[GRASSLIST:4727] GRASS 5 Dempster-Shafer modelling toolkit 1.02 available

[Benjamin Ducke]

Version 1.02 of the DST modelling tools for GRASS 5 is now available on
Michael Barton's web page:

http://www.public.asu.edu/~cmbarton/grass.htm

The toolkit can be used to create spatial predictive models with uncertainty
handling using Dempster-Shafer's Theory of Evidence.
Predictive maps are calculated by analysing the distributions of known sites
over landscape parameters encoded in raster maps and combining this 'evidence'
using Dempster's rule of combination.

For now, the programs can only be downloaded as a source code archive.
Instructions on how to compile and install them are included in the archive 
as well as full documentation in HTML format.
Everthing should work fine on any version of GRASS 5.X. Support for 5.7
is not very good at the moment as the old sites API is still used.
A "true" 5.7 version using vector points instead of sites is high up on
my priority list as are ready-to-install binaries.

If you have downloaded an earlier version of the DST toolkit or have
received one from me by email, please replace it with the new version
on Michael's page as it contains important bugfixes and enhanced
functionality as well as a lot of fixes for compiling on more recent
linux distros. I have not tried the programs on MacOS X or Win32/CygWin yet. 

You need the open source xml2 library (www.xmlsoft.org)
installed on your system (including the 'devel' stuff, i.e. C header files). 
Every Linux box has it. I think CygWin has it, as
well and so should MacOS X.
Try it and send me your reports if you like.

Thanks, Michael, for hosting the files.

Have fun,

Benjamin