Dao De Jing test bed: PublicDomain provenance fields, Legge slice, first golf-loop number #10
Loading…
Reference in a new issue
No description provided.
Delete branch "weksa/dao-de-jing-test-bed"
Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?
Dao De Jing as a lowering test bed. Credit: the analogy is Meta's, from #weksa:gamecult.org on 2026-11-06 (Matrix event
$2_P0mbtbnnsP4I61UhVMlatuDSTjFXMOuGa4SSbhGmc: many varied human translations as precedents, public-domain ones only as the scoring corpus), with the term-modelling note in$E4VHwS-zFEk9Fa-daFX8MkfoySn0iX0m1grltEJ9A8Y.What changes
Source::PublicDomainrequireslicence,jurisdiction,translatorandedition_yearbesides work, author and edition_url. An empty or unreadable string, or year 0, is refused by the store. CLI flags--licence --jurisdiction --translator --edition-year; tests instores.rsandcli.rs; corpus schema regenerated.data/daodejing/: a separate test-bed store set holding five Legge items (Gutenberg #216, 1891) with full provenance. Only Legge's words are stored: no Gutenberg header, licence text or trademark (1.E.2), and the eBook is not committed. Open ruling for a human: whether theedition_urlcitation of gutenberg.org counts as a "reference"; if so it moves to the print edition.weksa inventory daoandweksa packet dao.docs/targets/dao-de-jing-test-bed.md.First closeness number
Mean chrF 71.96, meaning cosine 0.925 over five items, against a mismatched-pair baseline of chrF 14.1 and meaning 0.458. This is not a measure of the interlingua: one author wrote packets, inventory and lowerer rows with each reference open, n = 5, and the item lowered by a Train-minted frame scored lowest (chrF 44.18, meaning 0.831). The scorer cannot see stance (the lowerer does not render it), a dropped negation or question force (the rhetorical question lowers to a statement), or "the heaven"-style table defects. The doc has the full list.
Not done here
polarityand cannot punctuate a question, so negation and rhetorical questions in the Dao cannot be lowered yet.scene_revision()is also numbered 3; the test-bed stores do not apply it, and it needs renumbering where the two meet.Test:
cargo test --workspacegreen.🤖 Generated with Claude Code
2308b91) into weksa/dao-de-jing-test-bed 8a78761a7afd18056, phones line) into weksa/dao-de-jing-test-bed 34ce2dc413Main (through #16 and #18) merged into the branch; Dao frames renumbered to 4..8 and concepts to 7..27 to clear main frame:3/concept:6; data/daodejing inventory and packets regenerated by the CLI. Head
34ce2dc: cargo test --locked --workspace all ok, run by Hands and independently by a Soul pass (holds; lows: no door enforces Gutenberg boilerplate absence, provenance checks shape not plausibility; id/revision overlap with future main revisions stays loud via the gloss check). Not run: scorer numbers (needs Ollama). Merging.