the
resulting map (Shannon index) http://www.faunalia.it/download/r_li/diff.png is the map of the
differences. All the best.

Nice.

what's the difference in running times?
G63> time r.le.patch map=rastermap
etc

when will this be ready for 6.3-cvs? (if that's your intention)

No speed test for now; theoretically it should be *much* faster (earlier
rough tests gave a 4x improvements), as it is multithreaded.
The main advantage is that the code is much easier to maintain, and
adding/modifying indices is easier (have you had a look to the code?).
For me it is ready for 6.3; only a few indices are implemented now, but
those who are should be ok. I would vote for having it included in trunk.
pc

Hamish ha scritto:

Paolo Cavallini wrote:

A new version of r.li (0.2.6) is available at the usual site: http://www.faunalia.it/download/r_li/
A bug preventing floating points calculations has been discovered and
fixed. First test on the comparison with r.le give very encouraging
results: numeric differences between the two programs are *very*
small, presumably due to rounding effects. A map of the differences is
available: http://www.faunalia.it/download/r_li/shannon.png is the
resulting map (Shannon index) http://www.faunalia.it/download/r_li/diff.png is the map of the
differences. All the best.

Nice.

what's the difference in running times?
G63> time r.le.patch map=rastermap
etc

when will this be ready for 6.3-cvs? (if that's your intention)