Rendered from docs/reviews/ste-editorial-pass-findings.md in the Headwater
corpus. Every document on this half of the site is typed by the taxonomy
the descriptor names: corpus.json.
STE editorial pass — source findings to address
Point-in-time record, 2026-08-10. During the ASD-STE100 house-profile pass over specs 0–12 (commits 786c5e2, d03d613, 651a302), the editors flagged the items below as possible defects or open decisions in the source content. Editorial policy was to report, not fix, anything that needs author judgment. Two objective typos found in the same pass were fixed directly and are not listed here.
Possible wording defects
| Location | Finding | Suggested action |
|---|---|---|
| spec 4, degraded-controls discussion | "An uncovered control cannot hide" — context (gaps attach to obligations; controls degrade) suggests the intended word is obligation, not control | Confirm and correct the noun |
| spec 4, same section | "A reader who cannot tell the maintainer … is a control that does not exist" equates a person with a control; the intended referent is probably the missing feedback channel | Decide whether the metonymy is intended |
| spec 9, Q17 | "a determined reader can difference" uses difference as a verb; reads as intentional diff-terminology but is nonstandard | Keep deliberately or change to "diff"/"compare" |
| spec 10, B.2 | Source read "every enum facet's value document how to choose"; the pass rendered it "every enum facet value must document how to choose it" — the original may have been a typo for "facet's values document" | Confirm the rendered meaning is the intended one |
| spec 3, absence-class sentence | "is the same absence class" was rendered "is in the same absence class"; flag in case the identity phrasing (a class, not membership) was intentional | Confirm |
| spec 11, §A intro | "the validate / audit split arrived at [spec 6]" is the source author's own construction and reads oddly |
Optional rephrase |
Terminology and spelling decisions (corpus-wide rulings needed)
The last three rows arrived later the same day, from the pass that built the glossary rather than from the STE pass. An index of every term is what makes a name collision visible, so the glossary found them. They are recorded here because this table is where corpus-wide terminology rulings are tracked.
| Question | Evidence | Options |
|---|---|---|
| American or British spelling | Corpus was British (organisation, artefact, customise) with drift: "knowledge organization" in spec 2 and spec 7; "judgment" and "judgement" both occurred. STE rule 1.14 prescribes American unless a directive says otherwise | Resolved 2026-08-10: American ruled (CLAUDE.md), corpus swept, reviews left as written |
| "register", two senses | Spec 1 uses register for the voice regime and for the register projection — one term, two concepts, a rule-1.11 violation in spirit, sharpened by the recent retirement of the standalone register concept | Resolved 2026-08-10: the voice sense renamed to "voice". The register projection and the registers it generates keep the word |
| "fails open" | Spec 5 says routing "fails open" meaning it stays silent below the confidence threshold; conventional security usage would call that failing closed. The usage is consistent with design principle 7 in spec 0 | Keep as project vocabulary (documented in principle 7) or align with conventional usage |
| "gap", two senses | A disposition on an obligation (no control discharges it yet) and a value of evidence_basis on a document (spec 3). Both senses are then said to "appear in the gap register", so the collision may not be only lexical: the specs do not say whether that is one register or two |
Resolved 2026-08-10: the evidence_basis value renamed to unevidenced, parallel to evidenced and reconstructed. gap is now only the disposition. Still open: whether spec 3 and spec 4 mean the same gap register, or two that share a name — recorded in spec 3 |
| "audit", two senses | taxonomy audit (spec 6) measures a schema against a corpus and is a CLI verb. A conformance audit (spec 4) is a periodic human sample that asks whether a specification is still true of the system |
Resolved 2026-08-10: the spec 4 activity renamed to "accuracy audit", which states what it checks. taxonomy audit keeps the CLI verb |
| "conformance", two senses | Spec 4 uses it for whether the corpus still describes the system. Spec 7 uses it for whether a consumer wired the published method, with headwater conformance as the command. The two ask unrelated questions |
Resolved 2026-08-10 with the row above: "conformance" is reserved for the spec 7 adoption sense, which owns the command name |