Editorial standard

How the dataset
decides

Principles for the public dataset: what gets included, how sources are evaluated, how relationships are described, and where the lineage view stays intentionally conservative. The narrow question throughout — which public people, places, works and institutions help explain how rope knowledge moved between eras and regions?

01What the lineage view shows

This is a public research view of shibari / kinbaku people, schools, studios, rope bars, venues, festivals, publications, platforms, material infrastructure and historical context. It is not a complete directory, a ranking, a certification registry, or a claim that every important person is visible.

Entries are added when they help explain transmission, scene infrastructure or historical context:

  • people with public rope identities — teachers, performers, organizers, authors, photographers, riggers, models, rope makers or lineage carriers — when that identity also has a graph-safe relationship or a clearly structural / historical role
  • recurring studios, dojos, festivals, conventions, communities and education platforms
  • publications, films, production labels and material projects when they shaped rope culture or connect multiple existing entries
  • historical or regional context when it explains why clusters moved or changed

The view is intentionally selective. A person or event may be important locally and still stay out of the visible lineage if the public sources are too thin, if inclusion would expose private information, or if the public evidence does not yet explain a wider historical or structural role. A source-rich current-person profile is not enough by itself; without a source-backed relationship, infrastructure role or historical bridge, it stays source-only / watch rather than becoming an isolated visible node.

02Evidence levels

The project separates evidence into three practical levels. What a source can support depends on which level it sits at.

Strong

Official bios, studio pages, event rosters, publisher pages, interviews, archived festival pages, books, public talks, or direct pages maintained by the person or organization. Supports entries, biography facts, source links and visible relationships.

Medium

Public social posts, directory rows, event mirrors, podcast pages, third-party writeups, or archived traces. Supports bounded public claims when its directness and limits are explicit.

Weak

Unsourced lists, reposts, dead pages without archive context, search snippets, anonymous claims, private screenshots, memory or rumor. Used as a research lead, not as a public claim.

If two sources conflict, the project prefers the more direct source: the person's own current public statement, then official studio / event records, then interviews and specialist archives, then broader secondary sources. Older sources are kept only when they explain historical context and do not conflict with a later correction.

03Source handling

Preferred sources are official pages, event rosters, studio biographies, interviews, archived festival pages, publisher pages, public talks and stable community documentation. Directory rows and social posts normally support bounded claims. A stable official social profile plus multiple distinct dated programme posts may establish a durable entity's public identity and recurrence; one isolated post or directory row does not establish durability.

A publicly visible platform-native coauthor or Collab label attributes that publication to every named coauthor for its explicit caption, graphics, date and coauthor label. The source remains affiliated rather than independent. Coauthorship supports joint publication and association with that specific activity; it does not by itself establish ownership, operation, hosting, teaching, venue identity, a durable partnership or a map relationship. Tags, mentions, comments, location labels, shared rosters, mirrors and reposts are not coauthorship.

Each visible entry should have at least one public link, unless the entry is broad historical context supported by the public source list. Repeated sources are kept in a shared registry so readers can check the evidence without the same link being repeated in every paragraph.

Source links support verification and attribution. They do not mean the linked source endorses Shibari Atlas, and they do not import the linked text into the project's license. For visible relationships, source records must support the relationship itself — a source that only proves two entries exist, or only places them in the same broad scene, is not enough.

04Documented corpus profile

These counts describe the public material currently documented by Shibari Atlas. They do not measure the size, importance or geographic distribution of the underlying rope scene.

The profile is a release snapshot, not a score. A high count can reflect accessible archives, language access or concentrated research effort; a low count can reflect missing metadata or documentation gaps. Unknown values stay visible instead of being inferred.

Source records
6,175
Cited entries
1,632
Source-to-entry citations
8,180

Source type

All public source records, grouped by their declared format.

6,175 known · 0 unknown
Source type, source records. Metadata coverage 100.0%.
CategoryCountShare
Archive 71 1.1%
Article 539 8.7%
Book 56 0.9%
Catalog 50 0.8%
Database 124 2.0%
Event 1,293 20.9%
Podcast 210 3.4%
Social 1,675 27.1%
Video 34 0.6%
Website 2,123 34.4%

Source language

Declared page metadata and unambiguous writing-system cues. Platform interface language and domain are not treated as content language.

