Software development
If computational modelling is my craft, then computational methods and software packages are my tools; and I believe that tools should be shared and reused whenever possible.
Initially, as a graduate student, I saw coding more as a means to an end and a potentially reusable by-product of my research activity, but with experience I became more appreciative of software carpentry and developed an interest in writing more robust code that others could use with ease.
Early experience
As a graduate student I’ve spent considerable effort implementing Bash and Perl scripts to pre-process input data and post-process the output from classical molecular dynamics (MD) simulations.
However, my first serious programming project was in Fortran, and it consisted of two stages:
- implementing a three-body interatomic potential for an in-house MD code; and
- parallelizing the entire code through atom-based decomposition and using Message Passing Interface (MPI).
Implementing
Having developed my own models and methods in code and then passed it on for others to use, as well as having contributed to a shared codebase in a multiple research groups,
As a computational modeller who enjoys implementing own models and methods in code, as well as relying on tools developed by others, I value
Less about a personal means of breaking down a problem, and more about building a reproducible product, often collaboratively as part of a group.