Matthew Turk
May 2008

This code (described below) has been modified to be wrapped as a shared library
callable from Python, as a part of the yt toolkit.

Stephen Skory
May/June 2007

This is a new implementation of hop for enzo datasets, to replace the
fragile 'enzohop.' enzohop uses the enzo grid functionality to extract
the particle data from the HDF5 datasets. newhop uses plain HDF5 C++ calls
to extract the data, which is then fed into the hop mechanism. As far as I
know, this version is fine with 64 bit integers/floats, which enzohop isn't.

There are a few versions of newhop which build on datastar just fine. I 
haven't tested it on other machines. The default build 'newhop' is for
packed datasets and will include both stars and dm in the grouping.
THe other versions are for non-packed datasets, or if you want to only
consider dm particles for grouping. Hop doesn't like datasets with too
many particles, (I've never done tests, but I know that 20 million
particles give hop problems) so sometimes dm-only is the only way to go.