4,194 known · 1,981 unknown
Source language, source records. Metadata coverage 67.9%.
CategoryCountShare
English (EN) 2,879 46.6%
Japanese (JA) 222 3.6%
German (DE) 215 3.5%
Spanish (ES) 207 3.4%
Russian (RU) 135 2.2%
French (FR) 115 1.9%
Italian (IT) 68 1.1%
Ukrainian (UK) 63 1.0%
Czech (CS) 49 0.8%
Portuguese (PT) 41 0.7%
Dutch (NL) 28 0.5%
Chinese (ZH) 28 0.5%
Danish (DA) 26 0.4%
Polish (PL) 16 0.3%
Hungarian (HU) 14 0.2%
Finnish (FI) 13 0.2%
Korean (KO) 13 0.2%
Hebrew (HE) 8 0.1%
Norwegian (NO) 8 0.1%
Serbian (SR) 8 0.1%
Greek (EL) 7 0.1%
Estonian (ET) 7 0.1%
Slovak (SK) 5 <0.1%
Catalan (CA) 3 <0.1%
Romanian (RO) 3 <0.1%
Thai (TH) 3 <0.1%
Basque (EU) 2 <0.1%
Croatian (HR) 2 <0.1%
Swedish (SV) 2 <0.1%
Turkish (TR) 2 <0.1%
Bulgarian (BG) 1 <0.1%
Slovenian (SL) 1 <0.1%
Unknown 1,981 32.1%

Cited-entry era

Source-to-entry citations grouped by the cited entry's declared start era.

8,180 known · 0 unknown
Cited-entry era, source-to-entry citations. Metadata coverage 100.0%.
CategoryCountShare
1800s 18 0.2%
1900s 13 0.2%
1950s 118 1.4%
1970s 65 0.8%
1980s 124 1.5%
1990s 218 2.7%
2000s 586 7.2%
2010s 2,763 33.8%
2020s 4,275 52.3%

Cited-entry country

Source-to-entry citations grouped by the cited entry's explicit nationality code. Scene locations are not substituted.

7,256 known · 924 unknown
Cited-entry country, source-to-entry citations. Metadata coverage 88.7%.
CategoryCountShare
Japan (JP) 1,257 15.4%
United States (US) 1,037 12.7%
Germany (DE) 641 7.8%
United Kingdom (GB) 318 3.9%
Russia (RU) 313 3.8%
France (FR) 302 3.7%
Italy (IT) 297 3.6%
Australia (AU) 281 3.4%
Spain (ES) 270 3.3%
Switzerland (CH) 181 2.2%
Canada (CA) 161 2.0%
Czechia (CZ) 150 1.8%
Netherlands (NL) 139 1.7%
Ukraine (UA) 118 1.4%
Thailand (TH) 114 1.4%
Belgium (BE) 95 1.2%
Denmark (DK) 91 1.1%
Mexico (MX) 77 0.9%
Colombia (CO) 76 0.9%
Poland (PL) 72 0.9%
Sweden (SE) 67 0.8%
Chile (CL) 65 0.8%
Finland (FI) 61 0.7%
Brazil (BR) 59 0.7%
Hong Kong SAR China (HK) 59 0.7%
Portugal (PT) 59 0.7%
China (CN) 52 0.6%
Austria (AT) 51 0.6%
Argentina (AR) 49 0.6%
Serbia (RS) 46 0.6%
South Africa (ZA) 45 0.6%
Israel (IL) 41 0.5%
Singapore (SG) 38 0.5%
Belarus (BY) 35 0.4%
Greece (GR) 35 0.4%
Costa Rica (CR) 34 0.4%
South Korea (KR) 32 0.4%
Philippines (PH) 32 0.4%
Taiwan (TW) 30 0.4%
New Zealand (NZ) 29 0.4%
Malaysia (MY) 28 0.3%
Norway (NO) 27 0.3%
Estonia (EE) 24 0.3%
Hungary (HU) 24 0.3%
Croatia (HR) 21 0.3%
India (IN) 19 0.2%
Uruguay (UY) 19 0.2%
Venezuela (VE) 19 0.2%
Romania (RO) 18 0.2%
Bulgaria (BG) 16 0.2%
Ireland (IE) 15 0.2%
Kazakhstan (KZ) 15 0.2%
Peru (PE) 12 0.1%
Lithuania (LT) 11 0.1%
Luxembourg (LU) 10 0.1%
Latvia (LV) 10 0.1%
Slovakia (SK) 10 0.1%
Georgia (GE) 7 <0.1%
Montenegro (ME) 6 <0.1%
Panama (PA) 6 <0.1%
Armenia (AM) 5 <0.1%
Ecuador (EC) 5 <0.1%
Réunion (RE) 4 <0.1%
Türkiye (TR) 4 <0.1%
Iceland (IS) 3 <0.1%
Martinique (MQ) 3 <0.1%
Cuba (CU) 2 <0.1%
Puerto Rico (PR) 2 <0.1%
Indonesia (ID) 1 <0.1%
Jamaica (JM) 1 <0.1%
Unknown 924 11.3%

