Legacy Function Interview¶
Stakeholder interview — instantiates Latent Constraint Preservation Audit
Recovers a structure's tacit function and hidden dependents by questioning the maintainers, operators, and long-time users who still carry the knowledge in their heads.
The Legacy Function Interview recovers why does this exist? and who leans on it? from the people who know, not from the archive. It is structured questioning of maintainers, operators, long-tenured users, and affected parties to surface the tacit knowledge that was never written down — the near-misses that shaped a rule, the edge case a step quietly guards, the downstream party who would be hurt if it vanished. Its defining move is harvesting living, tacit knowledge: it works precisely where documentation has decayed but memory survives, and much of what it recovers is knowledge people hold without ever having articulated it.[n1] The interview produces both a function hypothesis grounded in experience and a register of the dependents those experienced people can name from having watched the structure work.
Example¶
A newspaper is under pressure to eliminate its copy-desk fact-checking pass — a separate read in which a veteran editor verifies names, dates, and quotes before a story publishes. Online competitors skip it, it slows publication, and management sees it as a relic of the print era. Before cutting it, an editor interviews the three most senior copy chiefs. The conversations surface what the pass actually does: it is the last point where a libel-exposing misquote gets caught, the place where a reporter's transposed statistic is corrected before a correction is needed, and the informal check that stops one particular beat's recurring source-attribution error. The chiefs also name who silently depends on it — the legal team, who have never had to defend a suit partly because of catches at that desk, and two reporters who rely on it as a safety net. None of this was in any manual. The interview does not decide the pass's fate; it converts "print-era relic" into a specific list of functions and dependents that any replacement would have to cover.
How it works¶
- Pick the right informants. Target the people with the longest tenure and the most direct contact — maintainers, operators, frontline users, compliance owners — including those who have moved on but can be reached.
- Ask for stories, not verdicts. Elicit concrete incidents ("tell me about a time this caught something") rather than opinions ("is it useful?"), because tacit function shows up in episodes, not summaries.
- Probe for the invisible dependent. Explicitly ask who would notice or be harmed if the structure disappeared, drawing out low-visibility and downstream parties the informant has seen rely on it.
- Triangulate across informants. Compare accounts; a function multiple independent people describe is stronger evidence than a single recollection, and disagreement flags a contested or drifted purpose.
Tuning parameters¶
- Informant breadth — one expert versus many across roles. Breadth guards against a single person's blind spots and captures dependents no one individual sees, at the cost of time and scheduling.
- Question structure — open-ended narrative versus a fixed checklist. Narrative surfaces the unanticipated function; structure ensures coverage and comparability. Most interviews need both, in that order.
- Leading-question guard — how carefully the interviewer avoids planting the answer. Tight discipline prevents informants from simply confirming the interviewer's prior; loose questioning contaminates the hypothesis with the remover's assumptions.
- Recollection-decay allowance — how much weight to place on old, possibly reconstructed memories. Heavier skepticism resists confabulation but may discard a real, hard-won lesson.
When it helps, and when it misleads¶
Its strength is recovering function that exists only in people's heads — the lesson from a near-miss that never became a document, the dependent who never filed a ticket. It is the archetype's frontline defense when records have decayed, and it is often the only source for the informal, human side of what a structure coordinates.
Its failure mode is memory's unreliability: recollections are reconstructed, self-serving, and subject to hindsight, so an interview can confidently report a rationale that is embellished or plain wrong. It is also exposed to incumbent bias — the people most attached to a structure are the ones most available to interview, and they may defend it sincerely but selectively, so the hypothesis skews toward retention. A related trap is the leading question that lets an informant simply ratify the interviewer's assumption. The classic misuse is running the interview as a courtesy consultation — asking "any objections?" rather than "what does this catch?" — which harvests permission instead of knowledge. The guarding discipline is to seek disconfirming accounts, triangulate across independent informants, and, where the stakes justify it, corroborate the recovered rationale against the documentary record.
How it implements the components¶
original_function_hypothesis— its primary output: an experience-grounded account of what the structure does and why, assembled from concrete incidents informants recount.dependency_and_stakeholder_register— the dependents, downstream parties, and harmed-if-removed stakeholders that experienced people can name from having watched the structure operate.
It does NOT implement persistence_evidence_signal — that is its nearest twin, Historical Rationale Reconstruction's territory: reconstruction reads the documented trail of the structure surviving past reforms, whereas this interview trades in living memory. Where both fill the original function hypothesis, the interview sources it from people and the reconstruction from records — and where this interview names dependents from recollection, the Dependency-Tracing Workshop instead traces them through the system's connections.
Related¶
- Instantiates: Latent Constraint Preservation Audit — the interview is the archetype's living-memory recovery step.
- Sibling mechanisms: Historical Rationale Reconstruction · Dependency-Tracing Workshop · Chesterton's Fence Review Gate · Compensating Control Matrix · Constraint-Loss FMEA · Deprecation with Rollback Window · Post-Removal Sentinel Dashboard · Removal Sandbox Trial · Silent Dependency Survey
Editorial Notes¶
Form Classification¶
Form family: Communication, Facilitation & Learning
Rationale: Structured questioning elicits tacit function and hidden dependencies from maintainers, operators, and long-time users.
Nearest alternative: Assessment, Review & Assurance — The accounts are later assessed, but the mechanism itself is communicative elicitation.
Review outcome: Adjudicated after independent review; high confidence.
Origin Attribution¶
Primary origin: Ethnography & Qualitative Methods
Origin pattern: Cross-disciplinary synthesis
Present-day reach: Universal
Rationale: Requirements recovery provides the target context, but eliciting tacit function through stories, probing, and cross-informant triangulation derives from qualitative interviewing.
Related originating lineages:
- Engineering & Design — Retained as a formative lineage identified independently as primary: Requirements recovery and reverse engineering use stakeholder interviews to uncover hidden functions before modifying inherited systems.
- Human-Computer Interaction — Contextual inquiry contributed questioning maintainers and users about actual system function.
- Organizational & Management Science — Knowledge-management practice materially shaped recovery of operational rationale before change.
Review resolution: Requirements recovery provides the target context, but eliciting tacit function through stories, probing, and cross-informant triangulation derives from qualitative interviewing. The source supports the selected provenance; the retained alternates record documented formative or independently established lineages, not downstream applicability alone. origin_mode=cross_disciplinary_synthesis because the mechanism joins contributions across those traditions. domain_reach=universal records application breadth separately from origin.
Encyclopedia synthesis: The exact catalogued form synthesizes established practice rather than reproducing a single standard historical label.
Review outcome: Researched adjudication after independent review; medium confidence.
Sources consulted:
- https://onlinelibrary.wiley.com/doi/10.1155/2014/124701 — Peer-reviewed experience report recovering legacy software design from stakeholder interviews when documentation is unavailable.
Notes¶
[n1] Tacit knowledge — Michael Polanyi's observation that "we can know more than we can tell" — is the knowing-how and hard-won judgment people hold without ever articulating it. It is exactly the kind of function that never reaches documentation, which is why interviewing the people who hold it is a distinct recovery method from reading records. ↩