Collaborative science in the age of software

Science relies on software. Yet, we don't sufficiently make use of established tools for writing, collaborating on, and validating the software that underlies our empirical claims. Negative consequences of this blind spot manifest most saliently when data analyses go wrong, but also in severe inefficiencies in education, collaboration, knowledge sharing, and reproducibility. I outline ways in which our haphazard approach to software hinders the process of scientific discovery, and, perhaps more importantly, what we can do about it. A professional attitude, use of established tools for version control, collaboration, testing, and documenting, and an appreciation for software's importance to science is central in our quest for reliable, valid, and cumulative science.

Sponsorship of an event does not constitute institutional endorsement of external speakers or views presented.