Search and find. How we made MDN discoverable
summary
MDN moves from Google custom search to an in-house system: a Kuma (Django) site serving ~33,000 pages, 900 live demos and edits from ~14,000 contributors while ~12 staff maintain it. The product implements full-text search with faceting, JSON endpoints (append .json), Kumascript for dynamic content, a responsive 2013 redesign, a command/query prototype for pre-filtering, and docs.json harvesting. The claim is that these tools make results discoverable and editable in production and the features are already deployed and in active use.
AI-written summary · about
chapters (15)
- 0:00Intro
- 0:45Jannis background
- 1:18What is MDN
- 2:13Community platform
- 5:34The old site
- 8:22The redesign
- 10:15The search
- 11:11Public spaces
- 12:15Search pages
- 13:08Search features
- 13:53Documentation status
- 14:55Chromatic
- 15:32Firefox
- 16:50Code
- 18:25Questions
related talks
20mDocumenting the Web Platform
21mInteractive Documentation at Scale Using MDX & Next.js
58mFederated Search with Drupal, Solr, and React
22mAll about djangoproject.com
36mIf you build it, they will come: Our journey to unified (and discoverable) docs
37mThe Search for Search
20mSearch and Javascript Sites
29mSearch-First Writing for Developers