Benefits, limits and reasonable objections

Start with the misunderstanding you want to prevent

A connected model is not valuable simply because it contains many connections. It is valuable when people can answer a practical question with less guessing: which dwelling does this account concern, what does this amount represent, or which version supports this statement? The improvement is a clearer agreement about those questions, not a promise that every process will become cheaper or every piece of information more accurate.

Consider two fictional floor-area accounts for the same dwelling. Account A gives 68 m², excluding an enclosed 4 m² area. Account B gives 72 m², including it. These premises explain the difference without choosing a professional measuring standard. Now consider what happens when the accounts are exchanged.

The second exchange supports a better explanation because it preserves subject, unit, basis and source—not because it uses a particular file format. The measurements chapter develops this example, including a harder case where the missing evidence prevents the same conclusion.

Five benefits, with their conditions attached

Fewer repeated reinterpretations

If sender and recipient use the same understood distinction, a person should not need to explain it afresh at every handoff. The condition is that the distinction survives the exchange and the recipient uses it correctly. If “measured area” is flattened back into one unexplained field, a carefully defined model upstream will not rescue the lost meaning downstream.

Evidence that is easier to inspect

A statement linked to its source and version gives a reviewer somewhere useful to look. That requires maintained references, relevant passages and an honest account of what the source supports. A link to an entire document, or to an inaccessible version without an interpretable description, may leave the important question unanswered. Provenance makes investigation possible; it is not a certificate of truth.

Agreements that practitioners can challenge

A readable definition, example or relationship lets somebody say precisely what is wrong. That benefit depends on reviewable wording and a reasoned human response. If an essential exclusion exists only in code, a domain expert may never see it. If a polished diagram conceals an unresolved assumption, visual clarity can create false confidence rather than understanding.

Deliberate reuse

A well-defined concept can be reused where its meaning genuinely fits. Reuse requires comparing scope, identity, purpose and relevant constraints—not just matching labels. Reusing a term that almost fits can hide a consequential difference. Sometimes a qualified correspondence between distinct meanings is the useful result; merging them is not always an improvement.

A clearer account of change

If a definition or source changes, explicit connections help identify which interpretations need review and which can stay the same. This requires version information and an explanation of the change. Overwriting the previous definition removes the very evidence needed to explain why an earlier decision used it. Conversely, preserving a version does not establish that it remains appropriate for current use.

These are intended benefits and ways to test them, not measured performance claims for the current Property Pack candidate. No time-saving percentage or commercial return is implied.

What the approach asks of people

Definitions need owners and maintenance. Difficult cases need discussion. Sources need enough context to be interpretable and enough permission to be used. Boundary agreements need representatives who understand both sides. Generated representations need checks that they preserve the meaning and the constraints they claim to carry. These are real costs, not work that vanishes because an ontology exists.

Bounded work helps make those costs manageable. A group can address a specific question inside its documented scope, reuse an established meaning where it fits and record a justified exclusion. It need not model every possible property situation before producing something useful. But a smaller scope must remain visible: an agreement for one purpose should not quietly become a rule for all purposes.

Reasonable objections deserve more than slogans

“We already have a schema.”

A schema is valuable. It can describe a document's structure and constrain permitted content; its accompanying documentation may contain substantial domain knowledge. The question is whether those materials make the necessary subjects, relationships and conditions explicit across uses. A well-formed field called “date” still needs a meaning.

OPDA's earlier PDTF schema and its separately identified schema-derived ontology remain useful evidence for requirements, traceability and future migration analysis. The recorded decision basis does not treat either as an OPDA-endorsed predecessor scheme or as automatic authority for future domain meaning. The approach changes how that meaning is established; it does not discard everything learned from the existing material.

“Why not agree one definition of property?”

Shared definitions help when the intended meaning really is shared. But a building, a dwelling, a right or interest described by a title record, and a subject considered in a lending decision are not made identical by calling them “property”. OPDA works within documented contexts so useful local distinctions remain explicit. At their boundaries, people must establish what information another context needs and what a proposed connection preserves. This is not permission for arbitrary disagreement; it is a way to locate and examine the difference.

“Why not import an existing ontology such as FIBO?”

Reusing established work can avoid reinvention. ADR-0063 identifies the Financial Industry Business Ontology as a useful reference and potential source of concepts for Finance and Banking. It also says the group should establish its scope and questions before considering wholesale import. A term's existence in another ontology does not demonstrate that it answers this group's question with the right scope. Other modelling approaches also support evidence and participation; OPDA's choice is not a claim that all alternatives impose one top-down model.

“Is an identifier enough?”

An identifier helps us refer to something consistently. It cannot establish that two organisations assigned it to the same subject, or decide what happens when a dwelling is subdivided. Review the identity criterion and the evidence for applying it. The identifier supports the agreement; it does not replace it.

“Will this replace our forms or require one database?”

Not inherently. Forms and systems can offer useful views of a shared agreement, and organisations need not centralise all their records to share meaning. Implementation and migration still require separate choices and testing. A model does not automatically update an existing form, translate a historic field correctly or make every recipient adopt a new exchange.

“Is AI deciding the standard?”

No. AI can help examine evidence and draft candidates. Its proposals may omit context, inherit a source's mistake or produce plausible wording unsupported by evidence. Practitioners judge domain meaning; technical checks assess specified conformance; recorded OPDA authority governs promotion and release. Agreement between models does not substitute for those responsibilities.

“Can we contribute when the evidence is restricted?”

Often a bounded description of the question, an authorised excerpt or a carefully labelled fictional example can help. Describe the source's relevance and restriction without copying material you may not share. A fictional example can reveal an ambiguity, but it is not proof of a claim about real practice. The member guide is the home for current participation and evidence-handling instructions.