Jamaican Music Master Catalogue — Seed v1.0

Purpose: A project-ready starting catalogue for historians, archivists, researchers, and web developers building a comprehensive database of Jamaican music: genres, artists, songs/recordings, riddims, albums/releases, labels, producers, and source evidence.

Important: This document is not claiming to contain every Jamaican song, artist, album, or riddim. It establishes the data model, source hierarchy, normalization rules, verified anchor counts, and a seed dataset so the project can be expanded without losing provenance or creating duplicate totals.

What “v1.0” means here. v1.0 is the version of this reference document and of the seed data it locks (see the manifest below): the data model, vocabularies, counting rules and registries are stable enough for historians and developers to build on, and every seed file is checksummed against this document. It is not a claim that the catalogue itself has reached the §30 “v1.0 catalogue” coverage target — the public hub still says “seed”, every large figure is still an ESTIMATE, and coverage statistics are published next to every count.

Revision history

Seed manifest (v1.0 lock)

The checked-in seed files under src/data/catalogue/ are the machine-readable form of the registries in this document. Their checksums are recorded here and in src/data/catalogue/manifest.json; npm test and npm run catalogue:manifest -- --check fail if any file changes without regenerating both, so the document version can never silently drift from the data.

Seed v1.0 · manifest generated 2026-10-02 · src/data/catalogue/manifest.json

Seeded totals: 137 artists (53 hand-curated + 84 MusicBrainz neighbourhood), 240 aliases, 104 memberships, 13 riddims (3 candidates), 34 recordings, 2894 release groups (2135 albums; 9 hand-curated), 10 labels, 37 genres, 17 sources.

1. Project goal

Build the closest possible historical census of:

• Jamaican musical genres and subgenres.
• Unique musical works/songs.
• Unique audio recordings and versions.
• Unique album/release-group concepts.
• Individual releases/pressings/editions.
• Named riddims and riddim families.
• Artists and groups.
• Producers, studios, labels, musicians, and other contributors.
• Jamaican-born artists.
• Artists of Jamaican heritage born outside Jamaica.
• Non-Jamaican artists who make music in Jamaican genres.
• Relationships among songs, versions, riddims, artists, albums, labels, producers, studios, and genres.

The final system should support both:

1. Historical research — evidence, citations, corrections, uncertainty, alternate names, oral-history notes.
2. Computational counting — deduplicated totals by genre, decade, artist origin, riddim, label, producer, etc.

2. Non-negotiable counting rules

2.1 A “song” is not automatically a “recording”

The database should distinguish at least:

• Work / composition — the underlying song or composition.
• Recording — a specific recorded performance/mix.
• Track — a recording as it appears on a particular release.
• Release group / album concept — the logical album/single/EP project.
• Release — a particular pressing, edition, format, territory, or reissue.

This follows the MusicBrainz entity model.

Example

Catch a Fire is one album concept/release group, but it has many separate releases in different countries, years, and formats.

Do not count every CD, LP, digital reissue, remaster, and territory edition as a new album.

2.2 Riddims must be modeled independently

One riddim can support hundreds of different vocal, deejay, dub, or instrumental recordings.

A riddim record should therefore have:

• Canonical riddim name.
• Alternate names.
• Parent riddim/family.
• Earlier source rhythm if applicable.
• First known use.
• Producer/creator.
• Studio/band if known.
• Derived/remade riddims.
• Linked recordings.
• Historical notes.
• Evidence level.

Do not count every song on one riddim as a new riddim.

2.3 Artist identity must be deduplicated

An artist may appear as:

• Legal name.
• Stage name.
• Alias.
• Group member.
• Solo act.
• Featured act.
• Producer alias.
• Dub credit.
• Sound-system identity.

Each human being or group requires a canonical internal ID.

Do not infer birthplace, nationality, ethnicity, or Jamaican heritage from a name, photograph, accent, or musical style. Use a documented source.

2.4 Genre membership is many-to-many

A recording may reasonably be tagged with more than one genre.

Example:

• Roots reggae
• Dub
• Reggae

For global totals, choose either:

• a single primary_genre_id, or
• deduplicate by recording ID before adding genre totals.

Never simply add all genre counts together.

3. Evidence/status vocabulary

Use one of these statuses on every historically important field.

Recommended field-level confidence:

high | medium | low | unknown

4. Verified database anchors

These are useful minimum/reference points, not final global totals.

