writeonce/docs/examples/skill-catalog/README.md
shoney.arickathil 5941a092f7 feat(query-corpus): iteration 9g corpus #1 — skillhost needs no new query grammar
- resolved 9g's forks EMPIRICALLY against the running compiler:
  - count(<query>) and len(<query>) already work (fork 2 collapses to
    zero code)
  - skillhost's correlated NOT EXISTS is a backlink emptiness in
    writeonce (`where len(x.children) == 0`), using only 9b machinery
    (fork 1) — verified on a self-referential ?ref/backlink table
  => corpus #1 forces NO new grammar; per the method ("add only what a
     corpus uses"), exists/not-exists was NOT built
- docs/examples/skill-catalog: mirrors skillhost's `skills` table
  (name @unique, description/location/root, parent ?ref Skill, children
  backlink) and translates all five of its SQL statements 1:1
  (insert+dup-trap, get-by-name, list, roots via backlink-emptiness,
  count); scripts/skill-catalog-accept.sh 7/0, WAL-durable, dup trap
  persists across restart
- fixture run/db-query-corpus (count(query) + backlink NOT EXISTS);
  just skill-catalog module; target/ gitignored
- general exists/not-exists left unbuilt and recorded as "enters when a
  corpus forces a non-relation correlation"
- gates: oop-e2e 80/0, woc-test 566/0, skill-catalog 7/0; story + board
  record the finding

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
(cherry picked from commit 4c82461634d45f11eca1a252702031c03d999c7f)
2026-09-15 01:15:30 +02:00

30 lines
1.9 KiB
Markdown

# skill-catalog — the query grammar corpus (iteration 9g)
> Corpus #1 for the query-grammar method (story
> [09g](../../stories/language-runtime-database/09g-query-grammar-corpus.md)):
> take a real application backed by an embedded SQL database, translate its
> every statement to the writeonce query surface, and add only the grammar it
> forces. The application is `~/projects/skillhost` (a C++ MCP host whose
> in-memory SQLite holds its skill catalog).
**Result: skillhost forced no new grammar.** Its entire SQL footprint — one
table, a single-row `INSERT`, and four `SELECT`s (whose only non-trivial
features are `COUNT(*)` and a correlated `NOT EXISTS`) — is expressible on the
surface iteration 9b already shipped. The five statements, translated:
| skillhost SQL (`src/catalog/catalog.cpp`) | writeonce | mode |
| --- | --- | --- |
| `INSERT INTO skills (…) VALUES (?,…)` + `SQLITE_CONSTRAINT` dup check | `insert Skill { … }` + `try…catch` on the `@unique` trap | `seed` |
| `SELECT … WHERE name = ?` | `from x in Skill where x.name == n take 1 select x` | `get <name>` |
| `SELECT … ORDER BY name` | `from x in Skill order by x.name select x` | `list` |
| `SELECT COUNT(*) FROM skills` | `count(from x in Skill select x)` | `count` |
| `SELECT … WHERE NOT EXISTS (SELECT 1 FROM skills c WHERE c.parent = s.name) ORDER BY name` | `from x in Skill where len(x.children) == 0 order by x.name select x` | `roots` |
The one translation choice: skillhost's correlated `NOT EXISTS` (skills that
are nobody's parent) becomes a **backlink emptiness** — `Skill.children` is the
inverse of `parent`, and `len(x.children) == 0` is the childless test. No
subquery construct is needed; the general `exists`/`not exists` is deferred
until a corpus uses a correlation a backlink cannot express.
`skill-catalog seed | list | roots | count | get <name>`, WAL-durable under
`WO_DATA`. Acceptance: `scripts/skill-catalog-accept.sh`.