Adoption playbook

This page is about making the project spread.

The main distribution strategy

A research package becomes memorable when the first hour is frictionless.

That means:

1. a five-minute quickstart,
2. three canonical public examples,
3. one benchmark protocol that stays stable,
4. installation instructions that admit ecosystem fragmentation,
5. and outputs that are easy to cite in papers and talks.

What creates adoption

1. Teach from canonical cases

Germany, Prop99, and Basque should be the public face of the package. Researchers remember methods through recognizable examples.

2. Publish runnable tutorials

Every documentation page should map to a command or example script. If a reader cannot reproduce the page, the page is marketing, not documentation.

3. Keep the protocol stable

Changing defaults too often hurts trust. Add new methods, but keep the evaluation contract stable across versions.

4. Make outputs citable

Ship JSON specs, Markdown reports, and CITATION.cff so users can cite both the software and the benchmark protocol.

5. Support different researcher profiles

• method developers want sweeps and adapters,
• applied researchers want clear case studies,
• labs and teaching teams want offline-friendly snapshots and CI.

6. Show reliability in public

Visible tests, docs builds, release notes, and installation matrices create credibility.

Suggested launch materials

• one short announcement thread,
• one notebook or script per canonical study,
• one table comparing built-in models vs. optional adapters,
• and one tutorial aimed at coding-agent workflows.