Search and Documentation Scan¶
Evidence scan — instantiates Marked Default Audit
Sweeps a whole corpus of artifacts to quantify how often a baseline is silently assumed, so the audit rests on breadth of evidence rather than a single anecdotal label.
Search and Documentation Scan is a breadth instrument. Instead of reading one label closely, it sweeps many artifacts — docs, examples, tickets, records — to measure how pervasively an unnamed baseline is assumed. It assembles the corpus-wide comparison set of occurrences and surfaces the unmarked default precisely by its ubiquity and its lack of any qualifier. Its one idea is coverage: it answers "how often, across how much," never "what does this one name imply" or "what should we do." That makes it the corrective to anecdote — the reason an audit can say the asymmetry is systemic rather than a single vivid case.
Example¶
A developer platform's API documentation is scanned for OS and shell assumptions. The scan greps every code sample and prose page for path styles, shell commands, and platform notes. It finds 340 examples that write Unix-style paths and bash commands with no platform note at all, 18 that explicitly say "(on Windows, use ...)," and zero that mark the Unix case as a choice rather than the given. The comparison set is that whole population of examples; the unmarked default — Unix — is surfaced not by any single label but by its silent presence in roughly 95% of samples; and the frequency breakdown is the asymmetry evidence: one platform assumed by default everywhere, the alternative always tagged as the exception. That corpus-wide count, not any one page, is what a revision rule can rest on. The scan gathers; it interprets and rewrites nothing.
How it works¶
- Define the corpus and the query patterns — qualifiers, platform notes, tags, exception markers, and their absence.
- Sweep and count. Tally where a baseline is silently assumed versus where alternatives are explicitly marked.
- Assemble the occurrences into a comparison set and quantify the one-sidedness.
- Surface the default by its silence — the baseline that is everywhere yet never qualified.
Its distinctive quality is that it detects the default by absence of marking at scale, which no close reading of one artifact can establish.
Tuning parameters¶
- Corpus scope — one repository versus all docs, tickets, and records. Wider is more representative but noisier.
- Pattern precision — strict queries (few false positives, miss variant phrasings) versus loose ones (catch more, more noise).
- Aggregation unit — per file, per example, or per token. Changes what "how often" means.
- Frequency threshold — how dominant a silent baseline must be before it is flagged as the default rather than a common case.
When it helps, and when it misleads¶
Its strength is that it replaces "I noticed a bad label once" with a measured pattern — it is the corrective to the availability heuristic[n1], which makes one vivid anecdote feel like a trend. Its failure mode is that counts are not meaning: a high frequency of an unmarked baseline may be perfectly justified (the platform really is Unix-first), and the scan cannot tell justified ubiquity from smuggled hierarchy. The classic misuse is treating raw counts as a mandate to symmetrize everything that is common. The guarding discipline is to hand the frequency evidence to an interpretive mechanism — Naming Markedness Audit or Inclusive Language Markedness Review — and a justification check before any rule or rewrite follows.
How it implements the components¶
comparison_set— assembles the corpus-wide population of parallel occurrences to be compared.unmarked_default— surfaces the silently assumed baseline by its qualifier-free ubiquity across the corpus.asymmetry_evidence— quantifies how one-sided the marking is at scale, turning anecdote into a measured pattern.
It does not infer the status_implication of the pattern or execute any label_or_category_revision — it supplies breadth evidence to Naming Markedness Audit for interpretation and to Classification Markedness Audit for repair.
Related¶
- Instantiates: Marked Default Audit — grounds the audit in corpus-wide evidence rather than isolated labels.
- Sibling mechanisms: Naming Markedness Audit · Default Label Review · Inclusive Language Markedness Review · Classification Markedness Audit · Symmetry Labeling Matrix · Style Guide Revision · Stakeholder Interpretation Review
Editorial Notes¶
Form Classification¶
Form family: Assessment, Review & Assurance
Rationale: Search and Documentation Scan operates as a bounded evaluation of existing evidence or work that produces a finding or disposition because it sweeps a whole corpus of artifacts to quantify how often a baseline is silently assumed, so the audit rests on breadth of evidence rather than a single anecdotal label.
Independent corroboration: The frozen evidence defines Search and Documentation Scan as 'Sweeps a whole corpus of artifacts to quantify how often a baseline is silently assumed, so the audit rests on breadth of evidence rather than a single anecdotal label', so its operative form is Assessment, Review & Assurance.
Nearest alternative: Analysis, Modeling & Optimization — Search and Documentation Scan includes features of an analytical, modeling, inference, comparison, or optimization procedure that derives insight or a solution, but its defining operation is a bounded evaluation of existing evidence or work that produces a finding or disposition.
Review outcome: Independent reviewer agreement; medium confidence.
Origin Attribution¶
Primary origin: Library & Information Science
Origin pattern: Cross-disciplinary synthesis
Present-day reach: Universal
Rationale: Corpus-wide searching and evidence enumeration are information-retrieval practices.
Related originating lineages:
- Accounting & Auditing — Audit evidence gathering independently replaces anecdote with broad documentation.
- Computer Science & Software Engineering — Computer science and software-engineering practice supplies a parallel or contributing lineage for the mechanism's defining operation: sweeps a whole corpus of artifacts to quantify how often a baseline is silently assumed, so the audit rests on breadth of evidence rather than a single anecdotal label.
- Data Science & Analytics — Text analytics materially quantifies silent baseline assumptions at scale.
Review resolution: The blind reviewers agree that library_information_science is the primary origin and differ only on alternate origin disagreement, origin mode disagreement, domain reach disagreement. I preserve every independently explained alternate from both records rather than imposing a numeric cap. I retain cross_disciplinary_synthesis because the combined record shows material contributions from several lineages. The broader reach of universal records portability separately from historical provenance, and encyclopedia_synthesis=true preserves the affirmative synthesis judgment where either reviewer identified one.
Encyclopedia synthesis: The exact catalogued form synthesizes established practice rather than reproducing a single standard historical label.
Review outcome: Reconciled after independent review; medium confidence.
Notes¶
[n1] The availability heuristic (Tversky & Kahneman) — judging how common something is by how easily an example comes to mind. It makes one memorable label feel like a systemic pattern, and one absent-marking case feel like a fluke; a corpus-wide count is exactly what corrects for it. ↩