Open Questions

These are the parts I have not resolved. They are here because a playbook that admits its gaps is more trustworthy than one that pads them.

Solo versus organisational collections

This playbook's centre of gravity is organisational: multiple teams, multiple authors, a real yokoten problem. The solo practitioner's collection is the degenerate case where sense, extract, and codify apply fully but govern collapses to a single person's discipline. I do not yet know whether the improvement-to-standard loop has a meaningful solo form or whether solo governance is just honesty with a changelog.

The evidence threshold

Three uses is the default bar throughout this playbook for both codification and improvement proposals. It is a heuristic, not a finding. Higher-risk plays probably deserve a higher bar and cheap reversible plays a lower one, but I have not worked out how to state that without making the bar negotiable in exactly the situations where negotiation is most tempting.

What a library owes its forks

When a variation is a fork rather than an improvement, the honest move is a second play with different entry conditions. But a collection full of near-duplicate plays with subtle condition differences has a findability problem and possibly a credibility one. Where the line sits between honest forking and fragmentation, I do not know yet.

Playbooks and AI-assisted practice

Plays are, in effect, situated prompts: conditions, move, signals. Whether a well-codified play library becomes directly operable by agents, and what that does to the extraction discipline when the practice being codified is itself partly agentic, is an open question I am actively inside rather than observing.

Measurement

The signals in individual plays are observable. What I lack is a good signal for the health of a collection as a whole, beyond the negative checks: no retired plays means no governance, uniform dates mean a museum piece. A positive collection-level health signal would change how govern is practised.

Licensed under CC BY-SA 4.0.