improper cvff

I just wanted to add one final post to this thread to wrap this discussion up.

I have fixed the bugs in moltemplate which were causing both OPLSAA
and AMBER/GAFF to generate improper interactions with atoms in the
wrong order.

I also fixed many bugs and added a few new examples and features. (I
am actually pretty confident now that the OPLSAA and AMBER
force-fields are behaving correctly.)

If you use moltemplate, please download the latest version of
moltemplate at www.moltemplate.org.

Cheers
Andrew