This is the README file that describes the online files for Statistical Analysis and Data Display Richard M. Heiberger and Burt Holland This is the March 16, 2008 version of the R and S-Plus packages and the December 27, 2007 version of the HH online files from http://springeronline.com/0-387-40270-5 We distribute two sets of online files. The HH_2.1-9.* package contains S language functions for all chapters in the book plus some additional functions, complete documentation for all functions, and all datasets. Install them using the standard R or S-Plus installation procedures. HH_2.1-9.zip R package for Windows (Available on CRAN). HH_2.1-9_S_WIN386.zip S-Plus package for Windows (Available on CSAN). HH_2.1-9.tar.gz R and S-Plus source package (Available on CRAN). The zip file contains executable S-Plus and R and SAS source files for all examples from the book, all output transcripts, all figures as PostScript files, and all datasets as ASCII files. hh122707.zip: primarily for Windows users (December 27, 2007). hh122707.tar.gz: primarily for Unix users (December 27, 2007) 1. Recommended unpacking of zip file containing chapter examples for use with S-Plus, R, or SAS. The distribution zip and tar.gz files unpack into the hh directory. See HH Appendix B.1.1, B.1.4, C.1.1, C.1.2 of the book for details. A. Windows zip file: hh122707.zip For my Windows machine I think of the HH book's files as belonging to another user whose directories are parallel to my own. My home directory is c:\HOME\rmh For the book's files I use c:\HOME\hh You may choose a different location. A1. Option 1: Windows Explorer From the Windows XP Explorer Window, right click on the zip folder hh122707.zip. Select menu item "Extract files...". Then enter into the destination path "c:\HOME" (or some other location) and then click "OK" (leave all check mark options set to defaults). The extraction process will create the destination directory if it doesn't already exist. If you accept the recommended "c:\HOME", then the directory will be called c:\HOME\hh The directory contains contains subdirectories and files for Statistical Analysis and Data Display A2. Option 1: MSDOS Command line. You may use any unzip program that understands the *.zip file extension. Make sure you give it arguments to retain the directory structure when the file is unzipped. You will need commands of the form mkdir c:\HOME cd c:\HOME unzip c:\path\to\hh122707.zip If you accept the recommended "c:\HOME", then the directory will be called c:\HOME\hh The directory contains contains subdirectories and files for Statistical Analysis and Data Display B. Tar archive file for Unix or Windows: hh122707.tar.gz Unix users (or Windows users with the gunzip and tar commands) I recommend creating a new user hh and unpacking the file into that user's directory. For the book's files I use /usr/users/hh/hh From the shell, type cd /usr/users/hh gunzip < /path/to/hh122707.tar.gz | tar xf - Be careful in typing as the spaces are critical. Replace the phrase "/path/to" with the correct path on your machine. If you accept the recommended "/usr/users/hh", then the directory will be called /usr/users/hh/hh The directory contains contains subdirectories and files for Statistical Analysis and Data Display C. R packages: HH_2.1-9.zip and HH_2.1-9.tar.gz Install using normal R package installation procedures. D. S-Plus 8 packages: HH_2.1-9_S_WIN386.zip and HH_2.1-9.tar.gz Install using normal S-Plus 8 package installation procedures. library(pkgutils) install.packages(pkgs="/path/to/HH_2.1-9_S_WIN386.zip", repos=NULL) 2. Figures Figures are all gzipped PostScript files. There is no need to unzip them as they will display as is by dropping them into GSview. Here are the July 2005 URLs for GhostScript and GhostView for Windows. Both are needed. Newer releases are probably out now. Unix machine will already have either PostScript or Ghostscript installed. ftp://mirror.cs.wisc.edu/pub/mirrors/ghost/AFPL/gs851/gs851w32.exe ftp://mirror.cs.wisc.edu/pub/mirrors/ghost/ghostgum/gsv47w32.exe 3. More information Read the Preface for information on the file structure in the hh directory. Read HH Appendix B for information on using the hh directory with S-Plus and R. Although Appendices B.1.2.3, B.1.3.3, B.1.5.3, B.1.6.3, and Table B.1 in the book are obsolete, the rest of Appendix B is still current. Read HH Appendix C for information on using the hh directory with SAS. 4. Software 4.a S-Plus and R Appendices B.1.2.3, B.1.3.3, B.1.5.3, B.1.6.3, and Table B.1 in the book are obsolete. These obsolete sections describe access to the functions prior to their availability as a package. These obsolete sections are replaced by this README file. All our files have been tested with R-2.6.1 and S-Plus 8.0.4 on Windows. The package for S-Plus include replacements for several functions that are part of the S-Plus distribution. We informed S-Plus of the bugs and donated the fixes to Insightful. Fixes for S-Plus bugs belong to Insightful, not to us, and are therefore not covered by our copyright. 4.b S-Plus 8 from package HH_2.1-9 The HH functions are available as an S-Plus package from CSAN. If you have not yet installed the package utilities, then enter at the S-Plus command line install.pkgutils() library(pkgutils) install.packages("HH") Load the HH package with the S-Plus command library(HH) All our functions have been tested with S-Plus 8.0, specifically on Windows with 8.0.4. All the examples in the help files and all the examples in the hh/xxxx/code/*.s files work in S-Plus 8. Please write to me directly at rmh@temple.edu if there are questions. The HH package defines the hh() function to find the HH datasets that are in the package. The hh() function does not find the files in the hh122707.* distribution files. You will need to reference the text files hh/xxxx/*/* by full pathname. All our examples assume the HH package is loaded. See HH Appendix B as modified by this README for more details. 4.c. R from package HH_2.1-9 The HH functions are available as an R package from CRAN. Install the package once with the R command install.packages("HH") If you have an older version of the HH package, then use update.packages("HH") Load the package into the current R session with the R command library(HH) All our functions have been tested with R 2.6.1, specifically on Windows with 2.6.1. All the examples in the help files work in R 2.6.1. Most of the examples in the hh/xxxx/code/*.s files work in R 2.6.1. In some cases where the *.s examples don't work in R, equivalent examples are available in parallel *.r files. Some of the text examples have not yet been modified to work in R. Please write to me directly at rmh@temple.edu if there are questions. The HH package defines the hh() function to find the HH datasets that are in the package. The hh() function does not find the files in the hh122707.* distribution files. You will need to reference the text files hh/xxxx/*/* by full pathname. All our examples assume the HH package is loaded. See HH Appendix B as modified by this README for more details. We require the R user to download the following optional libraries from the R distribution: Hmisc Harrell's miscellaneous functions, specifically latex(). multcomp The glht() and related functions. mvtnorm Computes the multivariate normal and t distribution (used by multcomp). abind Combine multi-dimensional arrays. On Windows, from the R GUI, click "Packages", then click "Install Package(s) from CRAN...", then ctrl-click the above four package names. On Unix, enter (that is, pick up the lines from this file, and drop them in the shell window). The spaces and punctuation are critical. local({ a <- CRAN.packages() install.packages(select.list(a[,1],,TRUE), .libPaths()[1], available=a) }) 4.c. SAS. The user must enter the contents of the file hh/sas.library/code/hh.sas suitably customized with the pathname for the directory where you installed the HH files. If you accepted our recommended location, no changes are needed on Windows. You may enter the file by manually dropping the contents of the file into the SAS Edit window and then submitting it. Or you may modify the SAS icon to use this file as the autoexec file. This action defines the &hh macro which is used to locate all the hh datasets and example *.sas files. All our examples assume the hh macro has been defined. See HH Appendix C for more details. ##Local variables: ##mode: text ##End: