Thursday, September 2, 2010

System.out.println("Hello world");

Hello all!

This blog is intended to be a place for me to chronicle the work I will be doing over the next school year developing my senior project in the Computer Science department at Willamette University. With apologies to non-technical readers, this blog is likely to be somewhat dense and is not intended for a general public readership.

At this point, I am exploring different potential ideas for what I could be working on. I have two ideas at this point, neither of which are fully formed, but both would be enjoyable, interesting, and challenging.

Developing an Operating System and Multi-Threaded Virtual Machine
My first idea, which I have had since sophomore year when I took CS353 and developed the PC231, a simple virtual machine to develop assembly code on, is to expand upon the PC231 in the following ways:
  • Improve the architecture (larger word size, two's complement, more hardware commands, etc.) - potentially/probably trying to replicate a more common architecture
  • Develop a limited kernel to handle memory management and execution control
  • Develop a rudimentary operating system to allow the user to execute arbitrary commands
  • Develop a virtual file system to integrate with the virtual machine
This would be a fairly new field for me - outside of CS353 I have never done any kind of assembly code work or low-level programming. Exploring this area would be interesting, however my limited experience and knowledge makes it difficult to predict if this is a reasonable goal, and if the amount of work is feasible.

Develop A Lightweight Issue Tracker for Use With Distributed Revision Control Systems
Over the summer, I developed an extension for Mercurial, a distributed revision control system. This extension is a lightweight, command-line bug tracker called b. I would like to develop this into a free standing tool which can be used anywhere, with few system requirements, no setup or configuration necessary, and a minimal footprint. Specifically, I intend to do the following:
  • Remove Mercurial dependency and turn b into a standalone program
  • Enable it to integrate cleanly with the major distributed version control tools (Mercurial, Git, Bazaar)
  • Develop an internal web server to provide a web interface to view and manage issues
  • Develop plugins for major IDEs (Eclipse, others if time permits)
  • Develop a graphical front-end for non-command line use
The different facets of this project would allow me to explore many fields at once, ranging from Python programming to web development to software management concepts. This would be an interesting and valuable project, both personally - because I would love to use such a tool - and potentially others would find it useful as well.


And these are my two ideas at this point. Comments, questions, and alternative ideas are welcome!

2 comments:

  1. My two cents (or three):

    Try to provide more detailed description to bring out the features of b, compare with t and bugs everywhere.

    Prioritize the sub-tasks and assess what need to be done for each task.

    Re-phrase: "I would like to develop this into a free standing tool which can be used anywhere, with few system requirements, no setup or configuration necessary, and a minimal footprint."

    ReplyDelete
  2. I think it would be a great idea if you talked to the class once about versioning systems. They are used in the work world and presumably a Good Idea(tm) for large-scale projects, but not really gone over at Willamette. If you have experience with versioning software, it could be beneficial to everyone if you explained to everyone the benefits, features, etc, from what you've learned at least. You could also promote Mercurial at the same time - I hear you like it!

    ReplyDelete