Gazetteer Place

Place record identity and identifier minting

A gazetteer place is identified by a persistent, opaque identifier that survives renaming, reclassification and geometry revision. Authoritative master-system identifiers (GNIS Feature ID, INSPIRE inspireId, national gazetteer key) take precedence; a governed HTTP URI is minted for web resolution; only where no external authority exists does the adopting Dimension mint a UUID or ULID. A date, a name or a coordinate pair is never the identifier.

1. Which authoritative master system already identifies this place, and what is its identifier value and namespace? identity
2. What persistent HTTP URI resolves to this place record, and who guarantees that it will not be reused? identity
3. Does the identifier remain unchanged when the place is renamed, reclassified or re-delimited, and what evidence supports that? constraint
4. How are two candidate records determined to denote the same place rather than two distinct places? validation

Geographic identifier, location type and indirect spatial referencing

ISO 19112 positions a gazetteer as a directory of geographic identifiers describing location instances, where a location type defines the semantics of the instances it admits. The place record must therefore declare which gazetteer register and which location type it belongs to, and must support indirect referencing — resolving a label or code to a position rather than carrying coordinates alone.

1. Which location type does this place instantiate, and what definition and territory of use does that type declare? classification
2. Which gazetteer register is authoritative for this geographic identifier, and what is that register's declared scope? authority
3. What position does the geographic identifier resolve to, and is that a point, an extent or both? spatial
4. Which alternative geographic identifiers (codes, legacy keys, statistical codes) also denote this location instance? interoperability

Gazetteer membership and register item

ISO 19112:2019 defines a gazetteer as a register of location instances of one or more location sub-types containing some information regarding position, and recognises the gazetteer as a sub-type of ISO 19135-1 Register, with Location a RegisterItem. OGC WFS-G exposes gazetteer metadata and location instances over a feature service. A place may appear in more than one gazetteer.

1. Which gazetteer or names register holds this location instance? identity
2. What is the register-item status of this location in that gazetteer? state
3. Who is the custodian or overall owner of the gazetteer and of this location instance? ownership
4. What is the territory of use and language or character-set scope of the gazetteer? constraint

Toponym form, language, script and transliteration

A name is not a string: INSPIRE models GeographicalName with language, and one or more SpellingOfName each carrying text, ISO 15924 script and an optional transliteration scheme, plus optional PronunciationOfName with IPA and a sound link. Romanized forms must be traceable to the romanization system used, which is a standing UNGEGN concern.

1. For each spelling of this name, what is the text, the script code and the transliteration or romanization scheme applied? definition
2. What language is each name asserted in, and how is an undetermined or multi-language name handled? classification
3. What Unicode normalization, diacritic and case-folding rules apply before comparison or indexing? validation
4. Is a pronunciation recorded, and in what notation or audio form? evidence

Name status, nativeness and preference

Names carry a status (official, standardised, historical, variant) and a nativeness (endonym or exonym), and exactly one preferred label is normally selected per audience. GNIS permits only one official name per feature for Federal use with variants retained; Getty TGN requires a preferred name plus vernacular and historical variants; GeoNames types names as official, short, colloquial or historical.

1. Which name is official, under whose authority, and for which territory and usage domain? authority
2. Is each name an endonym or an exonym, and against which language community is that judged? classification
3. How is the single display label chosen when several names compete, and does the choice vary by audience locale? decision
4. Which variant, historical or deprecated names must be retained for search recall even though they are not displayed? retention

Historical, deprecated and sensitive names

TGN retains historical, misspelled, pejorative and avoid-use names for retrieval of legacy materials and flags them so they are not used for new indexing. GNIS marks obsolete names and vanished features as historical without implying age. GeoNames records historical and colloquial alternate-name types. These forms remain in the name set rather than being deleted.

1. Which names are historical, local-use, misspellings or otherwise non-current, and for what period were they used? temporal
2. Is any name flagged pejorative, deprecated or avoid-use, and may it still be used for retrieval of historical sources? constraint
3. Are indigenous, minority or local-use names recorded, and under what community or authority warrant? authority

Feature type, local type and classification change

Every governed place carries at least one type drawn from a resolvable vocabulary. INSPIRE separates a coarse NamedPlaceTypeValue from a free localType; LPF requires broad fclasses (A, H, L, P, R, S, T) plus optional vocabulary-typed types with source labels; GeoNames uses featureClass plus featureCode; Getty TGN requires a preferred place type. Types are temporally scoped because a village can become a suburb.

