The Carpentries Workbench

A re-imagining of The Carpentries Lesson Infrastructure that strips away the tooling from the content so lesson authors, maintainers, and contributors can focus on the information and not the formatting.

in Infrastructure

The Carpentries Workbench is a trio of R packages that deploy, validate, and style Carpentries-style lessons in a modular way. This design allows lesson developers, maintainers, and contributors to focus on the content, not the tooling.

Installation

To install the workbench, you need to have R, git, and pandoc installed. It is recommended to use RStudio, but it is possible to do so without RStudio.

Once you have these installed, you can install the Workbench packages like so:

install.packages(c("sandpaper", "varnish", "pegboard", "tinkr"),
  repos = c("https://carpentries.r-universe.dev/", getOption("repos")))

Usage

New Lesson

You can create a new lesson via the templates at https://bit.ly/new-lesson-md and https://bit.ly/new-lesson-rmd

Once you copy your lesson, move to the next section to build and deploy it locally.

Existing Lesson on GitHub

To build your existing lesson locally, you can clone it from GitHub and then use sandpaper::serve() to preview your lesson as you work on it.

usethis::create_from_github("datacarpentry/r-socialsci") # download and open the R for social scientists lesson
sandpaper::serve() # serve the lesson

History

I redesigned and transitioned the underlying build process for more than 50 open source and active lessons, which serve more than 10,000 learners annualy for The Carpentries volunteer community. This culminated in the creation and deployment of The Carpentries Workbench, a portable lesson infrastructure, which separated the content of lessons from the build and styling components that has significantly improved the quality of life for the more than 100 Lesson Developers and Maintainer volunteers in The Carpentries.

Products

  • The Carpentries Workbench — A suite of R packages and GitHub Actions to build, maintain, validate, and deploy accessible lessons for data and coding skills.
  • tinkr — An R package in collaboration with rOpenSci that transforms Markdown to XML and back again. This package is used for validation of markdown elements without regular expressions and is the engine behind the experimental babeldown package, which provides automated human language translation of Markdown documents using the DeepL API.
  • Carpentries GitHub Actions — Tools for securely provisioning and deploying Carpentries Lessons with arbitrary code on GitHub.
  • Lesson Transition Tool — Automated and flexible transition between Carpentries-style lesson (built with Jekyll) to The Carpentries Workbench. This rearranged folder structure, updated Markdown syntax, and removed unneeded tooling and generated output to reduce the size of the repository and allow the git history to reflect authorship.

Timeline

  • R&D (2020) — Research of the fragmented landscape of build strategies for lessons and, iterating on feedback from the community, designed an initial prototype for the lesson infrastructure with deployment strategies. (supervisor: François Michonneau)
  • Alpha Testing (2021) — Successful alpha test of prototype, directed initial design of website frontend, automated safe and reproducible rendering of literate programming in lessons, and implemented first iteration of workflow to automate transition of lessons to new infrastructure. (supervisor: François Michonneau)
  • Beta Testing (2022) — Released The Workbench to the community, acheived full feature pairity with the former infrastructure, gathered and iterated on feedback from early adopters, coordinated and initiated beta testing phase with official lesson Maintainers and Instructors, and strategically planned to seamlessly transition more than 50 official lessons by Spring 2023. (supervisors: François Michonneau, Kari L. Jordan)
  • Rollout and Capacity Building (2023) — Finalized beta testing phase, coordinated the seamless and secure rollout of The Workbench to all lessons in April–May 2023. Increased capacity within The Core Team by training three colleagues in R package development and creating comprehensive developer documentation about The Workbench design and testing practices. (supervisors: Toby Hodges, Robert Davey)
Posted on:
January 1, 0001
Length:
3 minute read, 573 words
Categories:
Infrastructure
Tags:
Carpentries R
See Also:
My life with the r-universe
Parsing GitHub Task Lists with {tinkr}
WTH is GitHub's GraphQL API even?