Claude [Wed, 11 Nov 2026 20:08:01 +0000 (20:08 +0000)]
docs(tmdm): coverage matrix, relationship readers, default name types; run report
Add docs/tmdm-coverage-matrix.md: clause-by-clause TMDM coverage with the
consulted public texts (URL, date, SHA-256), keeping norm coverage and API
apart. Extend docs/reading-type-relationships.md with the readers for
TMDM 7.2/7.3 and with the two documented default name type identifiers,
which stay different identifiers and are never equated automatically.
Point to both from docs/reading-typed-topic-names.md and README.md.
Add reports/core-relationship-readers: the launcher logs before and after
the changes, the exit codes, the digests of the report files and an
execution report. The full offline launcher passed before and after
(exit 0); the 25 earlier report files are byte-identical; the Python
evaluation of the eleven CXTM baselines agrees with the readers. The XTM
suites were not run (they need SBCL 2.5.4 and patched CXML).
Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_014ewLGkKYU24V8KWmZKpjGx
Claude [Wed, 11 Nov 2026 20:06:29 +0000 (20:06 +0000)]
feat(tmdm): readers for the TMDM 7.2/7.3 relationships; tests for both default name types
Add six read-only functions to topic-maps.tmdm that recognize the
type-instance (TMDM 7.2) and supertype-subtype (TMDM 7.3) Associations the
CTM and XTM readers already create and store: type-instance-relationship,
supertype-subtype-relationship, and four Topic readers that return the
Associations in which a Topic is instance, type, subtype or supertype.
An Association is recognized exactly when its type and its two Role types
carry the clause's Subject Identifiers and it has exactly two Roles. The
readers are added API, not a conformance repair: they intern no Topic,
change nothing, evaluate no Scope and infer nothing. Sort names (7.4) are
deferred.
tests/core-relationships.lisp argues each expectation from its clause
(50 checks, eleven original isa/ako inputs).
scripts/check-core-relationships.py counts such Associations in the
unchanged CXTM baselines; it evaluates serialized data and is not a proof
of conformance.
tests/default-name-types.lisp (21 checks) shows that both documented
default name type Subject Identifiers, .../name-type and .../topic-name,
are preserved on import and in CXTM, coexist, and are never equated
automatically. It needs no production change. An explicit merge of the two
type Topics and the unchanged Core Subject set are recorded as a
characterization of current behaviour, not as a rule.
Both suites are registered in topic-maps.asd and run by the offline launcher.
Verified with SBCL 2.4.10: the unchanged baseline passed before these
changes, and both new suites pass. The full launcher has not yet been
re-run with them; the XTM suites were not run.
Co-Authored-By: Claude Fable 5.2 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_014ewLGkKYU24V8KWmZKpjGx
rgb [Wed, 28 Oct 2026 15:42:36 +0000 (16:42 +0100)]
perf(tmdm): index association normalization merge candidates
Replace the quadratic cross-Association scan in normalize-associations with
an indexed search for the next merge candidate. Pair selection order and
merge semantics are unchanged relative to 6162ed68.
Adds tests/association-index.lisp with 29 fixtures. Independently reviewed.
Not verified: Linux execution, the Stage 2 CTM, peak RSS and suitability
for the constrained NixOS host, high-arity performance, and the historical
Python Unicode/NFC checker mismatch (pre-existing, Python 3.14 / Unicode 16).
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
rgb [Wed, 28 Oct 2026 06:01:47 +0000 (07:01 +0100)]
perf(tmdm): index topic normalization merge candidates
normalize-topic-map chose the next pair of Topics to merge by comparing
all Topic pairs; each pair without a shared identifier called
topic-reified, which enumerates every information item. Reading the
1,855-line HyperDoc project CTM did not finish in 120 s.
next-topic-merge returns the same pair through an index built for one
search and dropped with it: identifier keys (item and subject identifiers
in one namespace, subject locators apart) and a table of reified items for
the nonconflicting-reifier preference. A Topic that reifies two items, or
a listed Topic that is not its own representative, falls back to
scan-next-topic-merge, the unchanged pairwise search, so the conditions
and the moment they are signalled stay as they were. Nothing outlives a
call, so no invalidation rule exists.
The project CTM now reads in about 29-31 s (median of three fresh runs in
two independent reproductions); its semantic signature and CXTM are
byte-identical to the 5a4a668 reference. The Association pair search is
the remaining cost and is left unchanged.
tests/normalization-index.lisp (120 checks): CXTM and rejection outcomes
recorded from unchanged b2eecffa, staged-import reifier preference,
retained handles, index-versus-pairwise comparison on random maps, merge
walks and CTM documents, stale references, a positive control and scale.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
rgb [Tue, 20 Oct 2026 21:38:59 +0000 (23:38 +0200)]
Document normative conformance decisions for Slice 22C
Review 13 originals with edition-qualified evidence: one A, no unconditional B, twelve C. Preserve registrations, corpus bytes and all historical reports. Record the complete fresh offline regression, 124 reader diagnoses, 108 executor controls and final integrity audit.
rgb [Tue, 20 Oct 2026 05:53:57 +0000 (07:53 +0200)]
feat(xtm): support item identifiers on name variants
Assign Variant itemIdentity locators through the existing global TMDM API after canonical Variant construction. Preserve identifier unions, native conflict conditions and retained handles. Register both unchanged 666-byte original CXTM contracts.
Retain full fresh offline SBCL 2.5.4 evidence: native API preflight, all 124 diagnostic comparisons, identity/merge/syntax and atomic conflict checks, and 60 executor fault controls. XTM: 62 PASS / 0 FAIL / 62 NOT-RUN; CTM unchanged at 262 / 4 / 7. No TMDM or other protected production changes.
rgb [Tue, 20 Oct 2026 04:13:08 +0000 (06:13 +0200)]
feat(xtm): support anyURI resource data on name variants
Resolve textual anyURI resourceData through the existing XTM IRI operations and native Variant API. Reject child elements and invalid datatype IRIs, preserve resourceRef semantics and canonical identity, and register the unchanged 621-byte original CXTM contract.
Retain full fresh offline SBCL 2.5.4 evidence, native resourceData/resourceRef equivalence and merge checks, all 124 diagnostic comparisons and 50 executor fault controls. XTM: 60 PASS / 0 FAIL / 64 NOT-RUN; CTM unchanged at 262 / 4 / 7. Corpus and historical reports remain unchanged.
rgb [Mon, 19 Oct 2026 21:24:19 +0000 (23:24 +0200)]
feat(xtm): support resource references on name variants
Resolve Variant resourceRef through the existing XTM IRI path and add-variant API, preserving canonical parents, proper scope unions and anyURI values without resource retrieval. Register both unchanged original CXTM contracts.
Retain fresh offline SBCL 2.5.4 regression, native identity/merge/syntax controls and all 124 diagnostic comparisons. XTM: 59 PASS / 0 FAIL / 65 NOT-RUN; CTM unchanged at 262 / 4 / 7. Corpus, historical reports and protected production subsystems remain unchanged.
rgb [Mon, 19 Oct 2026 20:48:18 +0000 (22:48 +0200)]
test(xtm): register variant item-identity merge original
Register merge-itemid-with-variants through the existing ASDF conformance harness, with native identity/parent/variant assertions and the exact 919-byte original CXTM comparison. Preserve canonical retained handles and source-key independence.
Retain fresh offline SBCL 2.5.4 regression evidence: XTM 57 PASS / 0 FAIL / 67 NOT-RUN; CTM unchanged at 262 / 4 / 7. No production code, original corpus, baseline or historical report changes.