Embedding deep provenance tracking, reversible execution and mergeable histories directly into an interactive programming environment (a POSIX-like shell) greatly decreases the gap between exploratory scientific work and reproducible, publishable results; all whilst leaving existing workflows intact and being programming language agnostic.

• POSIX-compatibility: I'm interested in both being POSIX-compatible and thinking about the post-POSIX nature of the shell. But this bolsters claims about not breaking existing workflows and is relatively easy to measure.

• MRDT for Shell Histories: Spoke to KC about this, I think there is some interesting work here for developing a useful MRDT for this.

• Easy to use provenance tracing: as simple as searching your shell history!

• OS Design Discussion: Plan9, fork/exec, void processes, VURL etc.

1. Introduction: Setting the scene: exploratory programming, why a shell, scientific results etc.

2. Background

3. Design and Theory

4. Implementation

5. Evaluation

• Shell library

• Shelter