Clean Code for Good Science: Rust in Research and Health
summary
Science has a reproducibility problem—about 70% of biology papers fail independent reproduction, 60% authors can’t reproduce their own results, and ~74% of research scripts fail in a clean environment—so Rust is used to put correctness at the boundary: Serde validation and custom deserializers catch bad ingests, newtypes enforcing domain constraints prevent swapped-field and out‑of‑range bugs, Cargo gives reproducible builds and docs, Polars and extendr speed pipelines, and a Rust backend actually powers a simulation system and synthetic‑data generation.
AI-written summary · about
chapters (9)
- 0:00Rust in scientific research
- 1:07Reproducibility crisis in science
- 5:43Good science code requirements
- 7:54Encoding data with Rust types
- 13:49Validating data ingestion
- 17:42Rust tooling for reproducibility
- 20:15Rust libraries for workflows
- 21:41Open science in Rust
- 27:01Documentation for research code
related talks
25mContinuous compliance with Rust in automotive software
01h 08mFireside chat
21mRust - Cargo Cult?
01h 29mC++ Should Be C++
05mBenefits and Drawbacks of Adopting a Secure Programming Language: Rust as a Case Study
46mThoughtful Software Engineering
01h 00mDo not trust my or any computational research
26mMoving to modern languages in long-living projects