A better way to follow Drupal development 🔗 **Source:** [A better way to follow Drupal development](https://dri.es/a-better-way-to-follow-drupal-development) > [!note]+ Document Notes > ```markdown > > --- > tags: [drupal, open-source, ai-tools] > author: Dries Buytaert > source: dri.es > reference: [https://dri.es/a-better-way-to-follow-drupal-development] > date created: 2026-02-20 14:30 > > --- > > # 🏷️ Title: A better way to follow Drupal development > <!-- deep-read:sections start --> ## 🧭 Table of Contents - [📝 Summary](#-summary) - [📖 Main Content](#-main-content) - [🗒️ Notes](#️-notes) - [📚 References](#-references) - [🗣️ Paraphrase](#️-paraphrase) - [📰 Newsletter](#-newsletter) - [✅ Action Items](#-action-items) - [🧪 Hypotheses](#-hypotheses) - [📊 Methodology](#-methodology) - [❓ Questions](#-questions) - [🔮 Future Research](#-future-research) - [💡 Implications](#-implications) - [🧠 Learning Styles Assessment](#-learning-styles-assessment) - [📊 Accessibility Assessment](#-accessibility-assessment) - [🌍 DEIB Assessment](#-deib-assessment) - [✅ Logical Validity](#-logical-validity) - [🛡️ DARVO Analysis](#-darvo-analysis) - [🎯 Conclusion](#-conclusion) ### 🔗 References - Include links cited in or relevant to the article. ## 📝 Summary - What happened. - Dries built **Drupal Digests**, an AI-powered summarizer of key issues and commits across Drupal Core, Drupal CMS, Drupal Canvas, and the Drupal AI initiative. - Why it matters. - It reduces the cognitive load of tracking a large open source project and helps contributors stay current. - Key stakeholders. - Drupal core and contrib developers, maintainers, new contributors, and potentially the broader Drupal community. - Risks and tradeoffs. - Reliance on AI summaries could miss nuance or misinterpret changes; tool is experimental and may or may not become a long-term service. - Bottom line. - Subscribe to the RSS feeds, try the summaries, and give feedback so Dries can decide whether to invest further in the tool. ## 📖 Main Content - Explain the core arguments in plain language. - Dries has followed Drupal Core for 15+ years via commit RSS and manual issue reading. With Drupal Starshot broadening activity (Core, CMS, Canvas, AI), that approach no longer scales. He built **Drupal Digests**, which monitors relevant issue queues and, when something important is committed, sends the diff and discussion to an AI model. The AI generates a short summary: what changed, why it matters, and whether action is needed. Output lives as Markdown files in a GitHub repo, plus RSS feeds for consumption. - He argues that staying informed is one of the hardest parts of contributing to a big open source project, and these summaries can help new and experienced contributors alike. - Add sourced context and define terms. - **Drupal Starshot** is the umbrella initiative to accelerate and modernize Drupal’s experience (including Drupal CMS, Canvas, and AI). - **Issue queues** on [Drupal.org](https://www.drupal.org) are where bugs, features, and tasks are tracked. - AI summarization here is essentially LLM-based explanation of diffs and discussions. - Note uncertainties or contested claims. - Dries is still “tuning the prompts” and measuring usefulness informally. He notes the current cost (~$2/day) and commits to at least a year-long experiment, reserving judgment about its long-term value and possible expansion (search, filters, custom feeds). ### 🗒️ Notes - Dries describes a long-standing workflow: subscribing to Drupal Core commit RSS, scanning them daily over coffee, and opening issues on Drupal.org to read summaries and comments. - As his focus expanded from Core to **Drupal CMS**, **Drupal Canvas**, and the **Drupal AI** initiative under Drupal Starshot, this manual monitoring became unmanageable even though the work is public. - To handle this, he created **Drupal Digests**: - It watches Drupal.org issue queues for several key projects. - When “something noteworthy” is committed, the tool sends the issue discussion and code diff to an AI model. - The AI returns a human-readable summary of: what changed, why it matters, and whether any action is required. - Each summary is stored as a Markdown file in a GitHub repository, making it easy to browse and version. - Dries also publishes RSS feeds from these summaries so he can keep his “morning coffee and RSS” habit, but with higher-level context. - He notes that staying informed is one of the hardest aspects of contributing to large open source projects, especially for new contributors or busy maintainers. - He expects the digests to: - Help new contributors ramp up faster. - Help experienced maintainers notice important API or behavior changes relevant to their modules. - Support better collaboration across subprojects by surfacing relevant changes. - He is still refining prompt design and system behavior, emphasizing that this is an experiment. - Current operating cost is under $2 per day in AI tokens, and he commits to running the system for at least one year to gauge real