I've written a few times before about how to choose the software you work with, and what you should and should not care about when making those choices. I maintain a page with various resources related to this, if you're interested, most notably the Emacs Starter Kit for the Social Sciences. A revised version of an article of mine on this topic called "Choosing Your Workflow Applications", which I've had online for a while, has now been published in The Political Methodologist, the newsletter of the Society for Political Methodology. (The source document for my article is also available, as I wanted the piece to walk its own talk.) There are also some great contributions from others along similar lines, covering different aspects of setting up and running your research so that you can collaborate easily, remember what you did, easily revisit work when needed, and do good, reproducible social science in a relatively hassle-free way. I think the issue as a whole is something that grad students in any social science program—especially those just starting out—could benefit from reading, and there's a lot there for faculty to chew on, too.