1. Which controlled vocabulary and version supplies the place type, and what is the concept URI? classification
2. What local or source-supplied type label was recorded, and how does it map to the governed concept? interoperability
3. Can this place hold more than one type simultaneously, and which one is preferred? constraint
4. When the type changes, is that a real-world change or a correction, and how is each recorded? lifecycle

Administrative versus physical facet

TGN distinguishes Administrative records (inhabited places, nations, empires) from Physical records (mountains, rivers, oceans) and discourages a Both flag. Physical features that cross political boundaries are placed under the next containing hierarchical level. GNIS after 2021 focuses on natural features, unincorporated populated places and selected hydrographic and reservoir classes, and removed most cultural classes.

1. Is this place classified as an administrative entity, a physical feature, or another facet used by the gazetteer? classification
2. Does a physical feature cross administrative boundaries, and under which containing parent is it placed? relationship
3. If the gazetteer excludes cultural or constructed classes, is this entity in scope or must it be referenced in a sibling model? decision

Representative position and multiple geometries

A place may carry a representative point, a bounding box, one or more detailed geometries, or no geometry at all. GeoSPARQL distinguishes hasGeometry from hasDefaultGeometry and provides derived centroid and area; LPF permits several temporally scoped geometries with per-geometry certainty and citations, or a null geometry; Getty TGN records point and bounding-box coordinates.

1. Which geometry is the default for this place, and on what basis was it selected? decision
2. What role does each geometry play — representative point, bounding envelope, delimited extent or approximate footprint? composition
3. If the place has no geometry, what is the reason and what indirect position can still be resolved? exception
4. How certain is each geometry, and which citation supports it? quality

CRS declaration and coordinate encoding constraints

Coordinates are meaningless without a declared spatial reference. RFC 7946 fixes GeoJSON to WGS 84 with longitude before latitude; OGC API - Features defaults to OGC:CRS84 and requires explicit CRS negotiation for alternatives; GeoSPARQL carries CRS identity inside geometry literals; the SDW best practices require publishers to state how coordinate values are encoded and to choose a CRS suited to users.

1. In which coordinate reference system are the stored coordinates expressed, and where is that declaration carried? interoperability
2. What axis order, units and decimal precision are used, and is that consistent across every encoding offered? constraint
3. If coordinates were transformed from a source CRS, which transformation was applied and what error does it introduce? provenance
4. How are geometries crossing the antimeridian or the poles represented? exception

Positional accuracy, resolution and viewing-scale bounds

The SDW best practices require describing positional accuracy and providing geometries at an appropriate level of accuracy, precision and size; INSPIRE NamedPlace carries mostDetailedViewingScale and leastDetailedViewingScale to bound legitimate display. Precision of coordinates must never be read as accuracy, and generalized coordinates must be labelled as such.

1. What is the stated positional accuracy of the recorded position, and how was it determined? measurement
2. Between which most-detailed and least-detailed viewing scales is this geometry fit to be shown? quality
3. Does the number of decimal places recorded overstate the real accuracy, and has spurious precision been removed? validation
4. Has the position been deliberately generalized, and by what method and magnitude? provenance

Existence period and temporal uncertainty

Places begin, change and cease. LPF expresses temporal scope as timespans with start and end, each expressible as 'in', 'earliest' or 'latest', plus named periods, duration, label and certainty; GNIS marks features as historical when they no longer exist. The model must accept open, approximate and disjoint existence intervals without forcing a false precise date.

1. Over what interval did this place exist, and are the start and end known, approximate or open? temporal
2. Which named historical period vocabulary terms scope this place, and from which period authority? temporal
3. Is the place currently extant, and if not, what event ended its existence? state
4. Which calendar and time zone apply to the recorded dates, and how are pre-modern or proleptic dates handled? constraint

Name validity periods and renaming events

Names have their own validity independent of the place. Getty TGN records display dates with start and end dates on names; GeoNames types historicalName separately; GNIS records a BGN decision date. A renaming is an event with a decision, an effective date and a superseded prior official name that must remain searchable.

1. Over which interval was each name in use or official, and how precisely is that known? temporal
2. What renaming events affected this place, and what decision authorised each? event
3. How does the decision date differ from the effective date and from the date the record was updated? temporal
4. May two official names be valid simultaneously, and if so under which authorities or languages? constraint

Record version lifespan and change log

INSPIRE gives each spatial object version a beginLifespanVersion and endLifespanVersion; Who's On First carries a last-modified epoch and a geometry hash, and marks human-curated fields as controlled so automated processes must not overwrite them. Record time must be recorded separately from real-world time, both as RFC 3339 values.

1. When did this record version become current and when was it superseded? lifecycle
2. What are the event time, the observation time and the ingestion time for each asserted change? temporal
3. Which fields are human-curated and protected from automated overwrite, and who set that protection? authority
4. What changed between two versions, and is the change a correction, a real-world change or a re-derivation? provenance

Supersession, merge, split and deprecation

Places are discovered to be duplicates, are divided, are absorbed or are found never to have existed. Who's On First records supersedes and superseded_by as arrays, allowing one-to-many and many-to-one succession; Getty TGN defines rules for merging, moving and deleting records. Identifiers must not be silently reused or hard-deleted, because external systems already cite them.