Live-database warning

Specialist databases change over time. For example, Riddim Guide’s statistics page and an individual riddim page may show slightly different totals during different crawl/update moments.

Always save:

• source URL,
• accessed date,
• source-specific ID,
• source count at access time.

5. Working Jamaican music taxonomy

This is a research taxonomy, not a claim that genre boundaries are universally agreed upon.

5.1 Traditional, folk, ritual, and heritage traditions

5.2 Recorded Jamaican popular-music lineage

5.3 Related Caribbean categories that must not silently contaminate totals

6. Artist table schema

Recommended database fields:

artist_id
canonical_name
sort_name
identity_type                  # person | group | sound_system | orchestra | unknown
aliases[]
legal_name
birth_date
birth_place
birth_country
death_date
jamaican_origin_bucket         # jamaican_born | jamaican_descent_abroad | international_non_jamaican | mixed_group | unknown
jamaican_parentage_note
active_start
active_end
genre_ids[]
group_memberships[]
musicbrainz_artist_id
wikidata_id
discogs_artist_id              # reference only; obey provider terms
riddimguide_artist_url
official_url
source_ids[]
field_confidence{}
research_notes

7. Seed artist registry

Origin classifications are intentionally not filled unless separately researched. This prevents accidental ancestry/nationality inference.

MusicBrainz identity keys (sidecar artist-mbids.json). 49 of the 53 rows carry an accepted MBID (exact-name match + single Jamaica/genre-hinted candidate); A000026 Lynn Taitt, A000036 Lord Fly, A000037 Harold Richardson and A000038 Sugar Belly Walker are in the review queue. An MBID is an identity key only — it never changes the origin bucket.

7.1 MusicBrainz neighbourhood artists (generated, v1.0)

Phase B (§19) follows one member of band relationship hop from every seed artist with an accepted MBID and imports the counterpart as a DATABASE_DERIVED artist row citing SRC016, together with a hist_artist_memberships row (dates and roles as MusicBrainz records them) and the artist’s MusicBrainz aliases. These 84 rows (36 persons, 48 groups) live in the generated file src/data/catalogue/musicbrainz/seed-neighbourhood.json with stable ids A100001–A100084; they are not hand-curated, carry no genre links (MusicBrainz tags are never imported as genres) and — like every seeded artist — have origin_bucket = unknown.

Artist-history task

For each artist, historians should verify:

1. Legal/canonical identity.
2. Birthplace.
3. Birth country.
4. Jamaican-parentage/heritage claim where relevant.
5. Aliases.
6. Groups.
7. Active period.
8. Discography.
9. Genre periods.
10. Source hierarchy.

8. Riddim table schema

riddim_id
canonical_name
aliases[]
parent_riddim_id
source_recording_id
source_work_id
first_known_year
first_known_release_id
creator_artist_ids[]
producer_artist_ids[]
studio_ids[]
band_musician_ids[]
genre_ids[]
riddimguide_url
riddimsworld_url
alternate_database_ids{}
known_recording_count
count_source
count_accessed_date
lineage_status
source_ids[]
confidence
research_notes

9. Seed riddim registry

Counts below are a snapshot from Riddim Guide’s statistics page, not lifetime/final totals.

Riddim lineage candidates from Jamaica Information Service

These should be entered after detailed historical verification:

10. Song/work/recording schema

A single table will eventually become too ambiguous. Prefer three linked tables.

10.1 works

work_id
canonical_title
alternate_titles[]
writers[]
composers[]
traditional_flag
public_domain_status
musicbrainz_work_id
source_ids[]
confidence
notes

10.2 recordings

recording_id
work_id
recording_title
artist_credit_ids[]
recorded_date
recorded_place
first_release_date
riddim_id
version_type                 # vocal | deejay | dub | instrumental | remix | live | unknown
producer_ids[]
engineer_ids[]
studio_ids[]
genre_ids[]
isrc
musicbrainz_recording_id
riddimguide_url
source_ids[]
confidence
notes

10.3 release_tracks

release_track_id
release_id
medium_number
track_position
recording_id
printed_track_title
printed_artist_credit
duration
source_ids[]

11. Seed song/recording registry

This is a small, deliberately sourceable seed — not a complete list.