value. - If Drupal Digests proves genuinely useful, he imagines building a richer product: a proper interface with search, filters, and user-specific or custom feeds. - He ends with a call to action: subscribe to one of the feeds and provide feedback on whether and how it helps. ## 📚 References - [A better way to follow Drupal development (original article)](https://dri.es/a-better-way-to-follow-drupal-development) - [Drupal.org main site](https://www.drupal.org) - [Drupal Core issue queue](https://www.drupal.org/project/issues/drupal) - [GitHub](https://github.com) (platform hosting the Markdown summaries) ## 🗣️ Paraphrase - Paragraph A paraphrase. - Dries explains that his long-used habit of reading Drupal Core commits via RSS has broken down as his responsibilities expanded, so he created an AI-powered digest tool to summarize key changes across several Drupal initiatives and deliver them back via Markdown and RSS. - Paragraph B paraphrase. - He believes this tool can ease one of open source’s hardest problems—staying informed—especially for new contributors and maintainers, and he’s running a low-cost, year-long experiment to see if it’s worth turning into a more fully-featured service, asking readers to try it and share feedback. ## 📰 Newsletter Drupal folks aren’t the only ones drowning in commit feeds. In this piece, Dries shares how he built **Drupal Digests**, an AI helper that watches issue queues, reads diffs, and ships concise RSS summaries for busy contributors. For WordPress teams, it’s a great pattern to borrow: imagine AI-powered digests for Gutenberg, core tickets, or major plugins, feeding into your reader of choice. Pair something like this with make/core posts and Trac/GitHub activity to keep your devs in the loop without burning hours. Helpful links: - Drupal: https://dri.es/a-better-way-to-follow-drupal-development - WordPress core dev blog: https://make.wordpress.org/core/ - WP GitHub activity explorer (example): https://github.com/WordPress/wordpress-develop ## ✅ Action Items #### DevRel - Document the “AI commit digest” pattern and share with contributor communities as a case study. - Run a short workshop on using RSS + AI summaries to track fast-moving projects. - Create example scripts or templates for teams wanting to build similar digest tools. - Publish a blog or video walkthrough and measure adoption via feed subscriptions. - Collect qualitative feedback from contributors about clarity and usefulness. #### OSPO - Draft guidance on acceptable use of AI for code/issue summarization in internal and upstream workflows. - Ensure license compatibility and attribution for any content mirrored from project trackers. - Define a security review path if digest tools handle private or embargoed issues. - Add AI-digest services to software inventories and SBOM where relevant. - Explore co-funding with foundations for digest infrastructure that benefits multiple projects. #### Community - Invite maintainers to subscribe and report when a digest helped (or failed) them. - Add “How to stay up to date” sections to contributor onboarding materials referencing such tools. - Offer office hours on interpreting digests and jumping into related issues. - Consider small sponsorships to offset token and hosting costs. #### WordPress - Core: Prototype an AI digest that summarizes key Trac tickets and GitHub PRs per week. - Core Performance: Summarize impactful performance patches and profiling findings. - Core AI: Track AI-related issues and experiments with clear, human-readable summaries. - Core Program Team: Integrate digests into weekly coordination rituals. - Mobile: Summarize major app changes and API shifts affecting mobile clients. - Playground: Provide digests of environment and tooling updates for educators. - Community: Share “staying informed” guides for meetup and WordCamp organizers. - Training: Turn important digests into lesson snippets or case studies. - Marketing: Use sanitized digests to craft release narratives and feature highlights. - Support: Surface digests on changes likely to affect end-user support loads. - Design: Track UX-affecting changes and document them for design reviews. - Documentation: Use digests as prompts to update DevHub/HelpHub and examples. - Photos / TV: Curate media resources explaining major changes for non-dev audiences. - Polyglots: Highlight strings or features with translation impact. - Accessibility: Track accessibility-related commits and audits. - Meta: Digest .org infrastructure and contributor tooling updates. - CLI: Summarize new or changed commands with examples. - Hosting: Share digests of server-related or configuration-impacting changes. - Tide: Highlight rule changes, scoring updates, or false-positive fixes. - Openverse: Surface ingestion pipeline and moderation policy updates. - Themes: Track theme API changes and popular patterns. - Plugins: Summarize policy and security changes affecting plugin authors. #### The WPOCC - Pilot funding for AI digest tools that serve multiple open source CMS