Type and language use source records as their denominator. Era and country use unique source-to-entry citations, so one source cited by two entries contributes twice. Duplicate citations on one entry contribute once. Only citations attached to people and organizations participate; citations attached to individual relationships, works or programme sections remain separate.

Open the public Sources catalogue to inspect individual public references. The catalogue keeps its own pagination and search; this profile publishes aggregates only.

05From claim to visible map

Not every true or interesting fact becomes a visible relationship. Every claim lands at the strictest level its public sources support — most never reach the map:

  1. Not used An attractive inference that is misleading, unsafe, duplicate or unsupported.
  2. Research lead Worth revisiting later, but not presented publicly yet.
  3. Biography context Useful for understanding an entry but too narrow, private, ambiguous or weak to draw.
  4. Visible relationship A source directly supports a specific connection that explains transmission, collaboration, publication, teaching or lineage.
  5. Visible entry Explains several relationships or a structurally important bridge.

The project avoids hub inflation. A magazine, festival, podcast, adult-media studio, social platform or directory is not connected to everyone it mentions. Relationships are shown only when meaningful for the lineage view — not merely because two names appear on the same page.

The project also avoids person-node inflation. Public social profiles, link hubs, interviews, directories and photo credits can be valuable source context, but if they do not support a graph-safe relationship or structural role, they remain source-only / watch. The map should not contain weak standalone current-person nodes simply because those people have real public rope identities.

Non-person entities are different: a durable studio, community, recurring event or other structural project may remain visible without a relationship when it passes its own category and evidence gates. Graph degree is not an entity admission requirement, and the project does not invent an edge merely to avoid an isolated entity.

06Era placement

Each visible entry is assigned to an era as a start-era label — the earliest source-backed public period in which that subject becomes relevant to rope history in the form shown on the map.

  • A person is placed by the first documented public rope role — publication, performance, teaching, organizing, named studio role, or explicit entry into the public scene.
  • A studio, bar, venue, festival, platform or publication is placed by its first documented opening, founding, edition, issue or release.
  • The project does not place entries by birth year, by current fame, or by whatever decade they are most active in today.
  • Later activity does not move an older entry into a newer era. If a senior practitioner later studies with a younger teacher, that is recorded as later cross-era study; their era still reflects their own public starting period.

When one continuous node would collapse distinct historical phases into a misleading single band, the project splits that subject into separate era entries instead of dragging the whole history into the present.

07Relationship categories

The view distinguishes relationship strength with a small set of public categories. Each has a categorically different line so it survives even when the map repaints edges by school colour.

Deshi
Explicit formal apprenticeship, named successor, or sustained primary-teacher lineage. The only solid line.
Certification
Instructor license, school authorization or formal rank that is public — but not framed as apprenticeship. Shown as a double rail.
Influence
Workshop study, private lessons, named stylistic influence or non-formal learning. The lightest tie.
Founded / operates
Founded, owns, operates, hosts or anchors a structured school, studio, platform, event or media line.
Appeared / credited
Verified presenter, workshop, festival, stage, magazine, film, production or other credited appearance.
Collaboration
Co-created, co-organized, partner project, staff/editor role or durable rope partnership framed as real working ties.

Workshop attendance alone is not treated as formal apprenticeship. Event co-presence never creates a person-to-person collaboration edge by itself. To avoid hub inflation, broad teacher lists fan out only to teachers who already have a node or are independently graph-worthy.

How Trace Connection chooses a path

Transmission uses only person-to-person apprenticeship, certification and explicit study / style / mixed influence. All documented connections can also use collaborations, studios, events, media and context. The route is calculated from the complete public graph, so saved visual filters never change a shared trace URL.

