Proposed Modifications of the Artifact Ontology

Artifact

As information content entities are artifacts the class name of artifacts that are material objects will be changed to Material Artifact.

As debate sometimes ensues over whether a type of artifact is an object or object aggregate and as the benefit of classifying artifact as a subclass of object to reasoning or interoperability of data seems insignificant, the class of Material Artifact will be reclassified as a direct subclass of Material Entity (BFO_0000040).

As a guiding principle, the direct subclasses of (Material) Artifact are differentiated by a function that is borne by instances of the subclass. A few of the direct subclasses, those having “Component” in the class name, violate this principle. Instances of such classes did not come into being through intentional design to realize processes of making some other artifact (e.g. an hydraulic fluid reservoir was not brought into existence in order to realize the process of making an hydraulic power source). Another issue created by such classes is that they dispose the ontology to have multiple inheritance among asserted classes. Cases of this exist already as evidenced by the classification of the Hydraulic Valve class as a subclass of Hydraulic Power Source Component but not as a subclass of Valve. All classes with “Component” in their class names (and also “Element” and “Part” as these seem equivalent to “Component”) will be removed from the ontology. Subclasses of the classes to be removed will be refactored and placed into appropriate places within the remaining taxonomy.

The first level of descendants from the class Material Artifact form a very mixed bag from a perspective of specificity of purpose. Medical Artifact, Legal Instrument, Infrastructure, Tool, and Portion of Processed Material stand as siblings to the classes of Flywheel, Heat Sink, Laser, Filter and Tripod. This issue may get corrected as more types of artifacts are added to the taxonomy and more general (i.e. less specific) parent classes are found that group the more specific types. That solution would be facilitated by there being a methodology in place to consistently add new content, rather than the more haphazard one of adding content as needed to satisfy project requirements. A suggested route is to review the Nice Classification (NCL) (https://www.wipo.int/classifications/nice/en/). The suggestion is not that the Artifact Ontology replicate the hierarchy of the NCL but rather that the NCL is reviewed as a source for additional content that would fill in gaps of the ontology.