X-Git-Url: https://dreyeck.freedombox.rocks/gitweb/topic-maps.git/blobdiff_plain/66664dbff6275d60dab1bc6a232fa70f7b916b8c..8036a2b1ce009218f33d300edcc5f715ec1d1124:/README.md?ds=inline diff --git a/README.md b/README.md index 5316919..b933e95 100644 --- a/README.md +++ b/README.md @@ -1,23 +1,31 @@ -# Common Lisp Topic Maps — Slice 2 +# Common Lisp Topic Maps — Slice 4 -A working ANSI Common Lisp reader for the first CTM identity/default-name -subset, with same-map Topic/Name merging in a minimal TMDM model. HyperDoc -remains a possible future client and is neither the codebase nor a dependency. There is no CXTM -canonicalizer, association/template implementation or cross-document merge engine. +A working ANSI Common Lisp reader for CTM identity/default-name syntax and +unary/binary Associations, with same-map merging of Topics, Names, Associations +and Roles in a bounded TMDM model. HyperDoc remains a possible future client +and is neither the codebase nor a dependency. Slice 4 adds a pure, independent +CXTM canonicalizer for this model. Templates and cross-document merging remain +outside the implementation. Systems: `topic-maps` (the retained foundation), `topic-maps/tests`, `topic-maps/tmdm`, `topic-maps/ctm`, `topic-maps/ctm/tests` and -`topic-maps/ctm/merging-tests`. +`topic-maps/ctm/merging-tests` and `topic-maps/ctm/association-tests`, plus `topic-maps/cxtm` and +`topic-maps/cxtm/tests`. The core reader/model uses ANSI Common Lisp; the foundation/tooling uses ASDF/UIOP and Python's standard library. No Quicklisp packages are required. -Start with [Reading Topic Identity and Merging](docs/reading-topic-merging.md) +Start with [Reading Canonical Topic Maps](docs/reading-canonical-topic-maps.md) +for the CXTM API, normative clauses, base-locator policy and actual byte comparisons. +[Reading the First CTM Associations](docs/reading-first-ctm-associations.md) +for actual Association/Role items, late Topic merges and current subset boundaries. +[Reading Topic Identity and Merging](docs/reading-topic-merging.md) for the current API, reference replacement, identifier tails and exact subset boundaries. See [Reading the First Five CTM Documents](docs/reading-first-five-ctm-documents.md) for original sources, actual objects, normative clauses, baseline evidence and copy-pasteable DESCRIBE forms. [API and boundaries](docs/slice-1-api-and-boundaries.md) document the accepted Slice 1 baseline; the Slice 2 document supersedes its statements about absent identity tails, merging and immutable Name references. +The Slice 3 document supersedes historical statements about absent Associations/Roles. The pinned suite targets the **CTM draft dated 2010-03-31**. Passing it alone does not establish full ISO/IEC 13250-6:2010 conformity. @@ -40,6 +48,8 @@ In a new Common Lisp process, with this checkout selected: (asdf:test-system "topic-maps/tests") (asdf:test-system "topic-maps/ctm/tests") (asdf:test-system "topic-maps/ctm/merging-tests") +(asdf:test-system "topic-maps/ctm/association-tests") +(asdf:test-system "topic-maps/cxtm/tests") (topic-maps.test-foundation:write-validation-report *corpus* #p"/tmp/topic-maps-inventory.json") ``` @@ -81,19 +91,24 @@ CXTM baselines only under `ctm/baseline`. Auxiliary files are reported separatel ## Current validation and historical foundation The complete pinned corpus remains unchanged: 231 valid CTM inputs, 42 invalid -inputs, 231 CXTM baselines and 22 CTM auxiliary files. Sixteen selected semantic -cases pass; their imported canonical contracts remain NOT-RUN. Only -`ctm/invalid/prefix-undefined.ctm` is PASS under its CTM rejection contract; -272 imported contracts remain NOT-RUN. Unsupported constructs and I/O errors +inputs, 231 CXTM baselines and 22 CTM auxiliary files. Twenty-four selected semantic +cases now also pass actual CXTM byte comparisons. These 24 `:CXTM-EQUAL` +contracts are PASS, alongside `ctm/invalid/prefix-undefined.ctm` under its specific +CTM rejection contract. The remaining 248 imported contracts are NOT-RUN. Unsupported constructs and I/O errors cannot count as that required rejection. The full launcher runs the retained 54 foundation checks, two Python integrity -fixtures, 123 retained identity checks, 646 merge checks, imported-report validation and selected baseline -property evidence. A report path also produces `.semantics.json` with -separate model observations, plus `.merging.json` for the new cases. +fixtures, 123 retained identity checks, 646 merge checks, 657 Association checks, +imported-report validation and selected baseline +property evidence, 445 CXTM checks and 16,294 independent Unicode NFC oracle +comparisons. The oracle needs Python UCD 15.0 or 15.1 (tables use UCD 15.1.0). +A report path also produces `.semantics.json` with +separate model observations, plus `.merging.json`, `.associations.json` +and `.cxtm.json` (complete actual/expected XML and byte counts). The merge checks include 128 graph/order variants with an independent reachability oracle. Reports are deterministic across the tested SBCL versions and checkout -relocation. See [Slice 2 validation](docs/slice-2-validation.md), +relocation. See [Slice 4 validation](docs/slice-4-validation.md) for current evidence. +[Slice 3 validation](docs/slice-3-validation.md), [imported contracts](reports/slice-1/validation.json) and [semantic observations](reports/slice-1/semantics.json). @@ -104,4 +119,7 @@ about absent reader/model systems describe Slice 0, not the current branch. Slice 1 is built from `9c83db34eb1ba8091b004433f088056717dbaa7e` on `codex/ctm-slice-1`. Slice 2 starts from the accepted Slice 1a commit `fb60afd548d679d83fd8ec76bc5ff384c1e23e5f` on `codex/ctm-slice-2`. -Associations/Roles are not part of this slice; await review before continuing. +Slice 4 starts from `35648c4dcc71e8dc6448b64abcacb438e53d64a2` on +`codex/ctm-slice-4`. +Slice 3 starts from accepted `66664dbff6275d60dab1bc6a232fa70f7b916b8c` on +`codex/ctm-slice-3`. Await review before continuing to the next slice.