Industrial data exchange is routinely framed as a problem of format — EDI, XML, spreadsheets, APIs. That framing misses the point. Format is the easy part; meaning is the hard part. ISO 8000-110 specifies how characteristic master data shall be exchanged so it is not merely syntactically valid but semantically unambiguous — interpretable by any receiving system without human mediation. This article examines what the standard requires, why those requirements are technically necessary, and what implementation looks like in practice.
The formal scope of ISO 8000-110 is precise and deliberately bounded. It specifies requirements for three things that, taken together, define what it means for a master data exchange to be complete — not merely correctly formatted, but correctly interpretable.
Master data messages must validate against a defined formal syntax — structurally correct, machine-parseable.
Each property must reference a data dictionary entry by unambiguous identifier (IRDI) — not a free-text label.
Each message must reference computer-processable specifications governing which properties, values and constraints apply.
The standard applies specifically to characteristic data — data that describes the properties of items rather than transactions, events or relationships. It is the data that answers the question “what is this thing?” — dimensions, materials, performance characteristics, tolerances, compliance references. ISO 8000-110 addresses the exchange boundary: the point at which data pass from one system or organisation to another.
At the core of ISO 8000-110 is a concept simple to state but with profound practical implications. The standard establishes that the property-value tuple is the fundamental building block of electronically stored characteristic data. It is not the data value alone. It is not the field name alone. It is the combination of a value with an unambiguous reference to the concept — the property — that gives that value its meaning.
The alternative — the property label — is systematically inadequate. The string “INNER DIAMETER” may appear in one system as “inner diameter”, in another as “ID”, and in a fourth as a label in a different language. Even when identical, the meaning may differ: measurement method, reference standard, tolerance regime and unit convention may all vary without being visible in the label.
An IRDI reference to a defined concept in an accessible data dictionary.
The actual data — a number, a string, or a structured value tuple for quantities.
Expressed via dictionary reference, ISO 80000 / IEC 80000 symbol, or property definition.
Nominal, minimum, maximum or typical — explicit when design-critical.
The IRDI is not a column header that a person has written. It is a formal identifier referring to a concept defined independently of any organisation’s internal terminology, that any receiving system can resolve to retrieve the full definition. This is the transition from data a person can read to data a machine can understand — a categorical difference in what the data are.
The requirement for semantic encoding in ISO 8000-110 has two levels, distinguished primarily by the licensing conditions attached to the data dictionary used. At both levels, the requirement is that each property-value tuple shall use an unambiguous identifier to refer to a data dictionary entry that defines the applicable property. The data dictionary shall be accessible in electronic form — either through a downloadable database or through an application programming interface available over the internet — so that any party receiving the data can retrieve the definition of any property referenced in the message without prior knowledge of the sender’s internal systems.
The standard also addresses a subtlety that organisations managing large data sets frequently overlook: the intellectual property implications of the identifiers themselves. If the identifiers in a data dictionary are subject to copyright, incorporating them into a data message may create a joint work, with consequences for ownership of the resulting data. ISO 8000-110 requires that the data dictionary be one in which identifiers are either in the public domain or are specifically licensed for use without the creation of a joint work. This requirement is not a procedural formality. It is a protection for the data owner’s rights over their own product information — ensuring that encoding product data against a defined concept does not inadvertently transfer intellectual property rights to the dictionary provider.
One of the most practically significant sections of ISO 8000-110 addresses the representation of quantities — the most common category of characteristic data in industrial contexts and the category most frequently represented in a form that is technically ambiguous. The standard specifies that when a property-value tuple represents a quantity, the unit of measurement shall be expressed in one of a small number of unambiguous ways: by reference to a data dictionary entry for the unit, by a symbol defined in the ISO 80000 or IEC 80000 series, or by specification within the property definition itself. What the standard does not permit is a free-text unit designation — the kind of entry that leaves a receiving system unable to determine whether “in” means inches, whether “t” means metric tonnes or short tons, or whether a temperature expressed without a unit reference is in Celsius or Kelvin.
Qualifiers matter just as much. A value of 10 mm may mean a nominal dimension, a minimum, a maximum or a typical value. A maximum wall thickness of 10 mm is not the same design constraint as a nominal wall thickness of 10 mm. ISO 8000-110 requires qualifiers be expressed explicitly — resolving an ambiguity that costs engineering organisations significant time and error across every product data exchange.
The third pillar of ISO 8000-110 — conformance to the data specification — addresses a requirement distinct from semantic encoding but equally necessary for genuine interoperability. Even a semantically correctly encoded master data message may fail to meet the specific requirements of a particular trading relationship if it does not conform to a data specification governing that relationship. A data specification is a computer-processable set of rules about the format and content of a master data message: the requirements that govern which properties shall be present, what values are permissible, and what constraints apply.