The algorithm prefers direct transmission semantics and adds a kind-and-degree penalty to intermediate entity hubs such as festivals or broad context entries. Equal routes prefer fewer steps and then immutable relationship IDs. Source counts, confidence and editorial assessments are not route weights. Every step keeps its recorded direction, traversal direction, public role / year / note and evidence; absent facts are shown as not recorded rather than inferred.

08Living people & privacy

For living or contemporary people, the project uses public scene names by default. Birth years, legal names, private residences and adult-media aliases are not added unless the person uses them in a stable public professional context and the information is necessary for the public presentation.

Missing birth years are expected. They are not a data-quality failure. Locations are public scene anchors — a studio city, festival city, teaching base or country-level scene context — not private addresses.

If a living person asks for a correction, removal, identity split, privacy downgrade or safer wording, that request is handled as a high-priority signal. Public safety and consent matter more than preserving a visually dense display.

09Queer, gender & sexuality context

Queer, trans, nonbinary, women / femininities or marginalized-community context may be included when it is part of the public rope-facing identity, event scope, community mission, pedagogy or historical role being documented.

For people, use only self-description, a direct public source maintained by the person or their project, or an official event / studio / publisher bio. Keep the wording close to the source. Do not broaden a specific public statement into a larger identity label, and do not infer identity from participation in a queer event, partnerships, appearance or scene reputation.

The graph keeps no dedicated identity field or "representatives" taxonomy. Identity context belongs in biography / source context, and should never be used to rank, tokenize, out, deanonymize or expose sensitive identity beyond the source's own public framing.

10Self-reported information

Direct corrections or additions from the person an entry describes may be used when they clearly allow publication — by email, direct message, public comment or another first-person channel. Self-reported information is treated as first-party correction material, not as a public source; private emails, messages and screenshots are not published in the source list.

Allowed self-reported updates include preferred name, spelling, pronouns, role, location scope, public links or removal requests; corrections to biography, teacher framing, partner context, studio affiliation or event framing; identity splits where two public names were incorrectly merged; and privacy downgrades. For strong lineage, rank or controversial claims, a public source is preferred.

11Adult-media & sensitive sources

Adult-media sources are handled narrowly. They are included only when they support a rope-specific public role: rigger, bondage director, rope consultant, production rope educator, selected rope-video work, or a production project that materially shaped rope visual culture.

The project does not import filmographies, one-off performer credits, scraped adult databases, ambiguous aliases or links that expose private identity beyond a public rope role. Sensitive sources should never be used to out, deanonymize, shame, rank or sensationalize a person. If a fact is not needed to explain rope transmission or infrastructure, it stays out.

12Licensing & attribution

The public dataset is split by content type. Its structured facts can be released under CC0 where legally possible: public entry identifiers, names, aliases, dates, scene locations, entry categories and relationship categories. Any public export must exclude biographies and other expressive prose, contact and reference URLs, canonical source identifiers, source-registry metadata and internal editorial fields. Where a fact has public evidence, an export may carry only an opaque citation key that can be resolved through this site's public citation layer.

Third-party source material remains under its original license. Biographies and notes should be independently written factual summaries — the project does not intentionally import verbatim third-party prose or close translations. If copied prose is found, it is rewritten, removed, or attributed and licensed according to the source.

13Editorial checks

Each update is checked against the same editorial principles:

  • public links or source references for updated entries and relationships
  • conservative relationship categories with short explanatory notes
  • weak leads held back until better public evidence appears
  • source-rich but isolated current-person leads kept source-only instead of becoming weak standalone nodes
  • tempting but unsupported inferences left out
  • consistency checks for relationship categories, dates, locations and public links

Known uncertainty is preferable to false precision. The project should say less when the sources say less.

14Corrections

For an ordinary factual correction, use Suggest correction on the relevant entry or event page. It opens the contribution guide, which explains what information is useful and where to send it. Where possible include a stable public source and a specific claim — what is wrong, what should change, and whether the change affects dates, location, school / lineage or relationship type.

If the correction comes directly from the person an entry describes, a public source is preferred but not always required; the person should state explicitly what may be published. Preferred outcomes are precise: fix a bio sentence, downgrade apprenticeship to influence, remove an unsupported relationship, add a source link, split two merged identities, or hide a private detail.

A correction, removal, access or privacy request about a record describing you can be sent through the contribution contact. Mark that the request concerns you and state what may be published. The message is reviewed manually; it is not published as a source and the project does not promise a fixed outcome or response period.

Help build the atlas

Corrections, missing sources and requests about a record describing you use the contribution guide.

How to contribute