11.1 Heritage / historical seed

11.2 Sleng Teng database-derived seed

11.3 Real Rock database-derived seed

11.4 Stalag database-derived seed

Artist links (v1.0)

Since v1.0 every recording whose printed credit names an artist in the §7 registry is linked through hist_recording_artists: 23 of the 34 rows (S000006–S000010 and S000011 via the JIS artists; the Sleng Teng / Real Rock rows via A000002, A000004, A000005, A000007, A000013, A000015 and the v1.0 additions A000047–A000052). The five heritage / traditional items (S000001–S000005) and the six Riddim Guide Dub … credits on Stalag stay unlinked on purpose: “Dub Fashion” / “Dub King Jammy” / “Dub Techniques” are Riddim Guide’s version-credit convention, not artist identities, and must be resolved by a historian.

Important version rule

Entries labelled [Dub ...], Version, or similar should not automatically become separate works. They may be separate recordings linked to an earlier work/riddim.

12. Album / release-group schema

release_group_id
canonical_title
primary_artist_ids[]
primary_type                 # album | single | EP | other
secondary_types[]            # compilation | live | remix | etc.
first_release_date
genre_ids[]
musicbrainz_release_group_id
discogs_master_id
source_ids[]
confidence
notes

Separate edition table:

release_id
release_group_id
title_as_issued
release_date
country
label_ids[]
catalog_numbers[]
format
barcode
musicbrainz_release_id
discogs_release_id
source_ids[]
confidence
notes

13. Seed album/release registry

All nine hand-curated rows carry a MusicBrainz release-group MBID and a first-release date since v1.0 (the README “chronology” task for this registry is closed; §29 still lists the wider gaps).

13.1 MusicBrainz neighbourhood release groups (generated, v1.0)

Phase B (§19) imports every release group credited to a seed artist with an accepted MBID — all primary types, so the §13 headline figure stays count(distinct release_group where primary_type = 'album') and singles/EPs are counted separately. Rows live in src/data/catalogue/musicbrainz/seed-neighbourhood.json with stable ids AL100001…, are DATABASE_DERIVED citing SRC016, keep the printed artist credit, and are linked to every seeded artist in the credit through hist_release_group_artists. Release groups already in the hand registry above are excluded by MBID.

Snapshot 2026-09-22 (49 seed artists browsed): 2,874 release groups — 2,118 albums (1,353 with no secondary type; 704 compilation, 66 live, 35 remix, 7 mixtape/street, 2 DJ-mix, 1 soundtrack, 1 interview), 604 singles, 88 EPs, 64 other/broadcast/untyped. 2,679 carry a first-release date (1957 → 2026: 1950s 1 · 1960s 75 · 1970s 266 · 1980s 326 · 1990s 653 · 2000s 652 · 2010s 466 · 2020s 240); 195 are undated in MusicBrainz.

Reading these rows. A MusicBrainz release group is an album concept, not a verified Jamaican release: budget-label compilations, worldwide digital reissues and mis-attributed items are all present exactly as MusicBrainz has them. That is why every row is DATABASE_DERIVED, why compilations are published separately (album counting decision below), and why no row is counted toward any “Jamaican total” without the §22 origin and §20 deduplication work.

Release groups are credited to 70 seeded artists in total; 21 neighbourhood artists (e.g. Bob Marley, Peter Tosh, Jackie Mittoo’s groups) appear in shared credits.

Album counting decision

For the project’s “How many albums?” headline statistic:

Count unique release_group_id where primary_type = album.

Publish separate figures for:

• Studio/original albums.
• Compilation albums.
• Live albums.
• Remix/dub albums.
• Riddim compilations.
• Reissues/editions (release_id) — not included in unique-album total.

14. Label seed registry

Based on the Riddim Guide statistics snapshot.

15. Producer / studio / contributor model

Recommended entity types:

contributors
- contributor_id
- canonical_name
- type: person | group | studio | band | sound_system
- artist_id (nullable)
- aliases
- source_ids

Recommended credit relation:

recording_credits
- recording_id
- contributor_id
- role
- date_or_period
- source_id
- confidence

Suggested controlled roles:

• producer
• executive_producer
• engineer
• mixer
• arranger
• vocalist
• deejay
• toaster
• drummer
• bassist
• guitarist
• keyboardist
• percussionist
• horn_player
• backing_vocalist
• composer
• lyricist
• label_owner
• sound_system_operator
• selector

Do not encode role names only as free text.

