03 Sep 2009
So I had been hoping to have written tons of blog posts, posted solutions to NP-hard problems, and become an internet celebrity by the end of the 2009 summer. Unfortunately, something called a Masters Thesis took over my life and turned me into a zombie for all of 3 months. Only recently have I stopped walking in a staggered fashion with a dull look in my eyes.
Writing a thesis is like being a slave to yourself. When the initial excitement ends and the actual work begins, you begin to think about how you got into this predicament. It didn't seem so bad on paper. I mean, you've always done well in class, so why would a thesis be any different? Besides, you have a whole summer.
Two months in, you stare into the ceiling at night and immediately fall asleep due to mental exhaustion. You feel a huge elephant sitting on your brain-waves. Ideas once crystal clear are now hazy. You lose confidence in your writing. Coding standards get thrown out the window as the deadlines approach. You don't mind sleeping under your desk.
But then you ignore the doubts in the back of your head, focus, bite the bullet, and pull through. If you're lucky to have a great advisor like I did, the result is a great feeling of satisfaction as you finish and defend your thesis.
My topic was to get a better sense of how to do query optimization on SPARQL queries, the newly W3C formalized query language for RDF. I based my work on a novel indexing structure called TripleT developed by Dr. George Fletcher. I looked at the different kinds of SPARQL joins possible and for each general case, came up with a general game-plan to ensure good join ordering.
Take a look if you're interested: Cost Analysis of Joins in RDF Query Processing Using the TripleT Index. (40 page PDF)
Life after thesis involves working full-time. A new beginning of paychecks and no homework. Should be fun.