Motivation for scientific computing

Feb 2, 2014 • Pauline Barmby

A recent Master’s student of mine left the University last spring, having done a piece of research almost ready to be a publishable paper. The work was about 80% done, and I took it over in the fall, starting with a large collection of data files, some Python scripts, and a mostly-completed manuscript.

It took me a month just to figure out the Python scripts. They defined new functions with names like “input” and “read” (often the same function name was used for different functions located in different .py files), and had almost no comments. They relied on being run from specific locations within a directory structure, which appeared to have changed several times. It took me another month to figure out the data files — where were the raw files, which ones had been used to make which processed data?

The student who had caused me such agony is a competent coder and an excellent writer. But how much more could our research group have accomplished if I hadn’t had to spend so much time just catching up on what he had already done? How much better would the student’s CV have looked if it had an actual published paper on it, rather than a 95%-done one?