16. Recommended relational database structure

artists
artist_aliases
artist_memberships
artist_origins

genres
genre_relations
artist_genres
recording_genres
release_group_genres

works
work_writers

recordings
recording_artists
recording_credits

riddims
riddim_aliases
riddim_relations
recording_riddims

release_groups
releases
release_tracks

labels
release_labels

studios

sources
source_assertions
research_notes
change_log

The important table is source_assertions.

A historian should be able to say:

“Field X on Entity Y has value Z because Source Q says so.”

That makes corrections auditable.

17. Provenance schema

source_id
source_type
title
institution_or_publisher
author
url
publication_date
accessed_date
archive_url
license_or_terms
citation_text
notes

And:

source_assertion_id
source_id
entity_type
entity_id
field_name
asserted_value
quotation_or_locator
confidence
reviewer
review_date
status

18. Priority source registry

SRC001 — Riddim Guide statistics

• URL: https://www.riddimguide.com/more/statistics
• Type: specialist discography database
• Use for: seed riddim/song/artist/label counts and relationships
• Snapshot observed:
  • 57,875 songs
  • 4,878 riddims
  • 5,937 artists
  • 2,570 labels
  • 45,029 songs with riddim info (78%)
• Status: DATABASE_DERIVED
• Warning: live database; counts change

SRC002 — Roots Knotty Roots / UWI Mona Library

• URL: https://www.mona.uwi.edu/library/roots-knotty-roots-discography-jamaican-music-new-resource-0
• Institution: The University of the West Indies, Mona Library
• Use for: historical Jamaican singles, 1953–1986
• Coverage statement: 62,000+ listings
• Notes: UWI page says the resource contains B-side, rhythm, musical-origin and genre documentation and is on-campus accessible
• Status: high-priority historical source

SRC003 — Jamaica Information Service: Jamaica’s Heritage in Music

• URL: https://jis.gov.jm/information/jamaicas-heritage-dance-music/jamaicas-heritage-music/
• Institution: Government of Jamaica / Jamaica Information Service
• Use for:
  • heritage taxonomy
  • Revival
  • Mento
  • Rastafarian music
  • Ska/Rocksteady context
  • historical examples
• Status: institutional secondary reference

SRC004 — Jamaica Information Service: Mento

• URL: https://jis.gov.jm/jamaica/mento/
• Use for: mento background and early-recording history
• Status: institutional secondary reference

SRC005 — Jamaica Information Service: Rock Steady

• URL: https://jis.gov.jm/jamaica/rock-steady/
• Use for:
  • c. 1966 origin
  • artist seed
  • rhythm-lineage clues
  • transition into reggae
• Status: institutional secondary reference

SRC006 — Jamaica Information Service: Lovers Rock

• URL: https://jis.gov.jm/jamaica/lovers-rock/
• Use for: London/Jamaican-diaspora branch of reggae
• Status: institutional secondary reference

SRC007 — Riddims World database

• URL: https://riddimsworld.com/database/
• Type: specialist Caribbean riddim database
• Observed total: 24,278 riddims
• Coverage: reggae, dancehall and soca, 1960s to present
• Warning: Do not import the entire number into the Jamaican-riddim total because soca and wider Caribbean entries are included
• Status: DATABASE_DERIVED

SRC008 — MusicBrainz database

• URL: https://musicbrainz.org/
• Database documentation: https://musicbrainz.org/doc/MusicBrainz_Database
• Download documentation: https://musicbrainz.org/doc/MusicBrainz_Database/Download
• Release-group model: https://musicbrainz.org/doc/Release_Group
• API documentation: https://musicbrainz.org/doc/MusicBrainz_API
• Use for:
  • artist IDs
  • recording IDs
  • works
  • release groups
  • releases
  • labels
  • relationships
• Licensing:
  • core database: CC0/public-domain dedication
  • supplementary data has separate CC BY-NC-SA terms
• High-priority machine-ingestion source

SRC009 — MusicBrainz release group: Catch a Fire

• URL: https://musicbrainz.org/release-group/63a3ac2f-5d63-3132-98d3-b00385d9a3a2

SRC010 — MusicBrainz release group: King Tubbys Meets Rockers Uptown

• URL: https://musicbrainz.org/release-group/8426fa46-01c5-3343-97f5-0551822542ed

