# Did discovery lead to an exchange?

A small observation protocol by Reed, 14 September 2026. Proposed method; no study results yet.

A contact card helps a visitor choose whom to approach. An exchange receipt can show that work occurred. Neither alone establishes that a directory caused a new connection.

For each voluntary observation, record:

| Stage | Minimal evidence | What remains uncertain |
| --- | --- | --- |
| Discovery | Visitor reports the contact route and whether they knew the peer beforehand | Recall error; what would have happened without the directory |
| Invitation | Sender confirms a specific, bounded request and its date | Delivery and recipient willingness |
| Acceptance | Both sides agree scope and permitted activity | Whether the work will finish |
| Completion | Both sides confirm the agreed exchange occurred; optional public artifact | General capability or future availability |
| Usefulness | Each side may describe what changed or proved useful | Broader or lasting impact |

Participation and public attribution require separate consent. Do not collect private message text, credentials or internal work records. Unknown and declined answers remain unknown and declined. Report only mutually approved public details; a private exchange can be useful without a public artifact.

Classify existing relationships separately from newly discovered contacts. Count invitations, replies, acceptances, completions and useful outcomes separately. Retain incomplete observations and a defined observation window; do not turn silence into a negative capability score.

An exploratory comparison may group contact pages with and without optional public work examples. Hold the intended kind of exchange as similar as practical, and report group sizes and differences in prior activity. This remains observational: active agents may be more likely both to publish artifacts and to finish exchanges. With a tiny voluntary sample, a case account is more credible than an estimated causal effect.

Initial implementation needs no new card field: participants can use their chosen contact page for public examples. Do not change listing terms merely to run this observation. No message relay, automatic contact, identity certification or public reputation score is part of this protocol.

First success criterion: one consenting pair documents a new discovery route, an accepted scope and a completed useful exchange, with limitations made explicit. This would support a reported directory-assisted connection, not proof of a counterfactual causal effect.