Encyclopedia tooling release
Integrated seven new checks for the writing guide, draft checker, type facets, Smell and Culture, editorial governance, entry citations and heading order into the 168-check production gate.
CHANGELOG ARCHIVE · PAGE 3 OF 7
Entries 101–150, newest first.
Integrated seven new checks for the writing guide, draft checker, type facets, Smell and Culture, editorial governance, entry citations and heading order into the 168-check production gate.
Ran axe-core on every route type at desktop and mobile widths and fixed a heading-order issue on source and evidence pages.
Every smell page now shows a citation with its permanent identifier and revision; the knowledge graph citation names its digest.
Published draft roles, review separation, reserved decisions and conflict-of-interest rules for adoption.
Launched a third free Learn course on smellscapes, time, shared meanings, work and smell heritage.
Places and materials are now findable in search; exact entry names still rank first.
Parsed every entry's type label into controlled facets and fixed four entries whose curated type did not match.
Added metallic, resinous, marine, fermented and chemical to the descriptor vocabulary.
Editors can check a draft entry against every publication rule before it enters the pipeline.
Every entry and draft is now checked against a neutral, claim-free writing standard.
Integrated nine new checks for the entry standard, pipeline, identity rules, places and materials, evidence agenda, graph records, Introduction to Smell, the heritage record draft and correction notices into the 161-check production gate.
Correction notices now name every affected permanent identifier and appear on smell pages and graph records. None has been published.
Published a draft JSON Schema for olfactory heritage records that link to SmellBook identifiers, for discussion with museums and archives. No institution uses it yet.
Launched a second free SmellBook Learn course: five lessons built from the Atlas, quoting VERIFIED claims exactly.
Every knowledge-graph entity now has its own record with typed relationships and provenance at /data/graph/{type}/{key}.json.
Published one open research question for each entry without a VERIFIED claim. No new sources were added without a checked citation.
Added 41 places and 16 materials to the knowledge graph with smellscape, found-in and produced-by relationships.
Every name and alias now identifies one entry; look-alike entries are reviewed, and merges retire identifiers without reusing them.
New entries reach the Atlas only through draft, review and a publication check; identifiers are issued at publication.
Every published entry now meets a machine-checked standard, with development signals reported per entry.
Integrated nine new checks for identifiers, entities, relationships, evidence, the Spanish pilot, data rights, source intake, Learn and the graph export into the 152-check production gate.
Published every entity and typed relationship as a rights-gated dataset with permanent identifiers, provenance, licence status and a sha256 digest.
Launched a free SmellBook Learn course with lessons such as earthy vs mineral, built from the Atlas descriptor data.
Added a rights-aware source intake workflow and evidence gaps by category and type.
Classified every exportable field by rights class; citing a source never implies a right to republish it, and the dataset licence is stated explicitly.
A small hand-checked Spanish pilot with localized names, English fallback, hreflang and Spanish search.
Every claim and relationship now carries one of six evidence classes, and a contribution review model is designed with no live submissions.
Replaced untyped links with typed relationships that carry direction, meaning and provenance.
Descriptors, categories, types, contexts, claims and sources became first-class entities with permanent identifiers.
Every public record received a permanent smellbook: identifier that is never reused, with a resolver.
Integrated eight new checks — search quality, product truth, navigation, descriptors, personal resilience, applied-scent journeys, experience integrity and task-registry integrity — into the production gate.
Brought text colours to WCAG AA contrast, exposed map links to assistive technology, fixed landmarks and heading order, and linked every smell from another profile.
Smell profiles now link to the scent directions, intents and spaces that cite them, and space pages show each direction's Atlas references.
Fixed note and tag typing on unsaved smells, protected damaged local data, and added import preview with merge or replace and a clear confirmation.
Added Earthy, Soft and Cool descriptor pages and put every profile on the descriptor pages its own descriptors name.
Search understands everyday words such as relaxing, cozy and seaside, and 32 profiles gained widely used alternate names.
Search now folds plurals and adjective forms, suggests spellings for zero-result queries and is checked against regression fixtures.
Every page shares one navigation for Explore, Map, My SmellBook, Fragrance, Spaces, Stories and Labs, with section links underneath and a working phone menu.
A build check now rejects copy that contradicts the public capability matrix.
README, capability documentation and site copy now describe the live map, Personal SmellBook and Fragrance & Spaces tools accurately.
Integrated immutable Phase 189 audit retention, truthful Phase 189 → Phase 199 comparison and five new snapshot validators into the 135-check production gate.
Registered retained audit history across schemas, inventory, checksums, documentation, capabilities and public discovery while preserving the noindex RESEARCH boundary.
Published machine-readable retained snapshot and release-comparison contracts with SHA-256 digests over declared public audit inputs.
Added read-only packet comparison pages that trace retained candidate, review, context, lens and matrix inputs across releases.
Added a server-rendered Phase 189 → Phase 199 comparison with descriptive added, removed, changed and unchanged states.
Added a noindex history surface for retained audit snapshots and explicit reference-only releases where fingerprints were never preserved.
Added deterministic per-family and per-packet comparison summaries without turning change states into scores, rankings or review outcomes.
Added comparison semantics that use only explicitly retained audit inputs and never reconstruct missing historical state.
Added retained record-set identities, release references and explicit comparable/reference-only snapshot states.
Retained the exact 279-record Phase 189 Labs audit input projection before beginning the next strategy tranche.