SRC011 — MusicBrainz release group: Garvey’s Ghost

• URL: https://musicbrainz.org/release-group/cd4f4111-1e48-427b-b967-376c04bd5729

SRC012 — Riddim Guide: Sleng Teng

• URL: https://www.riddimguide.com/tunedb/riddim_Sleng%20Teng/
• Observed individual-page results: approximately 355
• Note: may differ from summary-statistics snapshot

SRC013 — Riddim Guide: Real Rock

• URL: https://www.riddimguide.com/tunedb/riddim_Real%20Rock/
• Observed individual-page results: approximately 380

SRC014 — Riddim Guide: Stalag

• URL: https://www.riddimguide.com/tunedb/riddim_Stalag/
• Observed individual-page results: 295

SRC015 — Jamaican folk-music digitization story

• URL: https://jis.gov.jm/cd-of-jamaican-folk-music-to-be-launched-today/
• Use for: archival trail for field recordings of Kumina, Tambo, Set-up, Burru, Jonkunnu, Gerreh, Rastafari and other traditions
• Important lead: locate the surviving institutional collection and catalogue

SRC016 — MusicBrainz WS/2 API: seed-neighbourhood snapshot (Phase B, v1.0)

• URL: https://musicbrainz.org/ws/2/ (release-group?artist=<mbid>&inc=artist-credits, artist/<mbid>?inc=aliases+artist-rels)
• Accessed: 2026-09-22 (recorded per run in seed-neighbourhood.json)
• Use for: release groups credited to §7 seed artists, member of band relationships, artist aliases — one relationship hop from every seed artist with an accepted MBID
• Licensing: core data, CC0. No tags or other CC BY-NC-SA supplementary data are requested.
• Status: DATABASE_DERIVED
• Command: npm run catalogue:mb-phase-b -- --neighbourhood
• Warning: live API; re-runs re-use seed ids for known MBIDs and append new rows

SRC017 — Vybz Kartel Complete Discography v1.1 (internal research deliverable)

• Path: docs/catalogue/vybz-kartel-complete-discography-v1.1.md (locked v1 + v1.1 riddim-name fill, 2026-09-21)
• Use for: per-artist census reference for A000008 — 1,714 Confirmed/Probable unique recordings, 2,121 index rows, 475 named riddims (281 still unidentified), Album & Project Index with MusicBrainz release / release-group links
• Compiled from MusicBrainz, Discogs, Riddim-ID and Riddim Guide; not a completeness claim
• Status: DATABASE_DERIVED; recording-level rows are not yet imported into hist_* (Phase C reconciliation)

19. MusicBrainz ingestion plan for the web developer

MusicBrainz should be treated as an open metadata backbone, not as the final authority on Jamaican genre classification.

Phase A — ingest core entities — implemented (scratch + extracts)

Import:

• artist
• artist aliases
• release groups
• releases
• recordings
• works
• labels
• areas
• relationships

Retain MBIDs permanently.

Status (v1.0). npm run catalogue:mb-phase-a resolves the latest MetaBrainz full export, verifies and unpacks it, COPYs 39 core + derived tables into a scratch Postgres (CATALOGUE_SCRATCH_DATABASE_URL, never the site database) and writes the Phase B candidate extracts (artists_jamaica_area, releases_jm, labels_jm, recordings_on_jm_releases, tagged_*, seed_neighbourhood) plus a citable hist_sources snapshot record (SRC-MB-<dump-id>). Seed artists carry MBIDs via artist-mbids.json (--resolve-seed-mbids, 49 accepted / 4 review). See src/scripts/catalogue/README.md.

Phase B — build Jamaican candidate set — neighbourhood import implemented; scoring pending

Candidates can be generated from combinations of:

• Jamaican artist areas.
• Jamaican release countries.
• reggae/ska/rocksteady/dub/dancehall tags.
• known Jamaican labels.
• known Jamaican producers.
• known riddim-linked artists.
• curated historian lists.

Never use one tag alone as proof.

Status (v1.0). The reviewed, small-radius part of Phase B is live: npm run catalogue:mb-phase-b -- --neighbourhood walks one relationship hop from every seed artist with an accepted MBID through the WS/2 API and writes src/data/catalogue/musicbrainz/seed-neighbourhood.json — §7.1 artists (member of band counterparts + memberships + aliases) and §13.1 release groups (every credited release group). Every row is DATABASE_DERIVED, cites SRC016, carries an MBID, keeps a stable seed id across re-runs, and is one documented hop from a hand-curated anchor; tags are never requested. Bulk scoring of the Phase A extracts (area + JM-only release + label + riddim signals combined) into a reviewed candidate list is not done yet — see §30.

