University Projects

These are references and projects related to my studies at the University of Namur (FUNDP). A list of my publications can be found here.

Grid application with load balancing and robustness

Work of 2006. For the distributed applications lesson of Prof. Englebert. (2nd year of the masters grade)

The task was to develop a grid computing application for ray tracing calculation, using the well known PovRay ray tracer. The network included an arbitrary number of worker stations, work dispatchers and clients. There was one active dispatcher who, if he crashed, would be replaced by a backup dispatcher. A client application could connect to the dispatcher and submit jobs, query job progress and download completed jobs. A job was actually a PovRay scene description which was to be rendered. An incoming job was cut into pieces by the leading dispatcher, and the pieces were then rendered by the registered workers. Crashed workers had to be managed as well. Completed jobs were sent back to the leading dispatcher, who put arriving pieces together. It was all written in Java and networking stuff was done with TCP, based on our own protocols. This project was really cool and interesting, we learned much about socket programming and threading in Java.

Stock market application

Work of 2006. For the software development lessen of Prof. Habra. (2nd year of the masters grade)

This was a huge project involving Java, CORBA, Oracle and a lot of time. This stock market application included the stock market server, a broker interface, able to connect to the stock market server. Then each broker could have his own server, offering services to his clients, for whom an interface had to be written, too. All programming was done in Java, networking stuff was handled by openOrb, an open-source Java implementation of the CORBA middleware framework. The databases were Oracle 8 databases, one for the stock market server, and one for each broker. The project was rather big (for a student project, that is), there were 5 people to work in one team for over three month. This was the first time we developed something bigger in Java, we thus gained much experience. I was responsible for user interfaces and merging the code of the different group members. As I didn't know that there was something like JFace, I started programming my own window manager, and form generator. I spent much time on all this, but it looked quite cool in the end (it looked like JFace indeed).

Compiler for a Pascal-like language

Work of 2005. Project for the syntax and semantics lesson of Prof. Schobbens. (1st year of the masters grade)

This was our first big project at the university. The task was to develop a compiler for a Pascal-like language, which was called LSD05; the LSD05 code was compiled into P-Code. The language was strongly typed, variable scope was limited to sub-procedures, it featured recursion and all common control structures. The compiler was developed in C, using Lex and Yacc (or rather flex and bison) in order to automate lexical and syntactical analysis. The classical approach of building a syntax tree, a symbol table, and then producing the output was followed. This project was really cool, much cooler than all subsequent projects, we didn't have to bother about users, interfaces and stuff like that (just kidding ), it was only brain work. We did this in groups of 3, I was primarily involved in the generation of the syntax tree, the symbol table and the type/scope verification.