1. Which records does this record supersede, and which records supersede it? relationship
2. Is the succession a merge of duplicates, a split, an absorption, a re-scoping or a retraction of an erroneous record? classification
3. What must a resolver return when a client requests a superseded identifier? interoperability
4. How are downstream consumers and concordance partners notified of a supersession? process

Naming authority and the bounded scope of officiality

Officiality is always scoped. GNIS records names that are official for Federal use, and the BGN cannot compel non-Federal parties to adopt them; UNGEGN standardises names nationally and promotes them internationally. The record must state which authority made a name official, for which territory, for which usage domain, and from when.

1. Which body has the authority to decide the official name of this place, and under what legal or institutional mandate? authority
2. On whom is the official name binding, and where is it merely recommended? constraint
3. Do several authorities claim competence over this place, and how is precedence determined? exception
4. Is naming competence delegated to state, regional, municipal or Indigenous naming bodies for this feature class? ownership

Name decision process, dossier and effective date

Naming decisions follow a governed process: a proposal with documentation, review against published policies (including a presumption against changing long-established names, commemorative-name restrictions and a prohibition on derogatory names), consultation with state, local and Tribal bodies, a docketed decision, and publication in the authoritative gazetteer.

1. Who proposed the name, what documentation accompanied the proposal, and when was it lodged? process
2. Which naming policies were applied to the proposal, and did any of them bar it? requirement
3. Which parties were consulted, and how were their responses weighed in the decision? process
4. What is the decision date, the effective date and the date the decision was reflected in the gazetteer? temporal
5. Can the decision be reconsidered or appealed, and what triggers a review? exception

Contested names, disputed places and neutral presentation

Some places carry competing names or competing sovereignty claims. UN M49 accompanies its groupings with an explicit disclaimer that designations imply no opinion on legal status or the delimitation of frontiers; Who's On First encodes parentage that is unknown, complicated or contested as distinct sentinel states rather than forcing a false parent. The model must represent dispute as data instead of silently choosing a side.

1. Which parties assert competing names or claims over this place, and what does each assert? relationship
2. Which name is presented to which audience, and what disclaimer accompanies the presentation? access
3. Is the place's parentage or sovereignty recorded as unknown, complicated, contested or multiple rather than resolved? state
4. Who adjudicates a naming dispute inside the adopting Dimension, and what is the escalation path? decision

Custodian, administrator, licence and update commitment

ISO 19112 attributes a gazetteer to a custodian and a location instance to an administrator, and INSPIRE and Getty both require identifiable stewardship. The record must state who maintains it, under what terms it may be redistributed, how often it is refreshed and how errors are reported.

1. Who is the custodian of the gazetteer and the administrator responsible for this location instance? ownership
2. Under what licence and attribution requirement may this record be redistributed, and do any source licences restrict downstream use? access
3. How frequently is the record reviewed or refreshed, and what is the maximum acceptable staleness? quality
4. How does a consumer report a suspected error, and what is the commitment to respond? process

Source attestation, citation and editorial lineage

INSPIRE carries sourceOfName on every geographical name; LPF attaches citations to names, types, geometries and relations individually and requires at least one citation or a record-level temporal scope per name; Getty TGN requires sources and contributors and retains merge and revision history; Who's On First marks human-curated fields as controlled. Assertion-level attestation, not record-level, is the governing pattern.

1. Which source attests each individual name, type, geometry and relation, rather than the record as a whole? provenance
2. Is the evidence a documentary attestation, a field observation, a derivation from another dataset or an editorial inference? evidence
3. Which contributor supplied the assertion, and who exercised final editorial control over it? ownership
4. When two sources conflict on a name, date or position, which prevails and on what recorded basis? decision
5. Which assertions were produced by automated extraction or model inference, and are they distinguishable from human-verified ones? quality

Partitive and administrative hierarchy

Places sit in hierarchies: GeoNames gives parentFeature and parentADM1 to parentADM4; Who's On First gives parent_id, a hierarchy of ancestor sets and a flattened belongsto; Getty TGN distinguishes preferred from non-preferred parents; schema.org offers containedInPlace and containsPlace. Multiple concurrent hierarchies (administrative, statistical, ecclesiastical, historical) must coexist without one being privileged silently.

