A project archive answers what was built. A knowledge archive should also answer what the work taught.
The useful transition is not from “few articles” to “many articles”. It is from project-specific description to reusable engineering principles.
Separate capability from authority
A model can reason without becoming the execution authority. A derived catalog can be useful without becoming the storage authority. A student report can be useful without becoming an official source.
Make meaningful state durable
Workers, processes and attempts are temporary. Runs, evidence, decisions and authoritative data need a place that survives their replacement.
Verify outside the producer
The component that creates an artifact should not be the only component allowed to declare success.
Let complexity earn its place
A plugin boundary, database, service or runtime should exist because ownership or failure behavior requires it—not because a larger architecture looks more advanced.
The Lab becomes more valuable as these principles accumulate beside the projects that produced them. That makes the archive a record of engineering judgment, not just a gallery of implementations.