New Model

[Christian Lebiere]

John Rieman just contributed to the ACT-R repository a model of
interface learning using analogy which he developed at the University of
Colorado.  A short description of the model is included below, followed
by instructions on how to retrieve the files.  Contributions of models,
data, code or papers to the ACT-R repository are always welcome.  Send
me email for instructions.

Christian Lebiere

===============================================================
ABOUT THE MODEL

This is a model of a Macintosh user who knows Microsoft
Word and is learning to use the Cricket Graph graphing program.
The user applies a "label-following" strategy to locate widgets
in the interface that might advance the task goals -- that is,
the user looks for widgets labeled with words that also appear
in their unsequenced set of goals.  The user then tries to
achieve their task using those widgets.  In the process, the
user learns, using analogy to experience with MS Word, about 
interface affordances and situation-specific procedures.

Four files are supplied:
  - README: this file
  - cg-model-description.rtf: a chapter from my dissertation 
    describing this model and several related models at a high level.  
    This is in rtf format.
  - act-r-code-changes: several minor changes to the ACT-R code that
    must be made before the model will run 
  - cg-model-code.3-25-10: the complete ACT-R model
  - cg-model-trace: a trace of running the model

The model was developed in the version of ACT-R that just preceded
release of the book (version 2.0), with the changes described in
the code-changes file.  It runs fairly efficiently on a PowerBook 160.

- John Rieman
  rieman at cs.colorado.edu 
  October 6, 1994
===============================================================

Access Instructions:
------------------
The latest version of ACT-R is available on-line at three locations,
listed here in increasing order of accessiblity.  If you have access to
the Psychology servers at CMU, the Mac version is on server Great
Britain, volume Paradise.  The full repository is on the Largo server,
Public volume.  If you have access to AFS, the directory is
/afs/andrew/hss/psy/act-r.  Otherwise, you can FTP it at
largo.psy.cmu.edu (128.2.248.186) by logging in as anonymous, in the
directory public/ACT-R.