This file provides a brief list of major feature enhancements associated with each package release. For a more detailed list of changes see the ChangeLog file. Version 2.0.6 ------------- Bug fixes: * gplots::overplot() - Troels Ring noted that overplot() was reporting Error in eval(expr, envir, enclos) : couldn't find function "lowess.formula" Fixed by changing reference to gplots:::lowess.formula in panel.overplot. Version 2.0.5 ------------- New functions: * gmodels::coefFrame() - Fits a model to each subgroup defined by \code{by}, then returns a data frame with one row for each fit and one column for each parameter. Function changes: * frameApply() and drop.levels() have been moved to the gdata package from the gtools package. * gtools::quantcut() now properly handles cases where where more than one quantile obtains the same value. Bug fixes: * Minor change to avoid 'R CMD check' warnings generated by example code for the depreciated gdata::elem() function. * Fixed a bug in gtools::mixedsort when only a single value is provided. Version 2.0.4 ------------- New functions: * ConvertMedUnits.R - Convert Medical measurements between International Standard (SI) and US 'Conventional' Units. New data: * MedUnits data set provides conversions between American US 'Conventional' and Standard Intertional (SI) medical units. Bug fixes: * Fix behavior of ll() with some data types. * Minor update to remove 'R CMD check' warnings. New sub-package: * Add a 'gregmisc' subpackage. This package simply loads all of the other packages. This should help individuals who are used to the old gregmisc package.. Version 2.0.3 ------------- New functions: * gdata::unmatrix - function for converting a matrix into a well-labeled vector * gtools::frameApply - Apply a functon the the rows of a data frame. * gtools::drop.levels remove unused factor levels from the factors contained in a data frame Feature enhancements: * gdata::ll - Now handles list-like objects, including data frames * gmodels::CrossTable - Added 'SPSS' format, as suggested by Dirk Enzmann * gplots::heatmap.2 - - Add new parameters for controlling color key - - New parameter to allow control of height and width of separator box * gmodels::CrossTable - Show Chisquare contribution of each cell, suggested by Greg Snow. * gdata::read.xls - Add ability to specify the perl executable and path. Bug fixes: * gplots::heatmap2 - - cell note layout was incorrect - Fix handling of rowsep::colsep. - Ignore missing values when computing symmetric range. * gplots::bandplot - Fix problem in passing graphical parameters * add import of 'reorder' and 'na.omit' from 'stats' package to satisfy packages depending on us. Other changes: * gplots::textplot - Code can be simplified because R now supports selecting a fixed width font. * gplots::plotmeans - Add more extensive example. * gdata::elem - depreciated in favor of gdata::ll Version 2.0.0 ------------- First release as a package bundle. Functions are now segregated into four topic-specific packages: gplots gmodels gtool gmisc Thanks to Nitin Jain for doing most of the work of the conversion to a package bundle. Version 1.11.2 -------------- - Simpify and fix the build process on both *nix and Windows by simply providing pre-built perl modules. Version 1.11.1 -------------- - Fix building of of gregmisc on Windows: Runtime Requirements: ActiveState perl in the PATH Build Requirements: ActiveState perl and Microsoft's NMAKE utility in the PATH (A self extracting archive file containing NMAKE is available from .) This was tested on Windows2000 Pro with ActivePerl-5.8.3.809-MSWin32-x86.msi - (Hopefully) Fix building on Linux with Perl 5.8.0. Perl 5.8.3 is reported to work without error on Linux, as does perl 5.6.1 on Solaris. Version 1.11.0 -------------- - Bump version number to 1.11.0 to better reflect the production status of the package. - Add read.xls(), a function to read Microsoft Excel files by translating them to csv files via the xls2csv.pl script, which has also been added to the package. I've provided scripts to run xls2csv: Unix systems: $PACKAGE$/gregmisc/bin/xls2csv MS-Windows: $PACKAGE$/gregmisc/bin/xls2csv.bat or you can execute the perl code directly: $PACKAGE$/gregmisc/perl/xls2csv.pl The perl script explicitly loads its libraries from the installed gregmisc so you don't have to worry about paths. To get usage and parameter information, simply execute the script via any of the above methods without providing any command-line arguments. - Improvements to CrossTable() by Marc Schwartz - Improvements to ooplot() by Lodewijk Bonebakker - plotCI() and plotmeans() now have improved argument handling. - The running() function now has an additional parameter `simplify' which controls whether the returned values are simplified into a vector/matrix or left as a list. - A makefile that will download and attempt to install all available packages from CRAN and Bioconductor is now provided in $PACKAGE$/gregmisc/tools/ - space() can now space points along the 'y' dimension. - Fix an error in the permutations code for repeats.allow=T and r>2. Both the bug report and fix are from Elizabeth Purdom . - Various fixes for compatibility with R 1.9.X. Version 0.10.2 -------------- - Fixed latex warning. - Fixed smartlegend() positioning when x and/or y axis are log-scale. - Add function remove.vars(). Version 0.10.1 -------------- - Fixed bug in textplot() reported by Kevin Wright . Version 0.10.0 -------------- - Now works with and requires R 1.9.0 - Added ooplot() function that mimics Open-Office style plots. Contributed by Lodewijk Bonebakker - Fixed bug in running() that arose when the called function really neaded a minimum number of elements to work on. - Added several new features to running(), it can now allow sequences shorter than the requested width to be present at the front, the back, or on both sides of the full lenth sequences. This allows one to align the data so that the window is before, after, or around the indexed point. - Add enhancements to estimable() provided by Søren Højsgaard : The estimable function now 1) also works on geese and gee objects and 2) can test hypotheses af the forb L * beta = beta0 both as a single Wald test and row-wise for each row in L. - Add colorpanel function, which generates a smoothly varying band of colors over a three color range (lo, mid, high).