2010-11-05
	* Begun version 0.7-1.
	* Added 'remove.unmatchables' option to pairmatch(); separated out
	tests of pairmatch().
2010-07-15
	* Simple mean fill-in for missing values, with missingness indicators
	(see fill.NAs).
	* bug fixes on mdist.function method: real opmatch.dlist object, labels
	preserved.
2010-01-26
	* Support for propensity models fitted using bigglm (new method for mdist).
	* Bug fix for maxControlsCap.
2009-05-18
	* revised summary.optmatch method.  
	* objects of class "optmatch" now carry matched.distances information, as attribute "matched.distances".  This is created by a call to matched.distances(), with default setting on preserve.unit.names.
	* summary.optmatch calls xBalance(), in package RItools, if RItools is loaded and it is given a second argument of class "glm"
2008-07-24
	* stratumStructure now doesn't include strata in which matching failed when calculation comparable number of matched pairs.
2008-07-01
	* fixed bug in [.optmatch (failed with recent versions of R)
2008-05-19
	* introduced optmatch method for subsetting, so you can retain key metadata (specifically contrast.groups attribute) when you subset an vector produced by fullmatch() or pairmatch(). 
	* added calculation of effective sample size to stratumStructure, with documentation.
2008-05-14
	* Mahal.dist now fails (with an error message) if you give it a factor Tx-vble (w/o also giving it an inverted covariance)
2008-05-09
	* Addressed bug in fullmatch() whereby it dies when given distance of mode logical.  
	* Started test file fullmatch.R 
	* Update to mahal.dist.Rd -- hint on how to trick it into doing scalar discrepancies
	* version 0.3-7
2008-05-09
	* Bug fix to mahal.dist(): it wasn't accepting an inverse.cov argument.  Also added a test of this (which I informally verified).
	* Added nuclearplants data frame to /data.  It is very similar to the nuclear data set in the boot package: see its docs.
2008-02-29
	* Bug fix to mahal.dist() : got it to accept formulas w/ ns() in them
2007-10 
	* New license.
	* More informative error handling for stratumStructure()
	* Updated maxControlsCap, minControlsCap to reflect v0.3 changes to how unmatched units represented.
	* Change to loop within fullmatch() to avoid stacking of evaluation frames in memory
	* Added some tests
2007-08-20
	* Bug fix in pscore.dist (accommodating glm's that had a subset argument)
	* Small bug fix in stratumStructure
	* Bug fix in Ops.optmatch.dlist
2007-07-23
	* Bug fixes in Ops.optmatch.dlist, having to do with handling of lists containing a dim 0xn or nx0 matrix.
	* Bug fix in mahal.dist, having to do with names of icv not matching those of the data frame after data frame names are checked and modified by make.names (called by data.frame()).
2007-07-17
	* Adjusted handling of frames in makedist, pscore.dist, mahal.dist
2007-07-15
	* Small change to pscore.dist -- error changed to warning; 
	bug fixes re use of structure.fmla in pscore.dist and mahal.dist
2007-07-09
	* Removed tabs from relax4s.f
2007-07-08
	* Fixed a bug in mahal.dist
	* reorganized /src to pass checks
2007-06-20
	* v0.3
	* Changed unmatched indicator to NA in fullmatch output 
	(line 188 of fullmatch200706.R)
	* Small changes to stratumStructure() to adapt to prev change
	* Added mahaldist, propensitydist functions
	* Various updates to help pages
	
2007-05-10
	* Added optmatch.dlist method for abs
2007-01-06
	* fixed small problem in fullmatch affecting stratumStructure
2006-11-14
	* stratumStructure function
2006-10-31
	* pairmatch function
2006-10-21
	* Introduced namespace, experimental Ops.optmatch.dlist method
2006-06-04
	* New function makedist().  (Renaming of distlist(), really.)
	Vignette mahalanobisMatching updated and listOfDistances vignette
	removed (as it is now obsolete).  Documentation of makedist()
	contains a propensity score example.
2006-06-01
	* Improved, documented distlist()
2006-04-12
	* Added alpha version of function distlist() to facilitate
	creation of distance matrices, using a formula interface.  Among
	other things, it stores row names of generating data frame as an
	attribute, "row.names".  Also added some lines at the end of
	fullmatch() to take advantage of this info if it is given.
2005-10-28
	* Modified interpretation of tol to prevent slowdown on large problems.
2005-10-03
	* Tweaked maxControlsCap in order to get it to handle distance
	matrices of dimension 1xn or nx1 w/0 dropping their dimensions.
2005-09-08
	* Added some detail to error message in fullmatch() for it is
	given lists that fail to be matrices.  The hope is that this will
	be helpful to people in funny situations that can arise when they
	have been manipulating lists containing a few "matrices" of
	dimension 0xn or nxO, as arises when a subclass contains only
	treatments or only controls.  fullmatch handles matrices like that
	the way it should, but it's easy to inadvertently coerce such a
	matrix to a vector of length 0 without a dim attribute. 
2005-09-08
	* Modified beginnings of fullmatch(), maxControlsCap() to remove
	NULL entries from lists of distances.
2005-03-15
	* Created summary method for optmatch objects: reports bound on
	discrepancy between match and optimal match, then calls
	summary.factor
	* Modified print.optmatch, removing report of bound on
	discrepancy between match and optimal match and setting up for
	printing more like that of print.factor (but without giving a
	list of levels separately)
	* Fixed example in matched.Rd
	* Modified fullmatch() to permit giving a list of matrices as
	first argument.  No testing of this new functionality yet, but
	confirmed it still works with older-style calls

2005-04-??
	* Introduced minControlsCap, maxControlsCap
2005-06-01
	* Removed plantdist.rda and instead defined "plantdist" 
	distance matrix in text of examples, in order to work around
	hangup during Windows install

2005-06-16
	* Slight modification to zzz.R
