filtered out all machines with just one state as they are irrelevant and gave weird results for some other filters

added more command line options.
created analysis code for analyzing the results and stats gathered by
the main algorithm.

this was required because running compare results on the entire
benchmark set generates gigabytes of data.

line option to use it in addition to the regular code and compare the
results.
the results of the comparison get added to a different file that can be
changed through command line options.

wrapped all the hybrid ads code in the hybrid namespace and all the
complete ads code in the complete namespace to prevent ambiguity and
unintended interactions.

added an extra command line option to calculate a test suite using both
sets of code and then compare the results. WIP.

separating family as state identifiers. No testing has been done as of
yet.

created readme for batch file. added benchmark folder for use with batch file. created 1 extra benchmark.

static linking introduced errors on some pc's, so i included a dll file
in the bin directory instead.

This removes support for machine definitions that aren't fully
reachable.

removed finished todos.
removed unused code.
created batch script to run the algorithm on a set of files sequentialy.

created test method for checking validity of separating_family.
removed a bunch of old, unneeded code.

results still very buggy. further testing needed before use.

algorithm.
changed the old algorithm to the point where it can do everything it
used to do with the new datatypes.

instead of the internal integer labels.

Created a way to print the readable splitting tree to a file in an
organized way to compare it with expected results.