A field guide to false friends in industrial standards

False friends in industrial standards are deceptive similarities that lead to semantic mismatches with high integration costs. These errors often escape initial detection because they are flagged lowest by automated matchers and reviewed least carefully by humans. Recognising the recurring patterns of such errors is critical for accurate ontological alignment. This lesson explores the most frequent false friend shapes to help practitioners avoid these costly misalignments.
Partition and Modelling Convention Clashes
Shape 1, the partition clash, occurs when two concepts are semantically incompatible despite superficial similarity. For example, ido:Object is not IfcObject. IDO declares Object disjoint from Temporal, thereby excluding activities. IFC places IfcProcess under IfcObject, which includes processes. This mismatch causes every process to become contradictory when equated. The partition clash is particularly pernicious because it challenges fundamental class hierarchies and can invalidate entire reasoning chains.
Shape 2, the modelling-convention clash, involves concepts that are not disjoint but are modelled differently. ido:System is not IfcSystem. IDO does not declare FunctionalObject disjoint from PhysicalObject, allowing one entity to be both. In IFC, IfcSystem is under IfcGroup, which is disjoint from IfcProduct. IDO allows function and substance to coexist; IFC does not. This difference affects how systems are modelled and can lead to incorrect semantic inferences when the two are conflated.
Category and Quality Clashes
Shape 3, the category clash, involves concepts that are fundamentally of different kinds. ido:Site is not IfcSite. IDO treats Site as a Location, which is disjoint from PhysicalObject. IFC treats IfcSite as an IfcProduct, a physical entity with geometry. This difference reflects contrasting industry perspectives: the process industry treats space as a frame of reference, while construction treats it as a modelled object. Misalignment here can result in incorrect categorisation and misinterpretation of spatial data.
Shape 4, the quality/datum conflation, is a frequent source of error in data alignment. ido:PhysicalQuantity is not IfcPhysicalQuantity. The former is a length borne by a pipe (a Dependent), while the latter is a recorded measurement value. The correct alignment is with ido:QualityDatum. This mismatch is often overlooked because lexical matchers rank it highly. It is the pair a lexical matcher identifies first, making it a particularly deceptive false friend.
Granularity and Container-Content Confusions
Shape 5, the granularity clash, involves concepts that differ in their temporal or structural resolution. ido:Event is not IfcEvent. IDO’s definition allows an Event to be temporally scattered (e.g. a week of heater cycles as one event), whereas an IfcEvent is a discrete trigger. This distinction is critical in simulations and scheduling. The mismatch arises from different assumptions about event structure and temporal granularity.
Shape 6, the container-content confusion, involves concepts that are semantically distinct in their roles. ido:Stream is the fluid, whereas IfcFlowSegment is the pipe. This distinction is critical in process engineering, where the substance and its containment are modelled separately. This type of mismatch is common in integration projects and can lead to incorrect process models if not carefully managed.
What to take away
False friends are most dangerous when they are not immediately detected by automated tools or reviewed critically by domain experts. The recurring patterns of mismatch, partition, modelling convention, category, quality, granularity and container-content, require careful attention during alignment. Each shape has a characteristic signature that can be identified through a structured comparison of class relationships and semantic definitions. Recognising these patterns early in the integration process prevents costly downstream errors.
Reference
| Lesson | 11 of 15 |
| Outcome | Recognise the recurring shapes of false friend, and check a candidate pair against each before accepting it. |
| Worked repository | industrial-ontology-crosswalks |
| Engine | open-ontologies |
