flombaum.com: a database becomes readable history

The surviving flombaum.com files contain a personal website from roughly 2001–2003: classic ASP pages, a Microsoft Access database, photographs and the small visual assets that gave the site its identity. Restoring it meant recovering the relationships inside that database and expressing them in a format the current application could maintain.

Explore the restored flombaum.com.

What survived

The restoration record describes a 2.6 MB Access database alongside the old webroot. Its tables held daily journal entries, essays, fiction, recurring story sections, chat logs, reader comments, users, pictures and photo collections. Writing bodies contained HTML and old relative image paths.

The original extraction record reports 315 writing files, 1,051 comments, 52 gallery photos and six collections. These are historical extraction counts, rather than a fresh inventory of the current site. They explain why preserving only HTML pages would have missed important structure. Comments refer to article IDs and types; photographs belong to collections; dates and authors shape how writing can be browsed.

Extract content and retain relationships

The extraction task calls mdb-export, parses the resulting CSV and converts writing to Markdown with YAML frontmatter. Metadata retains original identifiers, dates, titles and section information. Comments and user display information become YAML datasets. Sensitive user fields such as emails and IP addresses are omitted from the restored user export.

HTML conversion preserves ordinary paragraphs, emphasis, links and images in the new representation. Image paths are rewritten to the restored section, and separate tasks copy image assets and reference CSS. The pipeline is recorded in lib/tasks/flombaum_extract.rake:19 and :204.

Render through the current application

FlombaumContent reads and caches the file-backed content. The controller exposes journal, essay, fiction, chat and photo collection routes; Markdown is rendered within a dedicated layout. The old writing can therefore be read without recreating the ASP runtime or connecting the public application to Access. Implementation: app/services/flombaum_content.rb:58 and app/controllers/archive/flombaum_controller.rb:8.

This preserves content and selected presentation rather than claiming a byte-perfect replay of every original page. The October 2026 Start Here introduction is a modern editorial addition, distinguishable from the historical journal entries. It helps readers enter the collection without pretending to be an original post.

The case illustrates a durable restoration pattern: export the source database, retain meaningful identifiers and relationships, migrate presentation carefully, and keep the recovered assets alongside readable content. Continue with Tweefle’s frozen social dataset or the restoration workflow.