Automatic usability evaluation with ACT-R and CogTool
A problem with usability practice is the lack of fundamental knowledge to build upon. Specialists have proposed guidelines and principles, but these are often regarded as conflicting, too broad or narrow, or outdated. We would like to have a tool that is both fundamentally grounded in cognitive psychological research and easily applicable in practice.
The Cognitive Modeling Tools project describes the potential power of a set of tools that apply a computer model of human cognition to evaluate user interfaces. Specifically, it discusses CogTool, a usability evaluation tool that uses an ACT-R user model and an intuitive interface to set up usability tests.
This project comprised my Bachelor Thesis for the Innovation Science Bachelor at TU/e. You can find the manuscript at the CMT project page.
Modeling racial segregation
Forty years ago, Thomas Schelling invented an experiment with black and white checker-pieces on a checkers board. The pieces represented black and white households, the board a city. After placing the pieces at random on the board, Schelling swapped pieces, taking one rule into account: a piece would only move if it had fewer than two pieces of the same color in its "neighborhood" (the eight surrounding spots).
Initially, only a few pieces are in such a minority that they want to move. But because of these moves, new minorities are created, who in turn leave their spot. Eventually, complete segregation occurs almost without exception. The most interesting fact about Schelling's simulation is that the pieces are not at all xenophobe. On the contrary, they are happy even when they are in a slight minority. Still, due to its dynamics, the model ends up in segregation without exception.
Over the last forty years the dynamics behind Schelling's simulations have been researched extensively using computer models; the solution however seems to be remarkably robust against all kinds of possible modifications. At the same time, however, slight changes in the micro-level conditions can significantly influence the dynamics of the system. My research therefore analyses the robustness and practical value of the Schellingdale model.