ISO 8000-110 requires that each master data message reference the data specification or specifications to which it conforms, and that those specifications be available to all interested parties in computer-processable format. This means conformance can be checked automatically by software, not only by human review. The practical value of this requirement is most apparent in supply chain relationships where the same product data must simultaneously satisfy the requirements of multiple downstream systems: a customer’s ERP system, a regulator’s surveillance platform and a certification body’s data repository. A master data message that references its applicable data specifications enables each receiving system to validate it automatically against its own requirements, without requiring the sender to produce a separately formatted version for each recipient. retrieve authoritative product data without prior knowledge of the manufacturer’s internal systems.
The requirement that data specifications be expressed in computer-processable form is also the foundation for scalable data quality assurance. Manual review of data conformance is not a sustainable approach at the volumes that modern supply chains demand. The architecture of ISO 8000-110 — syntax reference, semantically encoded property-value tuples and explicit data specification references — is designed so that the entire exchange can be validated computationally, end to end, without human intervention in the conformance checking process itself.
ISO 8000-110 was first published in 2009. The second edition, published in 2021, reflects more than a decade of implementation experience and a significant evolution in the landscape of industrial data exchange. The changes in the second edition are not merely editorial. They reflect a matured understanding of how the standard is applied in practice and where the most consequential implementation failures occur.
The most substantive terminological change is the replacement of “property-value pair” throughout with “property-value tuple”, and the corresponding replacement of “data value” with “value tuple”. This change is not cosmetic. It reflects the recognition that a property value is not always a simple scalar. When a property-value tuple represents a quantity, the value component is itself structured: it may consist of a number, a unit of measurement and a qualifier of measurement, each of which may in turn be represented by a dictionary reference. The pair formulation implied a two-element structure that was insufficient to capture this complexity. The tuple formulation is precise.
The addition of Annex B, which references an example schema for exchanging master data that are characteristic data, provides implementors with a concrete reference point for data exchange format — a resource the first edition did not supply. The normative Annex A, which assigns an unambiguous object identifier to the document itself in accordance with ISO/IEC 8824-1, supports automated identification of the standard in open information systems. As automated conformance checking tools become more widely deployed across supply chain and regulatory contexts, the ability to identify unambiguously which version of a standard a data exchange conforms to becomes a practical operational requirement rather than a theoretical concern. The second edition was developed within ISO/TC 184, Subcommittee SC 4, Industrial data — the committee that has maintained technical authorship of the ISO 8000 series since its inception.
The architecture specified in ISO 8000-110 — property-value tuples with IRDI-based dictionary references, dictionaries accessible under licensing conditions that protect data portability, explicit unit and qualifier representation, and conformance to computer-processable data specifications — is not the architecture that most master data management implementations currently use. It is, however, the architecture that KOIOS K:spec implements from its foundation. That is not a positioning claim: it is a description of how the system works at a technical level.
Every property recorded in K:spec is expressed as a property-value tuple in the sense defined by ISO 8000-110. Every property identifier is an IRDI, globally unique and resolvable through the KOIOS concept dictionary — managed under ISO 29002-conformant principles and operated under licensing conditions that specifically protect the data owner’s intellectual property rights over their own product information. Every data exchange from K:spec references both the applicable formal syntax and the applicable data specification, enabling automated conformance checking by any receiving system without human mediation.
This is not an incremental improvement over conventional master data practice. It represents a different understanding of what product data are: not a collection of fields populated for a specific system’s operational purposes, but a set of formally defined, semantically portable, machine-readable assertions about a product that can travel across any organisational or system boundary and remain fully interpretable at the point of arrival. For organisations whose product data must be exchanged across supply chains, across jurisdictions and — increasingly — into the infrastructure of Digital Product Passport ecosystems, this architecture is not aspirational. It is the requirement. ISO 8000-110 defines it. K:spec implements it.
For organisations assessing whether their current data exchange practice meets the requirements of ISO 8000-110, the first question is not about technology. It is about the structure of the data themselves. Are property values accompanied by unambiguous references to defined concepts in an accessible data dictionary? Are units of measurement expressed in a form that any receiving system can interpret without human inference? Are qualifiers of measurement explicit? Does each data message reference a formal syntax and a data specification in computer-processable form? Are the dictionaries used for encoding accessible — at no cost to the receiver — under licensing conditions that protect data portability?
For most organisations, the honest answer to several of those questions is no. That reflects the reality of how industrial product data have historically been created: for internal use, in systems configured for operational purposes, by people whose primary task was not data governance. The question is not whether the current state is adequate for the systems currently in use. The question is whether it will be adequate for the data exchange demands now arriving from supply chains, from regulators and from the Digital Product Passport frameworks that are advancing through international standardisation. Those frameworks are not designed around the data that manufacturers currently hold. They are designed around the data that ISO 8000-110 describes.
The gap between the two is real, measurable and closeable. The standard sets out exactly what is required. The KOIOS K:spec platform is built to deliver it.