Context

Within a long running investigation process, specific acts of observation, assessment, orders etc are undertaken in order to advance the process toward its goal. Each such act generates more information. For each of these moments, there is specific context, i.e. details about when, where the act of observing or ordering occurred and by whom. This context is understood in CRO as a special Process subclass called Situation. An instance of Situation will connect the relevant participants, including resources, time and place together. Every epistemic act of information creation has an associated Situation.

We are concerned not just with how information was conceptually created in a mind or an xray machine, but the act of recording it within an information system. In SPLASH, we call this provenance. It is not specifically described by CRO, but is formally modelled in the SPLASH Reference Model in the version control part of SPLASH.

FIXME: ref above

A consequence of the distinction between the naïve and epistemic modes of representation is that time is represented in two ways.

Data committed to an information system at any moment is a snapshot of some real state of affairs at a moment (or extended interval) in time. Since the usual goal of an information system is to track the evolution of entities over time, the general structure of data (whatever its detailed form may be) is a history of snapshots that can be associated to reconstruct change to entities in the real world.

Snapshot histories take two forms, one for independent social / business entities (Person, Device, etc) and another for dependent real-time variables (heart rate, transformer voltage).

Snapshots of social entities are almost always achieved via routine versioning in a database or file-system, and it is nearly always only the most recent version that is of operational interest.

In contrast to social and business entities, histories of real-time variables are achieved in SPLASH by time-series data structures of various kinds. Thus, a time-series of a heart rate is not considered a series of 'versions of data', but rather a single history of a variable over a time period.

Real time variables almost invariably have temporal mathematical continuity, i.e. they vary in a way related to their underlying physical form, such as the pumping of the heart. They are therefore able to be represented as time-based trends. Changes in business entities on the other hand have no mathematical continuity: there is no way to graph changes of a person’s name or address over time, although other visual representations may be possible.

In SPLASH, real-time variable histories are represented as a special kind of Observation, which comes under a hierarchy of Acts within a part of the model addressing Investigation.

TODO: details.

TBD