# 010 internal

### Intro and welcome

- what this is about
- Anne joins
- reproducibility

### (Richard)

Somebody working on a paper and unable to reproduce own work

- environment.yml
- reprohacks

### (Anne)

- Jupyter book

### (Radovan) Jupyter, R Studio and Binder

- writing down dependencies
- demo
- requirements.txt
- R
- packrat vs renv
- Rmd vs Jupyter

### What are the little things that help a lot for reproducibility?

- Radovan
- keeping things together in one repo (data, code, issues, readme) instead of distributed over many emails and places

### environment.yml

- Project with uncontrolled dependencies
- Create a new environment, install stuff until you get it working
- freeze the yaml, look at it. Losen it up some
- env export --from-history
- how specific should it be?

### If we run out of topics

Cluster etiquette:

*would you think an episode on cluster etiquette/best practices would fit RSHour? Like under what circumstances would it be ok to just use the login node, when to use an interactive job, how to turn the interactive job into a script, what tool(s) you use to develop something that will eventually run on a cluster etc.*