Context Mind Mapping

Vary the target question's individual words and aspects to surface what context would help answer it, before checking what's actually obtainable.

One technique for discovering what dynamic context an application needs: write the target question in the middle of a mind map and vary individual words or aspects of it, spawning follow-up questions. For example, "What book shall I read next?" varied on "I" and on removing "next" leads to "What have I read last?" and then "And how did I like that?" — each variation surfaces a piece of context that would help answer the original question, whether or not it's actually obtainable yet.

This is the inverse of starting from what's obtainable by proximity and stability and asking what's relevant — mind mapping starts from what would be useful and works backward toward whether it's feasible to obtain. Actually gathering some of the surfaced context may be infeasible, or worth postponing to a later version of the application; the technique's value is in generating the full candidate list before triaging it, rather than only ever considering context that was already easy to think of.

A recommended combined approach: mind-map what the model might want to know, separately list what the application can actually find out, implement the most obvious overlapping sources first, then extend to more exotic sources as the project matures.

---

Sources

• Prompt Engineering for LLMs (John Berryman), ch. 5