How this conference knowledge base was built
Ray Voelker is a library technology professional and IUG conference attendee. He works with Sierra ILS, library data systems, and open-source tools like Datasette. You can find him on GitHub at @rayvoelker, or reach him at ray.voelker@gmail.com / ray.voelker@chpl.org.
At IUG 2026, Ray presented two talks at the Great ILS-Data Pre-Conference: Datasette at the Library and Building a Data Lake.
This is a conference knowledge base for IUG 2026 (Innovative Users Group), held April 12–15 at the Chicago Marriott Downtown Magnificent Mile. It contains:
Detailed notes from 15+ sessions across all four days — keynotes, breakout sessions, forums, and birds-of-a-feather discussions covering Sierra, Polaris, Vega, and cross-platform topics.
Deep-dive reference guides written up from session content — including a Cloudflare protection guide for Sierra, an SSO implementation guide, and a Suggest-a-Purchase comparison.
A gallery of 150 speakers with session history, stats, and rarity tiers spanning multiple IUG years.
Auto-generated llms.txt and llms-full.txt files so AI agents can navigate and understand the site content.
This site was built collaboratively with Claude Code (Anthropic's AI coding tool). The entire pipeline — from initial conference note-taking through the final published site — was a human-AI collaboration.
A Python static site generator: 22 markdown+frontmatter content files, 6 Jinja2 templates, a ~350-line build script, and a YAML site config. Content is the source of truth; presentation is separated into templates and CSS.
Speaker data was scraped from sched.com, parsed, enriched with rarity tiers and session history, and merged with hand-edited fields (quotes, affiliations). The result is a 150-speaker JSON database that powers the speaker cards gallery.
The build script outputs static HTML to a docs/ directory, served by GitHub Pages. A dev server with file watching supports live editing.
A pytest test suite verifies build completeness, catches HTML rendering issues, detects duplicate content, and validates all inter-page links and speaker references.
The source code is open: github.com/rayvoelker/iug2026-shared. Corrections and contributions welcome — file an issue or open a PR.
