PDTF schema lineage and semantic crosswalk

    The PDTF schema contains the source paths used to define the Property Pack. A separate draft ontology was derived from that schema, its dictionary, glossary and supporting evidence. The SPDTF Property Pack candidate reconsiders the bounded profile by contextual meaning; it is neither a renamed schema nor a revised list of terms.

    Bottom line

    The main change is modelling authority and evidence discipline—not a move from JSON to RDF. A separate graph ontology was derived from the PDTF schema, but neither artefact has scheme authority. The Property Pack candidate asks domain-led questions afresh, keeps source facts separate from semantic proposals and assigns every candidate resource one explicit semantic home.

    Crosswalk status

    All 451 required Property Pack paths occur in the PDTF schema. Only 34 have an exact full-path entry in the schema-derived ontology provenance index. Consolidation and one-to-many modelling mean this is not a semantic coverage percentage. The maintained 451-row crosswalk from schema path to schema-derived construct or gap, and then to the Property Pack candidate, remains incomplete. It must record each retained, revised, consolidated, split or missing disposition.

    This report distinguishes the PDTF schema, the separate schema-derived ontology, and the Property Pack ontology candidate. The candidate is an accelerated SPDTF component, not an adoption, release, replacement or external-authority claim.

    Quantitative snapshot

    Not a like-for-like size comparison

    The schema-derived ontology covers evidence from the broader PDTF schema and overlays. The candidate covers only the required Property Pack profile and consolidates repeated source paths into semantic patterns. These figures describe each corpus; they do not rank quality or completeness.

    Draft derived artefact

    Ontology derived from the PDTF schema

    A separate technical corpus organised through concern modules and industry views; it is not an endorsed scheme.

    321
    OWL resources
    41
    classes
    75
    object properties
    205
    datatype properties
    48
    value schemes · 319 concepts
    402
    SHACL shapes · 66 indexed constraints
    31
    overlay profiles
    17
    diagnostic exemplars
    Public review candidate

    SPDTF · Property Pack ontology candidate

    A separate candidate namespace with explicit context ownership and source-item traceability.

    159
    OWL resources
    53
    classes
    77
    object properties
    29
    datatype properties
    14
    value schemes · 85 concepts
    45
    SHACL shapes · 100 property constraints
    451
    required source data points
    413
    source-topic concepts

    Side-by-side comparison

    Dimension Schema-derived ontology · draft Property Pack ontology · review candidate

    Starting question

    Schema-derived ontology · draft

    How can the PDTF schema, glossary, dictionary and form overlays be expressed and checked as linked data?

    Property Pack ontology · review candidate

    What semantic model is needed to represent the 451 required Property Pack data points correctly?

    Why it matters The Property Pack candidate changes the modelling premise. It is not a mechanical refresh of the schema-derived ontology.

    Evidence: PDTF schema reference · ADR-0066 · closed Property Pack scope

    Business-data scope

    Schema-derived ontology · draft

    Broad PDTF schema v3.5.0 package spanning the transaction schema and form overlays.

    Property Pack ontology · review candidate

    Closed initial scope of exactly 451 workbook rows marked Required.

    Why it matters The Property Pack candidate has source-to-candidate trace coverage for its 451-item scope, not complete coverage of the broader PDTF schema or schema-derived ontology.

    Evidence: PDTF schema reference · Property Pack source coverage

    Source authority

    Schema-derived ontology · draft

    The PDTF schema, glossary, dictionary, ODR and council evidence all shaped the separate schema-derived graph.

    Property Pack ontology · review candidate

    The 451 rows fix coverage; domain evidence establishes meaning. The PDTF schema and schema-derived ontology are attributed evidence, not silent semantic seeds.

    Why it matters Resources, relationships, identities and constraints are open to re-decision rather than inherited by default.

    Evidence: Schema-derived ontology provenance · ADR-0066 · closed Property Pack scope

    Architecture

    Schema-derived ontology · draft

    One schema-derived ontology namespace organised into seven concern modules: foundation, property, agent, transaction, claim, governance and descriptive.

    Property Pack ontology · review candidate

    A separate candidate namespace with six bounded contexts, a deliberately small Common boundary and a DBT Smart Data scheme context.

    Why it matters The schema-derived modules and Property Pack semantic homes are different organising concepts, even though both are graph models.

