Building

Experiments in infrastructure libraries could own and govern. The decisions, the working pieces, and the parts still unproven.

What could libraries control together?

The model I am exploring is shared standards and development, with multiple independently governed implementations. Local control does not require every library to run its own servers.

These build logs test pieces of that possibility. They do not establish that the whole cooperative model works. Start with one bounded question, inspect what happens, and name what is still missing.

The reporting behind the work is in Overdue Conversations. The tests and trade-offs are below.

BLD-002 Building L30 series 2026.05 Cited · 3 sources

Building a MARC Cataloger From Scratch

MarcForge is an open-source experiment in a standalone MARC workflow. Here is the gap it tests, what exists, and what remains unproven.

About this experiment

The question: Can a cataloging workbench stand on its own and exchange records across systems?
What to inspect: MarcForge's approach to MARC editing, validation, authority linking, and record exchange.
The limit: The build log distinguishes plans from later corrections. Koha and Evergreen integrations remain unimplemented in the account given here.

CatalogingExperiments

May 10, 2026·9 min read

BLD-001 Building L30 series 2026.05

Why I'm Building a Library System From Scratch

L/30 is an open-source sandbox for testing what a library can control when the code, data model, and deployment are inspectable.

About this experiment

The question: What changes when libraries can inspect and change their own system?
What to inspect: L/30's design decisions, scope, and maintenance trade-offs in an open-source sandbox using synthetic data.
The limit: A sandbox does not establish readiness for real patron data or prove shared operation and governance.

Open infrastructureExperiments

May 10, 2026·5 min read