ecosystems. - Sponsor “governance and communication” sprints where digest tooling is evaluated. - Test microgrants for maintainers who integrate digests into onboarding and documentation. ### FAIR - Encourage local storage of digest metadata and signatures to avoid lock-in. - Define transparent funding and sunset policies for community-facing AI tooling. - Publish reusable governance artifacts (charters, MoUs) for shared digest infrastructure. ## 🧪 Hypotheses - H1: Contributors who subscribe to AI-generated digests will report higher awareness of important project changes than those who only follow raw commits. - H2: New contributors ramp up faster when they have access to curated summaries of recent changes. - H3: AI digests reduce the time maintainers spend manually monitoring issue queues without reducing their sense of control or understanding. - H4: The perceived usefulness of digests strongly depends on prompt quality and model choice. ### 📊 Methodology - Run an A/B study: one group gets AI digests; another follows current channels (RSS, mailing lists, chat). Survey awareness and confidence over 3–6 months. - Measure behavioral data: number of relevant issues/PRs acted on, time-to-first-contribution, and review response times. - Qualitatively code feedback from interviews to identify when AI summaries mislead or omit critical nuance. - Experiment with different prompt templates and models, tracking user ratings of summary accuracy and clarity. ### ❓ Questions - What level of detail in digests best balances signal and noise for different contributor roles? - How should AI summarization errors be surfaced, corrected, and learned from? - Can digest tools safely handle private or security-sensitive issues? - How do we ensure accessibility and internationalization of AI-generated summaries? ### 🔮 Future Research - Explore multi-project digest tools that cover entire ecosystems (e.g., Drupal + related PHP libraries). - Investigate ways to combine AI summarization with human curation for high-stakes changes. - Study long-term effects on contributor retention and burnout when digest tools are widely adopted. - Evaluate cross-language summaries to support non-English-speaking contributors. ### 💡 Implications - Practice: Maintainers can rely on structured digests to prioritize their attention and react to breaking changes faster. - Policy: Projects may formalize guidelines for AI usage in communication and documentation workflows. - Security: Summaries need guardrails to avoid leaking sensitive details from restricted discussions. - Education: Training materials can incorporate digest content to keep courses and docs current. - Funding models: Low-cost yet widely beneficial tools like this are good candidates for community or foundation sponsorship. ### 🧠 Learning Styles Assessment - Reading level estimate: Around 9th–10th grade (Flesch-Kincaid approximation; short, clear sentences with some domain jargon). #### 👀 Visual learners - Present: Clear conceptual structure (old workflow → problem → new tool). - Improve: Add architecture diagrams of the digest pipeline and example screenshots of summaries and RSS feeds. #### 🔊 Aural learners - Present: None explicitly. - Improve: Provide a short audio explanation or podcast-style segment describing the motivation and workflow. #### 🗣️ Verbal learners - Present: Straightforward narrative, clear definitions by example (e.g., what the tool does, why it exists). - Improve: Add more analogies and explicit comparisons with other notification systems. #### 🧩 Kinesthetic learners - Present: Implied “try it yourself” via subscribing to feeds. - Improve: Step-by-step setup guide and a small exercise (“follow one issue from commit to digest”). #### 🧠 Logical learners - Present: Problem-solution structure; mention of cost and time horizon. - Improve: Add metrics, simple flowcharts, and explicit success criteria. #### 🧑‍🤝‍🧑 Social learners - Present: Emphasis on collaboration and community benefits. - Improve: Encourage discussion in community channels and small group experiments. #### 🧘‍♀️ Solitary learners - Present: Blog format supports private reflection and experimentation. - Improve: Add reflection prompts like “How do you currently track project changes?” #### 🌿 Naturalistic learners - Present: None. - Improve: Use ecosystem metaphors (signal/noise, flows) or discuss energy/time “resource usage” analogies. ### 📊 Accessibility Assessment #### Screen reader semantics - Present: Blog likely uses standard headings and landmarks. - Improve: Ensure clear heading hierarchy and descriptive link text for feeds and examples. #### Keyboard navigation - Present: Typical blog layouts are usually keyboard-friendly. - Improve: Confirm focus styles, skip links, and logical tab order. #### Contrast and readability - Present: Dri.es generally uses clean, high-contrast design. - Improve: Validate contrast ratios and line spacing against WCAG 2.2 AA. #### Text resizing and reflow - Present: Responsive design expected. - Improve: Test at 200–400% zoom