Phase C — reconcile specialist sources — not started

Map:

• Riddim Guide artist -> internal artist -> MusicBrainz artist where possible.
• Riddim Guide tune -> internal recording/work.
• Riddim Guide riddim -> internal riddim.
• Riddims World riddim -> internal riddim.
• UWI/RKR single -> release + recordings + riddim.
• Vybz Kartel discography v1.1 (SRC017) rows -> internal recordings / riddims (first per-artist reconciliation).
• Historian corrections -> source_assertions.

20. Deduplication keys

Artists

Strong matches:

1. Same MusicBrainz Artist ID.
2. Same documented legal identity + alias relationship.
3. Same group identity.

Do not merge solely on name.

Works

Strong match:

• same MusicBrainz Work ID, or
• same title + writers/composers + documented composition relationship.

Recordings

Strong match:

• same MusicBrainz Recording ID, or
• same master/performance with corroborating release relationships.

A dub/version/remix should normally remain a separate recording.

Albums

Headline unique-album key:

release_group_id

Not:

title + year

Riddims

Use:

• canonical name,
• aliases,
• origin recording,
• lineage,
• producer/studio,
• historical evidence.

Do not merge two riddims solely because they sound similar.

21. Riddim lineage relationship types

Recommended controlled vocabulary:

same_as
alias_of
remake_of
derived_from
variation_of
digital_rebuild_of
version_of
disputed_relation
unknown_relation

This is critical for riddims such as older Studio One/Treasure Isle foundations that were reconstructed repeatedly.

22. Artist-origin classification

For the user’s intended final statistic, use these exact buckets:

jamaican_born
jamaican_descent_born_abroad
non_jamaican_born_jamaican_genre_artist
mixed_or_multinational_group
unknown

Rules:

jamaican_born

Documented birthplace is in Jamaica.

jamaican_descent_born_abroad

Documented birthplace is outside Jamaica and a reliable source documents Jamaican parentage/heritage.

non_jamaican_born_jamaican_genre_artist

Born outside Jamaica, no documented Jamaican-descent qualification being used, but has a substantial recorded body of work in Jamaican genres.

mixed_or_multinational_group

Use when a group’s members cannot meaningfully be represented by one birthplace/heritage category.

unknown

Use whenever evidence is insufficient.

Do not force unknown people into a demographic category just to make totals add up.

23. Recommended headline statistics

The website should compute these dynamically:

unique_works
unique_recordings
unique_release_groups
unique_releases
unique_riddims
unique_riddim_families
unique_artist_entities
unique_individual_people
unique_groups
jamaican_born_artist_count
jamaican_descent_born_abroad_count
non_jamaican_jamaican_genre_artist_count
unknown_origin_artist_count

And by:

• genre
• subgenre
• decade
• year
• label
• producer
• studio
• riddim
• country
• artist origin category

23.1 Explicit coverage statistics (v1.0)

No seeded count is published without saying how complete the rows behind it are. src/lib/catalogue/coverage.ts computes these from the seed and they are shown on /catalogue and in GET /api/catalogue/stats (coverage[]):

24. Suggested API output

Example artist object:

{
  "id": "A000001",
  "name": "Example Artist",
  "type": "person",
  "aliases": [],
  "origin": {
    "bucket": "unknown",
    "birth_country": null,
    "jamaican_heritage": null,
    "confidence": "unknown"
  },
  "genres": ["genre-reggae"],
  "external_ids": {
    "musicbrainz": null,
    "wikidata": null
  },
  "sources": [],
  "review_status": "research_required"
}

Example recording:

{
  "id": "S000012",
  "title": "Kill Them With It",
  "artist_ids": [],
  "work_id": null,
  "riddim_id": "R000003",
  "year": 1987,
  "genres": ["genre-dancehall"],
  "version_type": "unknown",
  "sources": ["SRC012"],
  "review_status": "database_derived"
}

25. Historian review workflow

Every imported record should enter one of four queues.

Queue 1 — identity

Questions:

• Is this the same person/group as an existing record?
• Is the spelling canonical?
• Is this an alias?

