I have just started a graduate assistantship with a professor doing research in the automatic generation of programming problems. The idea here is to define a template for a problem and generate (automatically) enough unique problems for a classroom, along with a testing module for each instance of the problem. It is possible that this will reduce cheating, and it will probably make life easier for professors.
There are several things that I can look into for this project, including extending the existing framework written in Java. The other possibility, which I'm looking into this week, is to continue a nice, mathematical, rigorous approach to studying programming problems. The student that started down this route used Haskell, and seems to have a firm grasp on category theory and functional programming theory (better then mine, but then I am taking up this project mostly to learn).
While I certainly wouldn't give up my evolutionary algorithm work to do research in this area, at least it ties in to some of my other main interests in computer science, namely mathematical foundations and function languages.
Hopefully I will know more about this subject soon, and I will post about it.
No comments:
Post a Comment