Discussion notes from 21 November 1994

"Contributions of a Social Science Based Evaluation
for Adaptive Design Projects"
by Cristoph Grueninger and Werner van Treeck
Presented by Stuart Nelson
Notes by Brad Fischer

Questions


1. The author makes the statement, "Hiding all technical aspects
of computers would hinder users to actually learn anything about
the true technological basis of the adaptive system", is this true?

  The consensus was that it's not true.  An analogy is: Can you use the
  clutch on a car if you can't explain how it works?  You really
  don't need to know the internal mechanisms.
  The issue is, really, should the user actually learn the internals,
  and can they?  Does the user actually want to learn the mechanism?
  In most cases, no; sometimes, yes.

  Designing an easier metaphor for the user to understand is harder
  on the designer.

2.How do you make good inferences as an adaptive system designer
regarding the user's needs for adaptation.

  "Guess" and evaluate was proposed as a method.  Query the user first,
  then have a series of iterations to evolve the system.

  How long would this take?  Three iterations should be good, getting
  95% of the task, in a non-adaptive system.

  Interviews and questionnaires should be done by experts.

  The dilemma between shipping a poor program now and shipping a
  better program in six months, possibly losing market share,
  enters into this question.  There is no "right" answer to this
  problem.

3. How do you formulate theoretical guidelines to determine whether
or not adaptive systems meet their adaptive requirements.

  Create adaptive guidelines!

  There really isn't a systematic way to formulate these guidelines.
  You could poll the users and find out what they hate about the
  interface.  You must differentiate between quantitative and
  qualitative goals.

  The determination has to be tied in to the goals of the adaptation.

4.What methods of adaptation can be used to facilitate learning?

  How do you know if someone has learned something?  You must answer
  this question before you can tell if the adaptation has helped
  the user learn.

  Adaptive help, popup hints, and status line indicators for the functions
  of buttons and hotkeys are possible methods.

  A system whose purpose is teaching could adapt by presenting harder
  examples if you are doing well.


Bonus question:  If you design a system based on these techniques,
how portable will it be?

  Moving from organization to organization, even within a company,
  might not be portable; moving from company to company, for example,
  in similar organizations would probably be portable.  You would
  have to adapt for different cultural biases at the different sites,
  however.
------------------------------------------------------
-- 
 /) _ _ _/       -- Heredity doesn't run in my family. --
/) / (/(/        Brad Fischer (brad@cc.gatech.edu)