People‎ > ‎

Charlie Magnuson, Undergraduate


Contact Information

University of Victoria 
Rigi Lab - ECS 412
Victoria, BC. Canada
E-mail: cmagnuso[at]

I am an undergraduate student in Computer Science at the University of Victoria and am involved with several interesting research projects:

Lively Web (Research page)

A web programming environment written entirely in JavaScript where the web application and the development environment are one and the same. Essentially the web application is created directly on the web application itself.

To start an application in a browser a blank html file is first created on a Lively server. By right clicking within the browser and selecting from a menu various "parts" can be chosen and then dragged to any location on the screen. These objects can be simple shapes, text boxes or more advanced items like tables or a Google map frame. JavaScript is then written directly in the browser which defines the behavior of these objects.

My work involves creating dynamic dashboards using Lively for use in cloud computing and green computing applications. The idea is that metrics such as cpu load, energy consumption, and network availability can be monitored and analyzed via a web application. There are many applications which do this but my work aims to make these dashboards dynamic. Depending on the type of data being monitored, the context of the data, and the requirements input by the user a unique dashboard would be created for that specific task.

Theory Contraction

Suppose we have an ideally rational agent who's current system of beliefs implies some proposition p. This proposition could be anything such as the earth is flat, the temperature outside is 0 degrees Celsius, or chocolate cake is good for you. Essentially p can be anything that is true or false. A system of beliefs implying a proposition p means that there is some supporting relationship between a set of premises and p. Also p itself could be a supporting premise for one or more other propositions. In this way an agents belief system can be modelled as a network of interdependent propositions.

Now suppose that our agent then is made aware that proposition p is false. That is, the agents belief system should include not p rather than p. This could be potentially catastrophic for the agent if p was used as a premise for many other beliefs. It's assumed that the agent is a logical paragon so he or she can't just keep believing a conclusion even though one of its supporting premises is shown to be false.

The problem of Theory Contraction is then how much do we need to contract the belief system such that the agent believes not p instead of p. This problem was shown to be NP-Complete via a reduction from Vertex Cover. My work involves researching different ways to model the problem and investigating whether it's fixed parameter tractable.