If you are working with protein structures, or any instance of the problem with a discrete reformulation, then MD-Jeep is a good choice. I would be delighted if you convince yourself more by knowing more about the software at:

https://team.inria.fr/genscale/protein-structure/md-jeep-distance-geomtry-solver/

However, here is one you may use:

http://trac.osgeo.org/geos/ (in C++)

... and I would insist you visit http://www.cgal.org/

Cheerio!

Hi David,

can Complearn be useful to your case?

Have a look here:

http://www.complearn.org/

The whole toolkit is free to download and makes use of a Normalized Compression Distance:

http://www.complearn.org/ncd.html

among objects, which are then visualized in a plane using a quartet tree heuristic for hierarchical clustering (a randomized hill climbing); e.g. see here:

http://arxiv.org/abs/cs/0606048

Here is an example of what you may get:

http://www.complearn.org/images/34mammals-unrooted.png

Ciao,

Sergio