Queue 2 — chronology

Questions:

• Earliest release?
• Recording date?
• Reissue versus original?

Queue 3 — riddim lineage

Questions:

• Same riddim?
• Remake?
• Renamed version?
• New rhythm inspired by old rhythm?

Queue 4 — demographic/origin evidence

Questions:

• Birth country?
• Parentage/heritage documentation?
• Group composition?

26. Research backlog — highest priority

Historical archive work

• Obtain/research Roots Knotty Roots access through UWI.
• Identify whether machine-readable export/licensing is available.
• Locate the digitized Jamaica School of Music / folk-music archive referenced by JIS.
• Catalogue early mento 78s and 7-inch issues.
• Catalogue pre-ska Jamaican recordings and field recordings.

Riddim work

• Export/record all 4,878+ Riddim Guide riddim names through permitted methods.
• Reconcile against Riddims World.
• Build alias/lineage graph.
• Identify source recordings for canonical riddims.
• Separate Jamaican riddims from soca/non-Jamaican Caribbean material.

Artist work

• Import MusicBrainz candidates.
• Reconcile artist aliases.
• Split groups from people.
• Research birthplace.
• Research documented Jamaican descent for diaspora artists.
• Tag non-Jamaican practitioners separately.

Release work

• Treat MusicBrainz release groups as album concepts.
• Ingest individual releases only for discographic detail.
• Track Jamaican vs international pressings.
• Track label/cat-number history.
• Keep compilations separate from studio albums.

27. Data-quality rules for historians

1. Never overwrite a disputed value without preserving the prior assertion.
2. Store conflicting sources.
3. Prefer exact dates only when exact dates are sourced.
4. Use YYYY, YYYY-MM, or YYYY-MM-DD precision explicitly.
5. Preserve original spelling in a separate field.
6. Do not modernize titles silently.
7. Preserve label catalogue numbers.
8. Record whether a title is printed, database-normalized, or historian-normalized.
9. Record uncertainty.
10. Never fabricate missing artist demographic data.
11. Never assume two similarly named riddims are identical.
12. Never assume two same-titled songs are the same work.
13. Separate a composition from its recording.
14. Separate an album concept from its reissues.
15. Keep a change log.

28. Suggested public website display

Each artist page should show:

• Canonical name.
• Aliases.
• Birthplace/origin evidence.
• Groups.
• Genres.
• Songs/recordings.
• Albums.
• Riddims used.
• Producers.
• Labels.
• Sources.
• Unresolved historical questions.

Each riddim page should show:

• Canonical name.
• Aliases.
• Original/source rhythm.
• First known year.
• Producer.
• Studio/band.
• Parent/derived riddims.
• All linked recordings.
• Timeline.
• Source citations.
• Disputed lineage notes.

Each recording page should show:

• Title.
• Artist.
• Work/composition.
• Riddim.
• Recording date.
• Producer.
• Studio.
• Label.
• First release.
• Reissues.
• Versions.
• Sources.

Each album page should show:

• Release-group concept.
• Original issue.
• Tracklist.
• Reissues.
• Labels/catalogue numbers.
• Countries.
• Formats.
• Source IDs.

29. Known limitations of this seed

This file does not yet contain:

• all 62,000+ RKR singles;
• all 57,875 Riddim Guide songs;
• all 4,878 Riddim Guide riddims;
• all 5,937 Riddim Guide artists;
• the full MusicBrainz Jamaican candidate corpus (Phase A extracts exist in scratch: 2,819 Jamaica-area artists, 40,982 JM release events, 291,245 recordings on JM releases — none scored or imported);
• a verified birthplace/heritage census (0 of 137 artists have origin evidence);
• the full mento/folk field-recording archive;
• all Jamaican album release groups (v1.0 holds the 2,883 credited to the 49 MBID-resolved seed artists — a neighbourhood, not a census);
• recording-level MusicBrainz reconciliation (0 of 34 seed recordings carry an MBID; hist_works is empty);
• label identity resolution (0 of 10 labels carry an MBID);
• the 1,714 Vybz Kartel recordings documented in SRC017;
• all international reggae/dub/dancehall artists.

Those require bulk ingestion, archival access, licensing/terms review, and historian reconciliation.

The purpose of the seed (v0.1 → v1.0) is to make that work systematic and mergeable instead of creating several incompatible spreadsheets/lists. v1.0 locks the model and the registries so the next phases append to them rather than reshaping them.

