Aligning outliers in constants/float16 with namespace majority patterns (random namespace pick, seed 20260429).

Description

This pull request normalizes package.json metadata across 15 outlier packages in @stdlib/constants/float16 so that they conform to the patterns observed in the rest of the namespace. All edits are purely additive metadata fixes — keyword additions and one trailing-period restoration — that do not change any observable behavior, public signature, or test expectation.

Namespace summary

• Target namespace: @stdlib/constants/float16
• Member count: 53 (no autogenerated members; all 53 eligible)
• 75% conformance threshold: 40 / 53 packages
• Features analyzed: file tree, package.json top-level keys, package.json scripts/stdlib subkeys, package.json keywords, package.json description shape, manifest.json shape, README ##/### section sequence, docs/types/index.d.ts shape, docs/repl.txt shape, dependencies / devDependencies sets.
• Features with a clear majority and detected drift:
  • package.json keywords (set-valued; 10 majority keywords, 14 outliers)
  • package.json description trailing period (scalar; 52/53 = 98% conformance, 1 outlier)
• Features with a clear majority and zero drift: package.json top-level key set (18 keys, 100% uniform); file tree of 8 universal files (100% uniform); empty dependencies / devDependencies / scripts / stdlib (100% uniform); docs/types/index.d.ts declare const FLOAT16_X: number; shape (100% uniform).
• Features excluded due to no clear majority:
  • README ## section sequence — only 26/53 (49%) follow [Usage, Examples]; remainder add C APIs (15) or See Also (13). The C APIs section is conditional on the package having a C native binding (manifest.json + include/*.h), which itself is intentional and not drift. See Also cross-links are optional and per-package.
  • manifest.json / include/*.h files (15-16/53) — present only on integer-valued constants that ship C bindings; legitimately absent on the floating-point constants. Conditional, not drift.
  • Constant-level JSDoc @see tag — borderline at 41/53 = 77%, but adding @see requires per-package research (Wikipedia / OEIS links) and is non-mechanical. Excluded as out of scope for an additive metadata pass.
• Features not applicable to this namespace: semantic features (public function signature, validation prologue, error-construction style, return kind) — constants/* packages export single values rather than functions, so no semantic majority vote was meaningful. Documented and skipped.

Per-outlier corrections

constants/float16/ln-two

Added float keyword (51/53 = 96% of constants/float16 siblings carry it). Closes a one-keyword gap; no other changes.

constants/float16/max-base10-exponent

Added float16 keyword (44/53 = 83%). Brings the package in line with the rest of the namespace; metadata-only.

constants/float16/max-base10-exponent-subnormal

Added float16 keyword (44/53 = 83%). Same drift pattern as the rest of the *-base*-exponent* group.

constants/float16/max-base2-exponent

Added float16 keyword (44/53 = 83%). Standardizes discoverability across the exponent constants.

constants/float16/max-base2-exponent-subnormal

Added float16 keyword (44/53 = 83%). Mirrors the fix applied across the *-base*-exponent* group.

constants/float16/max-safe-integer

Added mathematics and math keywords (51/53 = 96% each). Both safe-integer constants in the namespace had the same drift; this is the max- half.

constants/float16/min-base10-exponent

Added float16 keyword (44/53 = 83%). Same drift pattern as its max- counterpart.

constants/float16/min-base10-exponent-subnormal

Added ieee754 keyword (52/53 = 98%). The only package in the namespace lacking the ieee754 keyword; the constant is defined directly by the IEEE 754 standard.

constants/float16/min-base2-exponent

Added float16 keyword (44/53 = 83%). Same drift pattern as the rest of the exponent group.

constants/float16/min-base2-exponent-subnormal

Added float16 keyword (44/53 = 83%). Completes the exponent-group cleanup.

constants/float16/min-safe-integer

Added mathematics and math keywords (51/53 = 96% each). Companion fix to max-safe-integer.

constants/float16/num-bytes

Added float16 keyword (44/53 = 83%). The package describes the byte size of a float16 value; the keyword belongs.

constants/float16/precision

Added float16 keyword (44/53 = 83%). Describes the float16 significand precision; the keyword belongs.

constants/float16/sqrt-three

Added float keyword (51/53 = 96%). The package already carries float16 and floating-point; float was the lone gap.

constants/float16/ln-half

Added the trailing period to the description field (52/53 = 98% of namespace siblings end the description string with a period). Pure punctuation fix.

Questions

No.

Other

Validation audit

Three independent agents reviewed every flagged outlier before any commit was made:

• Semantic review (opus): for each outlier, inspected the package's lib/index.js and README.md and confirmed the missing keyword (or trailing period) reflects stale metadata rather than a deliberate semantic difference. Verdict on all 15 outliers: confirmed-drift.
• Cross-reference review (opus): for each outlier, verified that no test/test.js, examples/index.js, or sibling package documentation references the affected metadata in a way that would require a cascading change. Verdict on all 15 outliers: safe (mechanical, no cascade).
• Structural review (sonnet): for each outlier, confirmed the majority pattern is the right pattern to apply and that the "missing" item is not legitimately inapplicable. Verdict on all 15 outliers: confirmed-drift.

What was deliberately excluded from this pass:

• README section drift on C APIs and See Also — conditional or optional, not enforceable as a single majority pattern.
• Constant-level JSDoc @see tag drift — borderline majority and non-mechanical (would require per-package research for the right reference link).
• The single keyword inherited by ln-half and any other extras that are not part of the majority set — outside the scope of an additive metadata pass.
• All semantic features — not applicable to a value-exporting constants/* namespace.

The full local audit report (per-feature distributions, full outlier lists, drop reasons, and the random seed) is preserved at ~/drift-reports/drift-constants-float16-2026-04-29.md.

Process notes

This PR was opened on branch claude/compassionate-ritchie-CuGzQ because the harness pre-assigned that branch for this session; the routine's preferred philipp/drift-<namespace>-<date> form is documented in the local report.

Checklist

• Read, understood, and followed the contributing guidelines.

AI Assistance

When authoring the changes proposed in this PR, did you use any kind of AI assistance?

• Yes
• No

If you answered "yes" above, how did you use AI assistance?

• Code generation (e.g., when writing an implementation or fixing a bug)
• Test/benchmark generation
• Documentation (including examples)
• Research and understanding

Disclosure

This PR was authored end-to-end by Claude Code running the cross-package API drift detection routine. Random namespace selection, structural feature extraction, majority-vote computation, and three-agent validation (two opus, one sonnet) ran automatically; the per-package metadata edits, commits, and PR body were composed by Claude Code based on the validated drift findings. A human maintainer should audit before promoting from draft.

@stdlib-js/reviewers

Generated by Claude Code