Hidden Coupling Review Workshop¶
Facilitated workshop — instantiates Relational Grounding Verification
Convenes the people who each hold one piece of a system to surface the hidden couplings that no single view can see.
Some dependencies are invisible not because they are hidden in the code but because they are split across people's heads. Hidden Coupling Review Workshop is the facilitated human session that reassembles that scattered knowledge. It brings the owners of different subsystems, roles, and past incidents into one room and walks a claimed independence — "these two services are decoupled," "this metric is self-contained" — through structured questioning until the couplings that no single participant could see become visible. Its defining move is social elicitation: the mechanism's engine is a group of people whose combined memory holds the coupling, not an automated trace or a formal artifact. It works because the failure it hunts is exactly the failure that survives when everyone assumes their own piece is grounded and nobody compares notes.
Example¶
An engineering org has just had a baffling outage: a routine deploy to the recommendations service took down checkout, two systems everyone swore were independent. The reliability lead convenes a Hidden Coupling Review Workshop with the recommendations owner, the checkout owner, a database administrator, and the on-call engineer who fought the fire. Going around the room, each person's partial view is laid against the others'. Recommendations "just reads" from a shared cache — but the DBA notes that cache is backed by the same primary database checkout writes to. The on-call engineer recalls a near-miss last quarter with the same signature. Piece by piece the group traces a hidden coupling: a shared connection pool that recommendations can exhaust under load, starving checkout.
The output is a hidden coupling trace — an explicit "these are coupled through the shared pool, here is the path" — plus an archived counterexample: the near-miss and this outage, both filed as concrete cases where the assumed independence failed and under what load. Nobody in the room had the whole picture alone; the workshop's job was to make the invisible edge speakable so the org could stop treating the two services as grounded-independent.
How it works¶
- Assemble complementary vantage points. Deliberately seat people whose knowledge overlaps as little as possible — different subsystems, different tenures, whoever remembers the last strange failure. Coverage of blind spots comes from diversity of views, not headcount.
- Interrogate the claimed independence. Take each asserted "these don't affect each other" and push: what do they share — data, infrastructure, a vendor, a clock, a queue? The facilitator's job is to keep asking until a shared substrate surfaces or the independence genuinely survives scrutiny.
- Trace the coupling explicitly. When a shared dependency emerges, draw its path — X couples to Y through Z — so it becomes a nameable object rather than a hunch.
- File the war stories. Past incidents and near-misses that fit the coupling are archived as concrete counterexamples with their triggering conditions, so the coupling is anchored in real events, not just theory.
Tuning parameters¶
- Participant diversity — how far apart the seated vantage points are. Wider diversity finds more couplings but makes the session harder to facilitate and schedule.
- Facilitation pressure — how hard the facilitator pushes on comfortable "they're independent" answers. More pressure surfaces buried couplings but risks defensiveness and false positives conjured to satisfy the room.
- Scope width — how many claimed-independent pairs the session examines. A tight scope goes deep on a few; a broad scope skims many and may miss the subtle path.
- Evidence bar — whether a surfaced coupling needs a corroborating incident to be recorded or can be logged on plausible reasoning alone. A high bar keeps the trace credible; a low bar captures speculative couplings that may not be real.
When it helps, and when it misleads¶
Its strength is reach into knowledge that no artifact holds: the coupling that lives only in the fact that two engineers each remember half of it. It is the mechanism for the failure that is "misdiagnosed as a defect in the focal object rather than a defect in the assumed grounding context" — a common-mode dependency[1] hiding behind an assumption of independence.
Its failure mode is that a workshop is a social instrument and inherits social biases. The loudest or most senior voice can anchor the room; couplings owned by whoever is absent go unspoken; and under facilitation pressure a group can manufacture plausible-sounding couplings that do not actually exist. It also produces candidate couplings, not verified ones — a traced path is a hypothesis about the system, not a proof. The guarding discipline is to require corroborating evidence for high-stakes couplings, invite the vantage points most likely to disagree, and hand the traced couplings to a formal artifact or test rather than acting on the room's word alone.
How it implements the components¶
Hidden Coupling Review Workshop realizes the human-elicitation slice of the archetype — the components that depend on scattered tacit knowledge:
hidden_coupling_trace— its central product: the explicit path by which two supposedly independent things are actually coupled, reconstructed from the group's combined view.counterexample_context_archive— it files the past incidents and near-misses that exemplify the coupling, anchoring the trace in real events and their triggering conditions.
It surfaces couplings through people; it does not maintain the standing formal artifact. It produces no persistent relation_map of the system's dependencies and holds no current_grounding_context_record as a maintained model — that formalization is the Relational Dependency Graph, its nearest twin. The separating line: the Workshop is a one-time human session that discovers couplings; the Graph is the durable structure that holds them.
Related¶
- Instantiates: Relational Grounding Verification — the Workshop is the archetype's human-elicitation instrument for tacit, distributed coupling knowledge.
- Sibling mechanisms: Boundary-of-Validity Note · Context-Swap Protocol · Counterfactual Relation Probe · Invariance Test Suite · Measurement Reanchoring Checklist · Reference-Frame Matrix · Relational Dependency Graph · Schema Context Diff
Editorial Notes¶
Form Classification¶
Form family: Communication, Facilitation & Learning
Rationale: Hidden Coupling Review Workshop operates as a designed message, facilitated interaction, ritual, or learning activity that changes shared understanding because it convenes the people who each hold one piece of a system to surface the hidden couplings that no single view can see
Independent corroboration: The frozen evidence defines Hidden Coupling Review Workshop as 'Convenes the people who each hold one piece of a system to surface the hidden couplings that no single view can see', so its operative form is Communication, Facilitation & Learning.
Review outcome: Independent reviewer agreement; medium confidence.
Origin Attribution¶
Primary origin: Engineering & Design
Origin pattern: Cross-disciplinary synthesis
Present-day reach: Multi-domain
Rationale: Finding common-cause dependencies that defeat nominal redundancy is a core reliability and systems-engineering practice.
Related originating lineages:
- Organizational & Management Science — Cross-functional workshops materially supply the distributed-knowledge elicitation format.
- Systems Thinking & Cybernetics — Systems thinking supplies the relational search for indirect couplings across a whole system.
Review resolution: NASA reliability guidance analyzes common-cause failures created by shared dependencies, a direct engineering lineage for reviewing hidden coupling. Systems/cybernetics supplies the broader interdependence model, so the workshop is a cross-disciplinary synthesis rooted in engineering design. The retained alternate domains identify independent or materially shaping provenance, not downstream reach alone. domain_reach=multi_domain because the mechanism has independent established use in several fields. The encyclopedia entry deliberately composes those lineages.
Encyclopedia synthesis: The exact catalogued form synthesizes established practice rather than reproducing a single standard historical label.
Review outcome: Researched adjudication after independent review; high confidence.
Sources consulted:
- https://ntrs.nasa.gov/citations/20240013667 — NASA technical report on common-cause failure and shared-dependency analysis.
References¶
[1] Vesely, W. E., Goldberg, F. F., Roberts, N. H., and Haasl, D. F. Fault Tree Handbook. NUREG-0492, U.S. Nuclear Regulatory Commission (1981). Explains common-cause dependencies as dependencies that can defeat assumed component independence. registry ↩