1. Which parent is preferred for this place, in which hierarchy, and over which interval? composition
2. Which additional hierarchies does this place belong to, and are their ancestor chains stored explicitly? relationship
3. How is an unknown, complicated, contested or multiple parentage recorded without asserting a false parent? exception
4. What prevents cycles, orphans and type-inconsistent parenthood in the hierarchy? validation

Topological and associative relations

Beyond containment, places touch, overlap, adjoin, succeed and are functionally associated. GeoSPARQL provides Simple Features, Egenhofer and RCC8 relation families; schema.org exposes DE-9IM-derived geo relations; Who's On First records breaches where same-placetype records intersect; LPF relations carry a vocabulary relationType with temporal scope, certainty and citations.

1. Which topological relation vocabulary is used, and is the relation asserted or computed from geometry? relationship
2. Which non-spatial associative relations apply, such as capital-of, successor-of or seat-of? relationship
3. Do same-type places overlap where they should partition space, and how is that anomaly recorded? quality
4. Where no geometry exists, how is relative position expressed and against which reference place? spatial

External identifier concordance and match confidence

LPF distinguishes exactMatch, closeMatch, primaryTopicOf and seeAlso; Who's On First keeps a concordances hash of identifiers from other sources; schema.org uses sameAs. Asserting exact identity where only a close match is warranted is a common and damaging error, so match strength, matcher and evidence must be recorded and revisable.

1. For each external identifier, is the link an exact match, a close match, a topic-of link or merely a see-also? interoperability
2. What evidence supported the match, and who or what performed it? evidence
3. How is a concordance revalidated when the external record is superseded, deprecated or redefined? lifecycle
4. What happens when two external authorities that are both linked as exact matches disagree with each other? exception

Exchange encodings, profile conformance and query-service contract

The same place record projects into GML per the INSPIRE GN application schema, GeoJSON per RFC 7946, Linked Places Format as JSON-LD that is simultaneously valid GeoJSON, and RDF per GeoSPARQL — and is queried through interfaces such as OGC API - Features with /collections and /items paths, bbox, datetime and limit parameters and required link relations. Each projection loses or constrains something, and the losses must be declared.

1. Which encodings and profiles are offered for this record, and which is normative when they disagree? interoperability
2. What information is lost or flattened in each projection, such as temporally scoped names in plain GeoJSON? constraint
3. Which query parameters must the interface support, and how are temporal filters expressed? process
4. At which stable path or URI is an individual place item addressed, and how does that relate to the place URI? identity
5. How are model elements with no native slot in an encoding carried, and are those extensions documented? interoperability

Sensitivity classification, generalization and disclosure

GBIF defines graded sensitivity categories with matching coordinate generalization (region only, 0.1, 0.01 and 0.001 degrees), prefers grid-based generalization over randomization, requires that withholding and generalization be disclosed in dedicated fields, places the sensitivity decision with the data provider, and requires a recorded date for reviewing the sensitivity status. The same pattern governs culturally restricted, security-sensitive and privacy-sensitive places.

1. What sensitivity category applies to this place, on what grounds, and who decided it? security
2. What generalization or withholding was applied, at what magnitude, and where is the precise value retained? privacy
3. How is the fact of generalization or withholding disclosed to consumers without revealing the sensitive detail? access
4. Who may receive the unredacted record, under what agreement, and how is that release logged? access
5. When is the sensitivity status next reviewed, and what would justify downgrading or removing it? lifecycle

Validation, conformance testing and fitness for use

GeoSPARQL 1.1 ships SHACL shapes for RDF validation; INSPIRE publishes validation resources for the GN schema; LPF v1.3 makes fclasses required because it improves reconciliation; the SDW best practices call for exposing validation schemas online. Conformance to a schema is necessary but not sufficient — fitness for a stated purpose must be measured and published separately.

1. Which schemas, shapes or rule sets is this record validated against, and at which version? validation
2. Which elements are mandatory for a record to be publishable, and what happens to records that fail? requirement
3. Which semantic constraints beyond structural validity are enforced, such as type-consistent parenthood or geometry inside the parent extent? constraint
4. What quality measures are published for the dataset, and against which intended uses? measurement
5. Is a conformance claim to an external standard supported by executed test evidence, or is it only an alignment? evidence

Retention, tombstoning and lawful erasure

Gazetteer identifiers are cited by external systems, so removal is normally deprecation with a tombstone plus a superseded-by pointer, not deletion; Getty TGN treats deletion as a governed editorial action with retained history, GNIS retains historical features, and persistent-URI practice forbids identifier reuse. Genuine erasure obligations (for example personal data embedded in a note) must be satisfied without destroying the identifier or the succession chain.

1. How long are current records, superseded versions, change logs and evidence dossiers retained, and under what obligation? retention
2. What does the register return for a deprecated or erroneous record, and what minimum information does the tombstone carry? lifecycle
3. When a lawful erasure request affects part of a record, which elements are removed and which are preserved? privacy
4. Under what circumstances, if any, may a retired identifier be reused, and how is that prevented by default? constraint