and with reduced-motion preferences. #### Media and captions - Present: Mostly text; minimal media. - Improve: If screenshots are added, include alt text and captions. #### Forms and interactions - Present: Limited interaction beyond links. - Improve: For any future subscription or feedback forms, ensure proper labels and error handling. #### Standards reference - Map improvements to WCAG 2.2 AA criteria such as 1.4.x (contrast, resizing), 2.1.x (keyboard), and 3.3.x (input assistance). ### 🌍 DEIB Assessment #### Cultural breadth - Present: Focused on Drupal community, largely Western OSS culture. - Improve: Highlight contributors and use cases from multiple regions and languages. #### Inclusive language - Present: Neutral, non-gendered, and respectful tone. - Improve: Define jargon (“issue queues”, “Starshot”) for newcomers and non-native speakers. #### Neurodiversity support - Present: Short, structured article with clear sections. - Improve: More subheadings, bullet lists, and explicit summaries. #### Representation - Present: Primarily the author’s voice. - Improve: Include quotes or experiences from a variety of contributors. #### Psychological safety - Present: Encouraging, non-judgmental framing. - Improve: Mention feedback channels and norms for constructive critique. ### ✅ Logical Validity #### Strong reasoning - Clear causal reasoning: monitoring more projects makes manual RSS scanning insufficient, so automation and summarization reduce workload. This is plausible and grounded in experience. - Pragmatic evaluation: commits to a year-long, low-cost experiment before scaling up, which shows awareness of uncertainty. #### Fallacies - None obvious. Claims are moderate and experience-based rather than sweeping generalizations. #### Tone of voice - Tone is pragmatic, reflective, and community-oriented: “I built this to scratch my own itch, but realized it could help with something bigger.” ### 🛡️ DARVO Analysis - Deny: No denial or minimization of problems; the difficulty of staying informed is openly acknowledged. - Attack: No attacks on individuals or groups; the post is constructive. - Reverse victim/offender: Not present; Dries positions himself as a problem-solver, not a victim. ## 🎯 Conclusion The article introduces **Drupal Digests**, an AI-powered way to track important changes across several Drupal initiatives through concise summaries and RSS feeds. It frames this as an experiment to ease the burden of staying informed in large open source projects, inviting readers to subscribe, try it in their own workflows, and provide feedback that could shape a more robust future service. ## 🧱 Refactor Appendix ### 📚 Literature Split Pad - Purpose: Stage Konik CEE units that become per-source Literature notes. - Split level: **H3**. - Include verbatim quotes under 25 words and a [view highlight] link. ### [H1] AI digests summarize Drupal development > [!quote] [A better way to follow Drupal development](https://dri.es/a-better-way-to-follow-drupal-development) by [[Dries Buytaert]] via [[dri.es]]. Published 2026-02-19. > “When something noteworthy gets committed, it feeds the discussion and diff to AI, which writes me a summary.” Two lines: context and use. Cite data if relevant. [view highlight](https://dri.es/a-better-way-to-follow-drupal-development#H1) #xref #articleseed ID:rw_[MISSING] ### [H2] Staying informed is hardest in large OSS projects > [!quote] [A better way to follow Drupal development](https://dri.es/a-better-way-to-follow-drupal-development) > “Staying informed is one of the hardest parts of contributing to a large Open Source project.” Two lines: context and use. [view highlight](https://dri.es/a-better-way-to-follow-drupal-development#H2) #xref ID:rw_[MISSING] --- ### 🔖 Atomic Split Pad (Zettel Candidates) - Purpose: Promote reusable claims into Atomic notes. - Split level: **H4**. - Each item: one idea, present tense, specific, testable. #### [Z1] AI digests reduce open source info overload - Scope/condition in one line. - Pointer to page/time or dataset. - Links: [[Open-Source-Communication]] [[AI-Tooling]] #### [Z2] RSS remains valuable with higher-level summaries - Scope/condition in one line. - Pointer to page/time or dataset. - Links: [[Developer-Workflows]] [[Information-Filtering]] --- ### ✏️ Fleeting Queue (Extract by Selection) - Purpose: Capture quick actions or sparks for Fleeting notes. - Extraction: **Select one line** → “Refactor: Extract selection to new note (content only).” - Keep each line short and self-contained. - [F1] Try Drupal Digests and note whether summaries catch important API changes — context/https://dri.es/a-better-way-to-follow-drupal-development #fleeting #inbox - [F2] Prototype similar AI-based digest for WordPress core tickets — context/https://dri.es/a-better-way-to-follow-drupal-development #fleeting #inbox - [F3] Define metrics for evaluating AI digest usefulness for contributors — context/https://dri.es/a-better-way-to-follow-drupal-development #fleeting #inbox --- <!-- deep-read:sections end -->