    Evidence: Schema-derived ontology reference · Property Pack contextual boundaries

    Context ownership

    Schema-derived ontology · draft

    Industry contexts are perspectives derived across the shared model and overlay profiles.

    Property Pack ontology · review candidate

    Every OPDA-defined resource and every source item has exactly one proposed semantic home.

    Why it matters Property Pack makes domain accountability explicit while preserving differences between contexts.

    Evidence: Schema-derived overlay profiles · ADR-0067 · context-owned candidate

    From fields to meaning

    Schema-derived ontology · draft

    The draft index is property-heavy: 41 classes, 75 object properties and 205 datatype properties.

    Property Pack ontology · review candidate

    The candidate consolidates repeated source paths into 53 classes, 77 object properties and 29 datatype properties.

    Why it matters Property Pack leans more heavily on resources and relationships, but the count change is not a quality score because the scopes differ.

    Evidence: Schema-derived ontology reference · Property Pack ontology resources

    Common meaning

    Schema-derived ontology · draft

    A shared namespace is interpreted through concern modules and industry-context views.

    Property Pack ontology · review candidate

    Only meaning evidenced as genuinely shared belongs in the Common boundary; current cross-domain mappings remain unasserted.

    Why it matters Frequent reuse no longer makes a term common automatically.

    Evidence: Property Pack contextual boundaries · ADR-0067 · context-owned candidate

    Controlled vocabularies

    Schema-derived ontology · draft

    The indexed corpus contains 48 SKOS schemes and 319 concepts across the broader schema-derived scope.

    Property Pack ontology · review candidate

    Fourteen value schemes contain 85 curated concepts; a separate source-topic layer contains 413 concepts.

    Why it matters Property Pack separates values used in data from topics used to organise source questions.

    Evidence: Schema-derived ontology reference · Property Pack controlled vocabularies

    Validation

    Schema-derived ontology · draft

    The indexed corpus contains 402 SHACL shapes, 31 form-overlay profiles and 17 diagnostic exemplars.

    Property Pack ontology · review candidate

    The candidate contains 45 target-class node shapes and 100 emitted property constraints, plus deterministic candidate checks.

    Why it matters These totals count different shape populations. Neither total establishes semantic correctness.

    Evidence: Schema-derived overlay profiles · Property Pack SHACL shapes · Property Pack validation evidence

    Traceability

    Schema-derived ontology · draft

    Term-level provenance and RML connect ontology resources back to their schema locations.

    Property Pack ontology · review candidate

    Stable Property Pack item IDs trace all 451 source points to candidate constructs, with direct and structural evidence distinguished.

    Why it matters The Property Pack candidate strengthens source-versus-proposal separation; it does not yet provide a sanctioned schema-derived-ontology-to-Property-Pack migration map.

    Evidence: Schema-to-ontology RML mapping · Property Pack source coverage

    Outputs and compatibility

    Schema-derived ontology · draft

    OWL, SKOS, SHACL, RML, overlays, exemplars and dereferenceable documentation support the established JSON-schema ecosystem.

    Property Pack ontology · review candidate

    OWL, SKOS, SHACL, glossary, dictionary, coverage and validation artefacts form the current candidate package. JSON Schema and form projections are later outputs.

    Why it matters Existing implementers should not replace current schemas with Property Pack. Compatibility work requires an explicit migration decision.

    Evidence: Schema-to-ontology RML mapping · Property Pack standards profile

    Authority and status

    Schema-derived ontology · draft

    The PDTF schema is an existing technical implementation; the separate schema-derived ontology remains a draft reference for its committed corpus.

    Property Pack ontology · review candidate

    Public review candidate: 0.1.0-draft, machine-proposed and non-normative. Human review and recorded disposition remain required.

    Why it matters Publication makes the Property Pack candidate reviewable; it does not approve it, endorse a scheme or replace the PDTF schema.

    Evidence: Schema-derived ontology reference · Property Pack candidate overview · Property Pack validation evidence

    Concrete semantic differences

    Both models are graphs

    The schema-derived ontology moved beyond the JSON tree and made explicit identity, class and relationship decisions. The Property Pack candidate changes how those decisions are sourced and owned; it does not introduce graph modelling for the first time.

    The overlap is limited

    Only 20 of 159 candidate resources share an exact local name with a schema-derived ontology resource. Only 16 also retain the same broad resource kind. Names alone do not establish equivalence.

