From: lara@cc.gatech.edu (Lara Catledge)
Subject: Notes from Wednesday's Class


Class Notes		11/30/94
Adaptive User Interfaces - Katie Albers

Uwe Malinowski and Kumiyo Nakakoji: _Using Computational Critics to 
Facilitate Long-term Collaboration in User Interface Design_. 

Problem: Back and forth arguments between users and designers. Tastes 
great/less filling argument

Do we give ourselves (as designers) the ultimate power over users?
Each group will listen to each other (designers listen to designers, 
users listen to users).  

Intention is to create a common knowledge base.  
Otherwise, why use this kind of system.

Issue: How much does this go toward solving the problem.
<+>  it's a consensus builder
<->  if an end-user can change the system, what do designers need to provide 
as a environment?  Did they have to consider all possibilities?  Huge 
burden on original developers.
<->  What about new parts (Video camera for station 2)?  Can only change the 
interface, not functionality
<+>  Tries to keep track of changes and rationale of changes.  Keeps history 
and long term design rationale.
<+>  Perpetual Usability testing.

Issue: How do individual preference relate to group customizability?
Option:  Group rules.
Option:  Relative to the seriousness of the system (effects in a control 
environment)
Option: More preference oriented choices
Option: Grey out things that people don't use

Issue: Participatory Design vs. Jakob Nielsen vs. Joint Application 
Design (JAD)
<+> All open a dialog between users and designers
O:  Use one and don't forget about the about the options
<+> All create a full feedback loop
<+> advantage of this is rationale is forced and recored by users _and_ 
designers.

I: Rules for interface design:
Limited by application and application area.
The real issue is how do you represent ideals of functionality.

I: Portability
This system relies on rules within the system and assumptions that are coded 
into the program.  Has certain advantages and disadvantages.