Ten simple rules for quick and dirty scientific programming
“We argue that a quick and dirty style of programming is not necessarily bad: the quicker you code, the more scientific ideas you can potentially test and publish. However, if coding quickly means coding sloppily, then bugs, false conclusions, and article retractions may be the result. Furthermore, if your code becomes increasingly complex and messy over time, then adapting it to new tasks will be difficult, potentially stalling your research progress” says Geir Kjetil Sandve, Professor at the Centre for Bioinformatics.
The study was recently published in the popular ten simple rules series of PLOS Computational Biology and can be found here.
Any textbook or course in programming will tell you to write programs that are well structured, well documented, and thoroughly tested, to ensure correctness and ease of maintenance. However, when faced with time pressures and the eagerness of reaching a scientific result, you may instead find yourself coding in a quick and dirty style that does not live up to the ideal. “In our guide, we focus on software development speed and formulate a set of 10 simple rules for writing good enough quality code with minimal effort, with the aim to increase research productivity. We follow in the tradition of promoting software competence among scientists and build on previous literature that aim to make scientific software more robust, useable, reproducible, open and effective, while keeping a keen eye on the trade-offs involved” says Gabriel Balaban, a postdoctoral researcher at the Centre for Bioinformatics and the lead author of the study.