worcs: Workflow for Open Reproducible Code in Science

Introductions

AB: Andreas Brandmaier AP: Aaron Peikert CJ: Caspar van Lissa

Mission statements for our projects

WORCS: * One-click solution to get everything set up on YOUR SYSTEM for a project that meets all TOP/FAIR requirements for Open Science * Minimal effort, just good enough

REPRO: * Start from the goal of strict reproducibility, then work to make it more and more userfriendly Has a one click solution, like worcs Also has use_this type solution Checks for dependencies and recommendations on how to install them (see: has_git() in worcs?) automate commands: yaml headers for what data and scripts they need * make file linked to Build button to run entire project

How do these packages relate to each other

worcs interfacing with REPRO: * CJ - Picking up that project and dropping it into a time capsule - I think people are likely to move from worcs to repro - I think it's less likely that people will move the other way * AP: - People should start with worcs (should be standard) - Not ideal to switch package - Move internals to repro in modular fashion

Modularize worcs
Absorb into repro
Make worcs package depend on repro and be a "facade" for it

Next steps

Prepare codebook for compatibility with user-defined description, and machine-readable variable categories (talk with Frank Bosco from MetaBUS)
remove git dependency (CJ: Perhaps I disagree... git is pretty essential to a reproducible workflow)
redoc for collaborators
Use intall (the package) instead of remotes

Troubleshooting: ZIP: Aaron has linux, install zip lib on travis Tests - failing, cleanup examples * Icon, great idea