Physical Twin

A comprehensive description of some other entity as it is or was in tha past, is an existential instance of the CRO category Physical twin (PT). The latter carries the property has_referent whose range is Physical entity, meaning that a PT can describe a material entity, a process, or an agent.

A Physical twin attempts to be a comprehensive description of its referent, meaning that in general it consists of numerous items of information relating to the sub-parts of the root entity, as well as characteristics (i.e. emergent behaviour) at any granular level. In the case of manufactured artifacts, a representation of the parts structure is very common, which may be achieved with an instance of the Meronomic model category, described below.

1. Individual Twin

CRO breaks the Physical Twin category into the following sub-categories:

  • Individual Descriptor

    • Agent descriptor

    • Material entity descriptor

    • Process descriptor

  • Population Descriptor

Consider first the two top level categories, which address description of individuals and populations respectively. An individual descriptor either has an individual referent, such as a particular person, device, process etc, or else a 'lot' or a 'package' of identical individuals, as in the case of purchased vaccines, packets of Aspirin and similar. A 'lot' is usually understood as a manufacturing level concept, whereas packets are usually consumption level. For practical purposes, they function as individuals, since the contents are assumed to be completely identical.

2. Population Twin

In contrast, Population descriptors have a collection of individuals as the notional referent. Concretely, the referent is stated via the selection criterion applied to some larger population, such as 'all pregnant women in the London area', or 'crude carrying ships with >= 1.7m barrels capacity'.

A Population descriptor differs in important ways from an Individual descriptor:

  • the referent has no identity as such, and it may be a privacy requirement to obscure or remove all identifying information of the individual level input data;

  • the base data are generated by aggregating functions, e.g. sum(), avg(), count() etc;

  • more advanced characteristics can be generated by statistical functions, e.g. variance, standard deviation etc.

Accordingly, a Population descriptor provides a general picture of its population, rather than information about any specific entity within it.