The Fortitude Linter: Don't Blame your Tools; Build Them

Abstract

Research Software Engineers are prolific users of software development tools: testing libraries, documentation generators, linters, formatters, debuggers, continuous integration workflows, and so on. They’re a powerful force multiplier, and have an outsized impact on the quality of research software in the wider ecosystem.Fortitude is a new Fortran linter that helps developers to catch bugs early, modernise old code, and keep their code clean and stylish. Starting as an internal tool, it has rapidly become the premier open-source linter for Fortran, featuring around 100 linting rules, robust Fortran parsing, automatic fixes, and IDE integration. It’s also fast – up to two orders of magnitude faster than its closest competitors. With users across six continents, we’ve been able to (indirectly) contribute to a huge number of Fortran projects and make a real dent in technical debt.This was made possible by piggybacking on the popular Python linter Ruff. Not only was it used as design inspiration, in some cases we were even able to use the code itself to bootstrap our implementation. This propelled us straight to the cutting edge with a fraction of the effort it would have taken to write a linter from scratch. This shows how permissive licensing and code reuse can accelerate our work.This talk will discuss how research software engineers can achieve greater impact by contributing to tooling. Using the example of the Fortitude linter, we will discuss how we can create or contribute to tools in order to improve software sustainability for everyone.