    Exact local-name matches with the same resource kind

    This is a mechanical identifier check: each row uses the same local name in both models and is classified at the same broad level, such as a class or an object property. It identifies a useful starting point for crosswalk review; it does not establish that the definitions, identity rules, relationships or constraints mean the same thing.

    16 names to review. Compare the meaning and evidence behind each pair before treating it as continuity.

    Local name Schema-derived ontology Property Pack candidate

    Address class

    class class

    applicationDate datatype property

    datatype property datatype property

    builtForm object property

    object property object property

    hasAddress object property

    object property object property

    hasRegisteredTitle object property

    object property object property

    heatingType object property

    object property object property

    inclusionStatus object property

    object property object property

    mediaUrl datatype property

    datatype property datatype property

    ownershipType object property

    object property object property

    postcode datatype property

    datatype property datatype property

    priceQualifier object property

    object property object property

    Property class

    class class

    propertyType object property

    object property object property

    RegisteredTitle class

    class class

    Search class

    class class

    titleNumber datatype property

    datatype property datatype property

    Exact local-name matches that were reclassified

    These names also occur in both models, but their broad resource kind has changed. For example, a term may be a class in one model and a property in the other. This is a clear signal not to infer a mapping from spelling alone: the modelling role has already changed and needs an explicit decision under the Category 8 mapping method.

    4 reclassification questions. Each row needs a decision on whether the two uses are related, distinct or simply coincidental.

    Local name Schema-derived ontology Property Pack candidate

    applicationType datatype property

    datatype property object property

    councilTaxBand datatype property

    datatype property object property

    designationType datatype property

    datatype property object property

    leaseTerm object property

    object property datatype property

    A substantive identity question remains open

    The schema-derived ontology defines Property through spatial-material continuity and explicitly treats UPRN as a contingent identifier, not the identity criterion. The Property Pack ontology candidate Property candidate currently associates continuity with the same assigned UPRN. That is a semantic difference for working-group review—not a settled SPDTF decision.

    Present in the schema-derived ontology, outside the initial Property Pack scope

    Examples include Transaction, Person, Organisation, Buyer, Seller, Claim, Evidence and Trust Framework. Their absence from this candidate does not retire them; the 451-item Property Pack ontology candidate simply does not yet justify every broader schema-derived concept.

    Prominent Property Pack ontology candidate patterns

    Boundary Statement, Disclosure Statement, Fixture Schedule, Register Extract, Search Finding, Utility Service Record and Data Sharing Authorisation make Property Pack meaning more explicit within proposed semantic homes.

    What carries forward from PDTF schema

    • The PDTF schema remains useful implementation and migration evidence, while its derived ontology remains qualified semantic and provenance evidence.
    • The linked-data standards family continues: RDF, RDFS, OWL, SKOS, SHACL, Dublin Core and provenance vocabularies.
    • Existing definitions and design decisions can support a Property Pack decision when they are cited and reviewed.
    • The build, validation and documentation lessons from the schema-derived ontology are reused without silently importing its semantic structure.

    What this means in practice

    Working groups

    Review meaning, not ontology syntax

    Test definitions, identity criteria, context ownership, relationships and constraints against real domain evidence.

    Current implementers

    Keep using the current contracts

    The Property Pack ontology candidate is not a replacement schema or migration target. Current integrations remain anchored to the published PDTF schema implementation.

    Interoperability group

    Challenge the Common boundary

    Review genuinely shared meaning and propose mappings without overriding each bounded context's internal vocabulary. Use the governed mapping workflow.

    Model team

    Make migration explicit

    Record reviewed equivalence, transformation and projection decisions before generating replacement schemas or forms.

    What the comparison does not prove

    • A smaller or larger term count does not demonstrate better modelling: the business scopes and counting rules differ.
    • Lexical matches do not establish identity, equivalence or a migration mapping between the two namespaces.
    • Passing parser, SHACL and regeneration checks establishes technical integrity, not domain approval.
    • Covering all 451 required Property Pack data points does not cover every schema-derived concept, PDTF schema field or overlay.

    Evidence and method

    The figures are calculated at build time from the current generated ontology model, the committed schema-derived ontology profile and exemplar directories, and the isolated Property Pack ontology candidate projections. The lexical comparison uses exact local names only and deliberately makes no equivalence claim. Narrative conclusions are constrained by the recorded scope and architecture decisions.

    Comments

    Loading comments…