[IMP-dev] More small patches

[Daniel Russel <drussel@gmail.com>]

First, there were some incorrect indexes in MD (not a problem in  
practice). Second I loosed the tolerances on one of the tests since it  
was often outside the tolerance if anything changed with the state of  
the random number generator.

Attachment: md_patches
Description: Binary data

I added a call to seed the random number generator with the current  
time and a comment about how to set the seed if you prefer something  
else.

Attachment: seed_random
Description: Binary data