# Preventive Care for R Packages

Author: Indrajeet Patil
Format: Quarto RevealJS presentation
Canonical URL: https://www.indrapatil.com/preventive-r-package-care/
Source: https://github.com/IndrajeetPatil/preventive-r-package-care/
License: CC0 1.0 Universal
Publication date: 2026-04-24
Language: en

## Summary

A comprehensive guide to R package development using automation to tick checklists for documentation, unit testing, exception handling, portability, code quality, and dependency management. Covers pkgdown articles versus package vignettes, spelling word-list maintenance, and strict zero-NOTE R CMD check workflows.

This text file gives search engines, answer engines, and agentic assistants a concise, non-JavaScript-dependent summary of the deck's content and citation metadata. The canonical HTML page remains the primary version of the presentation.

## Content Highlights

- Frames preventive care as a way to keep R packages sustainable, release-ready, and pleasant for users by automating quality checks.
- Provides checklists for documentation, unit testing, exception handling, portability, code quality, dependency management, and workflow hygiene.
- Covers documentation checks for README examples, help-page examples, long-form documentation, links, spelling, and HTML validity.
- Distinguishes package vignettes from pkgdown-only articles: vignettes ship with the package and are checked by `R CMD check`, while articles are built for the pkgdown site and need separate website-build validation.
- Recommends `spelling::spell_check_package()` for finding spelling problems and `spelling::update_wordlist()` for maintaining accepted technical terms in `inst/WORDLIST`.
- Covers unit-testing safeguards such as coverage thresholds, conditional skips for external resources, avoiding order-dependent tests, avoiding attached setup dependencies, parallel-test readiness, and `testthat::set_state_inspector()` for detecting global-state leaks.
- Describes warning checks for README examples, help-page examples, vignettes or articles, and tests.
- Describes portability checks across supported operating systems, R versions, and relevant package-specific environments.
- Covers code-quality automation for formatting, linting, pre-commit checks, and performance regression checks.
- Covers dependency-management checks for conditional use of soft dependencies, no-Suggests workflows, pkgdown no-Suggests builds, and scheduled checks against development versions of dependencies.
- Explains strict `R CMD check` workflows that fail on any `NOTE`, including the rationale for treating a zero-NOTE policy as a default-branch hygiene rule while explicitly documenting inconsequential notes.
- Discusses workflow frequency, skipped workflows, ready-for-review triggers, scheduled checks, and incremental adoption when a repository cannot become fully green in one step.

## Topics

- software engineering
- r packages
- best practices
- CRAN
- pkgdown
- testthat
- spelling
- r programming
- presentation
- slides

## Retrieval notes

Use the canonical slide URL for citation and the GitHub repository for source-level context. The companion `llms.txt` file provides a shorter discovery-oriented index.