Version 0.6.4 (2023-02-04): - Fix C integer types - Modernize CITATION file - Integrate testthat tests - Create extract_conditions() to extract treatment conditions from assignment object - Update assg2xBalance() to pass a nonbinary treatment to xBalance() - Replace deprecated is.tibble() with is_tibble() - In assignment(), add warning about length of user-supplied column names if it doesn't match n.tr - Adopt [major].[minor].[patch] convention Version 0.6-3 (2 December 2016): - Update assg2xBalance() to handle when factor names are characters as well as numbers. - Fix block() issue when reporting factor unit names - Fix block() to handle data frames stored as tibbles - NAMESPACE updates Version 0.6-2 (8 January 2015): - In block(), drop variables where variance is 0 overall and warn the user. - In block(), eliminate removal of blocking variables with 0 variance within a group, since scaling matrix is not caculated for each group. - In block(), fix instances where optgreed.c and naive.c returned errors for groups with only one unit. - Bug fixes and documentation editing for seqblock() and block2seqblock(). - Updated Maintainer and Author affiliation. - Update title capitalization and replace require() libraries with :: to conform to Writing R Extensions guidelines. Version 0.6-1 (22 May 2014): *** New *** - Added function seqblock() for sequential experiments - Added function invertRIconfInt() to create confidence intervals by inverting the randomization test - Added function block2seqblock() to translate output from block() to input for seqblock(), for when experiment starts with many units, then adds new units sequentially after initial set of assignments. - Added manual files for seqblock(), invertRIconfInt(), and block2seqblock() Version 0.5-8 (8 April 2014): - In assignment() and demo/block.R, remove dependence on library(gtools) by replacing odd() and even() with (… %% 2) == 1 and == 0, respectively. - Move library(MASS) from Depends: to Imports: - In block(), fix bug to correctly attach namesCol values to output object. - In assignment(), correctly attach namesCol values to output object and provide new length options. - In block(), fix a bug in valid.var that incorrectly led to an error message. Version 0.5-7 (30 July 2013): - Remove .First.lib (zzz.R) - In block(), reformat output from algorithm = "optimal". - In block(), changed default optfactor to 10^7 so that vanilla blockings of x100 with groups run successfully. - In block(), now print error when user provides invalid "algorithm" argument. - Added option block(distance = "euclidean"). - Changed packages nbpMatching and xtable from "Depends" to "Suggests", to only invoke when needed. - Added function createBlockIDs() to output vector of integers to uniquely identify blocks. - Added function assg2xBalance() to ease balance calculations for assignment objects where n.tr = 2. - Added unfactor() function (not exported) to assist createBlockIDs() and assg2xBalance(). - In block(), added argument namesCol, an optional vector of column names for the output of block(). - In block(), drop variables within groups where variance is 0 and warn. - In block(), if a variable has IQR 0 and distance = "mvd" or distance = "mcd", then block() now returns a warning and proceeds using the nonresistant distance = "mahalanobis". - Changed behavior of block() and assignment() to output any unblocked units and incomplete blocks. For consistency with the motivation for the level.two argument, if level.two = TRUE, then unblocked units are listed, but unblocked subunits in units that have a subunit blocked are not listed. - Simplified C files' structure - In block(), added error message when valid.var is specified, but valid.range is not. - Added CITATION file. Version 0.5-6 (1 August 2012): - Display bug fix in file block.R (wherein units with distance zero were blocked as desired, but reported as having distance NA). Version 0.5-5 (11 June 2012): - Memory allocation bug fixes in files mahaloptgreed.c and optgreed.c. Version 0.5-4 (7 May 2012): - In outTeX(), added arguments enabling user to specify output file name(s) and caption(s) directly, as lists of length(block.obj$blocks) or length(block.obj$assg), depending on whether block.obj is output from block() or assignment(). File names are reused as table reference labels, in the form "t:filename". - In outCSV(), added argument enabling user to specify output file name(s) directly, as list of length(block.obj$blocks) or length(block.obj$assg), depending on whether block.obj is output from block() or assignment(). Version 0.5-3 (4 March 2011): - In block(), repaired valid.var and valid.range arguments for algorithms other than "optimal". These worked in 0.4-1, not 0.5-1 or 0.5-2. - In block(), added warning that valid.var and valid.range arguments are ignored for algorithm = "optimal". Version 0.5-2 (16 November 2010): - In block(), added warning for attempts to use algorithm = "optimal" with n.tr > 2. (Optimal algorithm only implemented for n.tr = 2. After warning, block() returns optimal blocks for n.tr = 2.) - In block(), corrected class of display of blocks for algorithm = optimal. - In block(), corrected display of blocks for n.tr > 3 (from 0.5-1). - Updated CHANGELOG format. Version 0.5-1 (5 October 2010): - In block(), blocking now done in C for all algorithms except algorithm = "optimal". - In block(), toggled to check.names = FALSE in data.frame() creation of output object. - In block() manual page, added MCD and MVE citations and clarification of n.tr > 2 algorithms. Version 0.4-1 (28 October 2009): - In block(), added algorithm = "optimal" option. - In block(), added optfactor argument. - Updated manual reference and maintainer information. - Made minor manual edits, such as adding appropriate \dQuote{}'s. Version 0.3 (25 April 2009): - In block(), added weight argument. - Updated maintainer information. - Added CHANGELOG.