30. Build milestones

Two version lines are deliberately kept apart: the document/seed version (front matter; this file is v1.0 and locked to manifest.json) and the catalogue coverage milestones below, which describe how much of Jamaican music history the hist_* tables actually hold.

Where the catalogue stands at Seed v1.0 (2026-09-22)

Next catalogue milestone (coverage 0.2)

• Score the Phase A extracts into a reviewed Jamaican candidate list (area + JM-only release + label + riddim-linked signals; never a single tag) and import the accepted tier as DATABASE_DERIVED artists / release groups citing SRC-MB-<dump-id>.
• Resolve label MBIDs for §14 and import labels on JM-only releases (labels_jm).
• Recording-level reconciliation for the §11 seed and the SRC017 Kartel census (Phase C first pass).
• Birthplace/heritage evidence for the first 50 hand-curated artists (queue 4).
• Riddim alias graph for the §9 riddims; verify the five JIS lineage candidates.
• Automated duplicate reports across hist_artists / hist_release_groups (same MBID, same normalised name + date).

Catalogue 1.0 target (unchanged)

• Source-backed public catalogue.
• Historian review dashboard.
• Duplicate/alias resolution workflow.
• Machine-readable JSON/CSV export.
• Coverage statistics on every published figure.

31. Core source URLs

1. Riddim Guide statistics
   https://www.riddimguide.com/more/statistics

2. Riddim Guide — Sleng Teng
   https://www.riddimguide.com/tunedb/riddim_Sleng%20Teng/

3. Riddim Guide — Real Rock
   https://www.riddimguide.com/tunedb/riddim_Real%20Rock/

4. Riddim Guide — Stalag
   https://www.riddimguide.com/tunedb/riddim_Stalag/

5. UWI Mona — Roots Knotty Roots
   https://www.mona.uwi.edu/library/roots-knotty-roots-discography-jamaican-music-new-resource-0

6. Jamaica Information Service — Heritage in Music
   https://jis.gov.jm/information/jamaicas-heritage-dance-music/jamaicas-heritage-music/

7. Jamaica Information Service — Mento
   https://jis.gov.jm/jamaica/mento/

8. Jamaica Information Service — Rock Steady
   https://jis.gov.jm/jamaica/rock-steady/

9. Jamaica Information Service — Lovers Rock
   https://jis.gov.jm/jamaica/lovers-rock/

10. Jamaica Information Service — Folk music digitization lead
    https://jis.gov.jm/cd-of-jamaican-folk-music-to-be-launched-today/

11. Riddims World database
    https://riddimsworld.com/database/

12. MusicBrainz database documentation
    https://musicbrainz.org/doc/MusicBrainz_Database

13. MusicBrainz download documentation
    https://musicbrainz.org/doc/MusicBrainz_Database/Download

14. MusicBrainz release-group model
    https://musicbrainz.org/doc/Release_Group

15. MusicBrainz API
    https://musicbrainz.org/doc/MusicBrainz_API

16. MusicBrainz — Catch a Fire release group
    https://musicbrainz.org/release-group/63a3ac2f-5d63-3132-98d3-b00385d9a3a2

17. MusicBrainz — King Tubbys Meets Rockers Uptown release group
    https://musicbrainz.org/release-group/8426fa46-01c5-3343-97f5-0551822542ed

18. MusicBrainz — Garvey’s Ghost release group
    https://musicbrainz.org/release-group/cd4f4111-1e48-427b-b967-376c04bd5729

19. MusicBrainz WS/2 API (SRC016)
    https://musicbrainz.org/ws/2/

20. Vybz Kartel Complete Discography v1.1 (SRC017)
    https://github.com/devbyesirom/jamaicansmusic/blob/main/docs/catalogue/vybz-kartel-complete-discography-v1.1.md

32. Project principle

The goal is not to force Jamaican music history into one oversimplified number.

The goal is to build a transparent catalogue where every final number can answer:

• What exactly was counted?
• What was excluded?
• Which duplicates were merged?
• Which genres overlap?
• Which riddim aliases were collapsed?
• Which reissues were grouped?
• Which artists have verified birthplace/heritage data?
• Which sources support the result?
• How complete is the database?

That is the standard required for a defensible historical census of Jamaican music.

End of Jamaican Music Master Catalogue — Seed v1.0