Skip to main content Skip to navigation

Getting Llama

Before you start

Get Xlisp-Stat for your computer, if you do not already have it installed. If you already have Xlisp-Stat installed, make sure it's the most recent version or at least that it is later than version 3.50.

Downloading Llama

The current version of Llama is Version 0.16. The Llama sources and examples are plain (ASCII) text files containing Lisp code.

  • llama.zip is the Llama package for Windows or Unix (or a Macintosh with "Zipit" installed)
  • llama.sea is the Llama package for Macintosh
  • array.lsp (by Jan DeLeeuw) supplies some essential array-specific functions
  • rc-model.lsp is an example showing how to fit a simple row-column association model
  • yaish.lsp is an example showing how to fit the "uniform difference" model used in analysing social mobility tables
  • cautres.lsp is a related example with two uniform-difference terms

Installing Llama

What should you do with those files? Make a new folder (directory) and put them in it.

  • Macintosh: double-click llama.sea to extract the various files
  • PC/Windows: use unzip.exe or pkunzip.exe or somesuch to extract the files
  • Unix: type unzip llama.zip

Using Llama

Under Windows: the neatest way is to create a Windows "association" between .fsl files and the wxls32.exe program; then Llama.fsl should be double-clickable. This happens automatically on the Mac.

Otherwise, start Xlisp-Stat, and use "Load" from the file menu to load either llama.fsl or llama.lsp.

It's usually best to set up data in a file rather than type it directly into the Listener. Three example files are made available above; the file "yaish.lsp" has been commented extensively to help introduce the syntax for entering data and setting up table and model objects. New users should probably look at yaish.lsp, RC-model.lsp and cautres.lsp, in that order.

To read any of the example datasets into Xlisp-Stat/Llama, use "Load" from the File menu. Then use the "Model" menu to fit a model, inspect a model or re-specify a model.