One of the reasons I became a blogger was to become more involved in a community which could offer me advice and support on the problems I face in graduate school and as a woman in STEM. In many cases, others have faced the same types of problems I’m encountering and have found possible solutions. So what does this have to do with open source software? In many cases, there are communities of users that will offer beginners advice and support on starting out with a new software package or help troubleshoot problems.
One of the open source software packages I use pretty much daily is LAMMPS (Large-scale Atomic/Molecular Massively Parallel Simulator), developed and maintained by a team at Sandia. While it’s a very well-documented code, the error messages can be vague. This is where the support community comes in. The lammps-user mailing list archive is usually my first stop, to see if someone else has already encountered this problem. If there’s nothing relevant in the archives, the mailing list is full of brilliant scientists who are great at helping solve problems (even if it’s a typo in your input script). The main developers are very active in the user list. It’s also a great place for inspiration.
I also use Python, and its extensions, NumPy, SciPy and Matplotlib for analysis. It’s a much more diverse user group than LAMMPS, and so the community is a bit more spread out. I most often find answers at StackOverflow, but there are many blogs and forums that talk about different ways of using Python. The #python tag on twitter is also a great way to find other users.
Without open source code, I couldn’t do most of my research. Open source code allows me to tinker, to see the guts of what’s really happening, and has a support community that doesn’t charge for their help. I’m lucky that in my field, some of the best code available is open source. What are your favorite open source tools?
I’ve successfully used Octave in place of Matlab for homework (I’m an EE major), which I know is not a shining endorsement for open source, but so far, I have yet to find anything that Matlab can do that Octave can’t do as well, other than plotting, but that’s probably my fault.
I have Python, NumPy, SciPy, and all those, but I never really dived into them.
SciLab is another Matlab-alike. It’s not code compatible like Octave is (most of the time), but I like its interface better. Mostly I stick to Octave, though.
Every time I hear “software X is the industry standard” for packages costing thousands of dollars, I cringe. It makes me cringe because what that’s really saying is “small shops and individuals are locked out of the game because they can’t afford the shiny bats.” All I can say is, thank goodness there are people who develop these OSS alternatives and make them usable.