Drawing on a decade spent bridging the Earth Science & open-source worlds, I argue that the Earth Science scientific community has much to steal from the computer science & engineering community — specifically, the concept of “rigor” broadly defined at the storage, compute, and science layers of the scientific software stack.

1. We have recently made progress toward increasing rigor at the storage layer with Icechunk, taking inspiration from transactional database systems.
2. We still lack analysis libraries and idioms that make it easy to articulate simple, convenient, and correct expression of complex scientific ideas.
3. We should be doing more to apply reliability and correctness approaches (e.g. property testing) in scientific software.

At the workshop, I am keen to explore any of these ideas with the group in the room.