6/8 - So over the weekend, nobody was able to understand what the problem with the Harvard Matlab server was, so we tried our hardest to find some reverse engineering tools that would help us with our project, but to no avail. Instead, we found a few (quite a few, about 80 pages worth) research papers that outlined the conceptual and concrete architectures of both Mozilla and Apache, so that was good news! After reading those, I think we can get along well at the meeting tomorrow about what we still need to accomplish.
6/9 - Meeting day! On the positive, we talked about progress. On the negative, nothing was really decided except to meet with Georgi to go over his work on the Apache Ant proect and that there will be another meeting on Thursday. Joy.
6/10 - Met with Georgi today to go over how he mined the SVN repository and developer mailing lists of Apache Ant. Ritika and I were pretty much ultimately confused, but by the end it all made a bit more sense. We're hoping to get the Perl scripts he used to mine the databases, but unfortunately, any new scripts are supposed to be written in Ruby. I suppose it doesn't matter because I don't know Ruby OR Perl, but I would have preferred to just modify the scripts that were already written. He also gave us the database schema for how to set everything up. Should be interesting.
6/11 - Another meeting today, this one longer. Was about an hour long, and they went over the plan for the next three years. A lot of it wasn't relevant to me because I will only be here another six weeks (and I think I'm glad for that...), but we did learn that we're going to be doing the dynamic analysis of the Firefox and Apache source code with Ed, one of the other grad students. Although we are told this is easier than the static analysis (which is going to be done by Basil, a star student, and Rob, a grad student, because they're available to work on the project over the next year), I'm still not put at ease at all. But Ed isn't available to help us with anything dynamic analysis related until Monday, so that gives us some time to get our repository and mailing list mining into some semblance of order... right?
6/12 - Hooray for us! Ritika and I are positively moving forward and wrote most of our Ruby script for mining the CVS log for the source code repository today. It looks shiny and beautiful and it makes me happy that we did it almost entirely on our own. I feel very accomplished. Now, we just have to get mySQL up and running, create our database and tables, and test the thing. Once it passes the tests, we can actually do something worthwhile! Wouldn't that be exciting? Of course it would! I'm very excited. |
