2009-08-12 09:38  Bill

	* common.c, freqbkgnd.c, freqbkgnd.h, parameters.c: removed command
	  line option for nmer background - now entered through prior file

2009-08-11 11:16  Bill

	* seq.weights.pl: perl program for calculating sequence weights

2009-08-10 14:19  Bill

	* common.c, gibbsmpi.c, gibbsmpi.h: gibbsmpi was sending data that
	  exceeded the MPI maximum buffer sizewhen using -bayes and 20
	  processors. Patched tosend in maximum 1MB blocks

2009-08-07 18:22  Bill

	* freqbkgnd.c: fixed possible unitialized variable

2009-08-07 17:10  Bill

	* species_sample.c, species_sample.h: sample by species

2009-04-23 14:43  Bill

	* NonGUI.inc, NonGUI.nompi.inc, centroid.c, centroid.h, column.c,
	  common.c, fels.c, fragment.c, gibbsmpi.c, impsamp.c, impsamp.h,
	  maximize.c, mem_mgmt.c, model.c, motif.c, nearopt.c, perror.c,
	  probability.c, probability.h, recur.c, rsample.c, sampling.c,
	  subopt.c: fixed problem with rdirichlet returning 0 probabilities

2008-11-06 09:14  Bill

	* centroid.c: fixed incorrect option on aling_centroid -m-y => -m
	  -y

2008-10-29 13:22  Bill

	* centroid.c: fixed problem with aling_centroid. It was using the
	  freq solution when a tree was inclusded in the prior.

2008-09-16 15:50  Bill

	* centroid.c, initseq.c: fixed problem in init_seq that was causing
	  program to incorrectly flag positions as not legitimate starting
	  positions

2008-07-05 16:04  Bill

	* prior.c: Added error message for sequences in MASS not being the
	  same length

2008-06-24 18:24  Bill

	* recur.h: cygwin update

2008-06-24 18:16  Bill

	* common.h: update for cygwin

2008-05-09 15:48  Bill

	* initseq.c, recur.c: fixed problem from Valeries Storms - clades
	  not being counted properly in phylogenetic sampler

2008-04-15 12:26  Bill

	* counts.c: added error message when nucleotide data is used
	  without -n

2008-01-16 16:59  Bill

	* INSTALL: Added note about MS compiler

2008-01-11 10:01  Bill

	* bayes.c, centroid.c, common.h, gibbsmpi.c, model.c: add motif
	  model to info saved for each sample by centroid samplerrr

2007-12-12 14:48  Bill

	* maximize.c: fixed probel when frequency solution was max. It was
	  sometimes not setting the background count properly.

2007-12-12 14:22  Bill

	* parameters.c: -l now sets -F

2007-11-21 09:15  Bill

	* bayes.c, common.c, common.h, gibbsmpi.c, hier.c, model.c,
	  model.h, parameters.c, rsample.c, rsample.h, sampling.c,
	  sampling.h: fixed CLOCKS_PER_SEC problem - was always 1000000

2007-11-19 12:46  Bill

	* NonGUI.inc, NonGUI.nompi.inc, bayes.c, centroid.c, centroid.h,
	  common.c, common.h, freqbkgnd.c, freqbkgnd.h, gibbsmpi.c,
	  gibbsmpi.h, hier.c, impsamp.c, make.mpi.opteron, make.opteron,
	  make.opteron.debug, mem_mgmt.h, model.c, model.h, prior.c,
	  prior.h, probability.c, probability.h, recur.c, rsample.c,
	  rsample.h, sampling.c, subopt.c, width.c, width.h: changed
	  CLK_TCK to CLOCKS_TO_SEC added n-mer background model added dump
	  of models

2007-08-24 17:57  Bill

	* phyllip.c: add '|' to list of legitimate characters for species
	  name

2007-08-16 10:50  Bill

	* centroid.c, parameters.c: made align_centroid search for version
	  with no .MPI at the end by default fixed minor problem with
	  option dispaly - it was showing credibility limits on npn-bayes
	  runs

2007-07-30 10:36  Bill

	* parameters.c: added error message for motif length == 1 and
	  fragmentation

2007-07-24 13:06  Bill

	* gibbsmpi.c, mem_mgmt.c: Fixed MPI problen where sample files were
	  being deleted before all processes completed

2007-07-20 10:19  Bill

	* fragadj.c: fixed column shift problem - was allowing negative
	  positions if site was at start of seq 0

2007-07-18 11:35  Bill

	* ChangeLog, Makefile, Makefile.am, README, configure: new build
	  7/17/07

2007-07-16 13:13  Bill

	* centroid.c, centroid.h, common.c: fixed problem in centroid
	  calculation. It was not taking all possible widths into account.
	  align_centroid now uses phylogenetic Gibbs to align sites

2007-07-16 13:11  Bill

	* fels.c: removed unnecessary normalization of substitution matrix
	  to speed up calculation

2007-07-16 13:10  Bill

	* prior.c: fixed normalization of BLOCKS parameter when number of
	  elemnets doesn' match -E parameter

2007-07-16 13:07  Bill

	* phyllip.c, phyllip.h: deleted unused routines

2007-07-16 13:06  Bill

	* initseq.c: allow program to continue even if it can't can't
	  initialize all motif sites

2007-07-16 13:04  Bill

	* rsample.c: fixed problem with zero probabilities in BLOCKS

2007-07-16 13:03  Bill

	* counts.c, probability.c: fixed problem with background
	  probability not summing to 1

2007-07-03 11:54  Bill

	* README: update readme

2007-07-02 11:32  Bill

	* README: updated README

2007-07-02 11:08  Bill

	* common.c, recur.c: updated version

2007-07-02 10:12  Bill

	* README: added info about phylogenetic sampler

Gibbs 3.03.003  2007-06-30
centroid.c
Fixed problem in centroid calculation. It was not taking all possible widths into account in some cases.
align_centroid now uses phylogenetic Gibbs to align sites.
fels.c
removed unnecessary normalization of substitution matrix to speed up calculation.
counts.c probability.c
fixed problem with background probability not summing to 1 in some cases
initseq.c
Gives warning instead of fatal error if it can't can't initialize all motif sites.
parameters.c
remove max iterations output for centroid mode
phyllip.c
deleted unused routines
prior.c recur.c rsample.c
fixed normalization of BLOCKS parameter when number of elemnets doesn' match -E parameter
