todo.txt

add more checks to arg parsing

start thinking about how to combine alignments from the Pareto front
to create local ones. Should try to take from the best representatives of
the whole front-- might want to look at distance in alignment space as well
as objective space to do so.

add logging of mean and max over time for plotting purposes

make sure asserts are turned off in optimized version

write tests for v1Unaligned being correct after all modifying ops.

consider abandoning size optimization if total alignment optimization
performance continues to be much better. Instead, might want to do partiality
as a post-processing step of extracting high-confidence sub-alignments.

try ACO. 

add the ability to specify a list of fitnesses to dominate as a stopping condition.
Maybe go another 10000 generations after dominating the last one.

add seed-and-extend with random restart

add an alternative crowding distance calculator that uses alignment space distances

add the ability to load all alignments in a given directory as the initial population.
Just do a non-dominated sort on them after loading them, then take popsize many of them
as the first population.

try getting rid of edge set in Network class. Not really used anymore, except in slow ICS, which
is used to initialize conserved count.

figure out why best s3denom is always reported as 0

add swapping alnMask values to local search when partial alignment is enabled

REFACTORING
-----------
clean up the Makefile. Currently hideous.