The structure/behavior dichotomy
This article discusses how we can describe the state and progress of things by carving the world into discrete entities and events, and describing the effects of the events on the entities.
Contents: The dichotomy. Describing entities, events and effects. Composite entities and events. The life history of an entity.
The dichotomy
We cannot know things in the world directly; we can know only the models or descriptions (neural and/or verbal) we make of them. At the end of the day, a description of a thing, even a mathematical one, will always be a description, not the thing it describes.
When describing things in the world, we divide them into two categories.
The universal nature of this dichotomy is evident in the many words we use to differentiate structures from behaviors, such as: actor-act, entity-event, object-operation, form-function, product-process, and state-state transition. Also, stocks and flows in System Dynamics.
The dichotomy may be generalized as structural-behavioral, spatial-temporal, continuant-occurrent, or endurant-transient. And it may be associated with the noun-verb distinction in verbal language. It seems fundamental. But is it?
The universe is an ever-unfolding process. Every structure we observe at the present moment is a transient side effect of that process.
(At the level of atomic particles, a structure such as the moon changes from moment to moment. But at the level of abstraction we describe it, it changes so slowly that change is imperceptible. We observe, envisage and describe those structures that persist long enough to be of interest to us.)
We live in spacetime. Time passing = change happening. Everything changes. At the level of atomic particles, even a structure as apparently persistent as the moon changes from moment to moment. However, at the level of abstraction we describe it, it changes so slowly that change is imperceptible. So, although all structures are transient, what we observe, envisage and describe as structures are those things that persist long enough to be of interest to us.
This two-by-two grid maps the structure-behavior dichotomy to the description-reality dichotomy.
Describing entities, events and effects
Entity event modeling regards an information system as a collection of related entities that are affected by events, and defines the effects of those events on those entities. The diagram shows how one effect relates one event to one entity, and so resolves the many-to-many relationship between events and entities.
This section discusses how we may describe structural entities, behavioral events, and the effect of events on entities.
Entities
As sentient beings, we identify things are "out there". Nature gives us a hand in the form of phase boundaries, which separate solid structures from the space that surrounds them. We see stones on the ground, trees in a forest, fishes in the sea, plants and animals in the atmosphere, and bodies in outer space as distinct entities.
Entity: a thing that is persistent; it as has state/structure that may be changed by events.
Note that the boundary we draw around an entity may be fuzzy. The boundary that separates a mountain from surrounding hills may be unclear. And an entity we describe, such as a flock of birds, may be divided into parts that are distributed in space.
Terms and concepts used in describing entities include
State: the features of an entity at a moment in time (as may be recorded in a memory as the values of state variables).
Part: a component of an entity that can be described as an entity in its own right.
Invariant: an attribute whose value does not change over the life of an entity, or the time period of interest to us.
Variable: an attribute whose value changes over time. Variables can include:
The current state of an entity is a "side effect" or result of past events (as the surface of the moon shows the effect of being struck by asteroids in the past).
Event sourcing: generating the state of an entity at a particular time by replaying past events stored in an event log.
Hysteresis: the situation where the current state of an information or real-time process control system lags behind the current state of the real world entities that the information represents.
Events
Event: a behavior, something that happens and may affect one or more entities. An event is a discrete state change that moves a system from a prior state to a successor state. Think of an activity, operation, process, method or transaction.
In information system, a event is represented as an input or message carrying parameters.
Parameters: the features of an event, the data it conveys about the effect of the event on the entities in the system.
Effects
Effect: the state change that happens when an event affects an entity. For example, the effect that a business event (such as order, invoice, payment, or complaint) has on a business entity (such as customer, order, product type, of warehouse).
The effects of one events on one or more entities may be defined declaratively in a contract that includes precondition and post conditions
Preconditions: the state of entities relevant to an event, before the event.
Post conditions: the state of entities affected by an event, after the event.
Entity event modelling
To define the structure and behavior of an information system, we can encapsulate its internal components and processes behind event or service contracts and interface definitions. The grid below is variation of the generic meta model in the ArchiMate system architecture definition standard.
For more, read the article on entity event modeling.
Composite entities and events
Composite: a whole divided into parts, or a multi-level structure that divides a whole into successively smaller parts, down to parts regarded as atomic.
The granularity of an entity depends on the interests of observers. We may define both atomic entities (not further decomposed in the domain of interest to us) and composite entities.
Composite entity: an entity that relates smaller entities in a structure.
Microstate: the state of an atomic entity's parts, such as position of a molecule in a gas cloud, or the mass of a predator or prey animal in an ecology.
Macrostate: the state of a composite entity, For example, the state of an ecology may be represented by total biomass, and the populations of predator and prey species
In practice, when describing a large entity we may define it at several levels of granularity, by composition and decomposition. Consider for example, the management structure of a business. And consider how the current state of a tennis match entity is described by game, set and match scores..
Similarly, the granularity of an event depends on the interests of observers. We may define both atomic events (not further decomposed in the domain of interest to us) and groups and sequences of events.
Composite event: a group of atomic events that places no constraint on their sequence. For example, the event "annual service" comprises several atomic events such as replacing a car tyre, and replacing a battery.
Process: a pattern that sequences two or more events in time, and may have parallel paths.
Circular process: a process that returns to the same state, or near to it.
Progressive process: a process that continually advances to a given end.
In practice, when describing a long process we may define it at several levels of granularity, by composition and decomposition. For example, the process of a tennis match may be divided into sets, then games, then points, then ball strikes.
The life history of an entity
In an information system, a data entity (identified by a primary key) is created, changes state over time, and is deleted. This life history may be represented in a state chart.
In biology, an organic entity (identified by its DNA) is born, changes state over time, and and dies. You can see the structure of your body as a side effect of the extraordinarily complex biological process that starts with the fertilization of an egg, and ends in death.
In engineering, a machine (stamped with a number) is assembled, changes state over time, and is scrapped. During its life, the machine may be modified in some way, to perform a different function.
In every day discussion, when an entity of interest is born and dies, and when it mutates into a different kind of entity, may be fuzzy or undetermined.
How is an entity born?
The term "instantiate" is used ambiguously, to mean either "display" or "create". We may say that an entity instantiates (displays) the features of a type; or say that a process instantiates (creates) an entity.
An entity is created by a process of
An entity may be manufactured by
To manufacture a cake requires that a baker to performs a process (baking) by following a recipe of some kind. Biological evolution works by incrementally changing the DNA recipe that each new generation of an entity uses to manufacture itself.
How can an entity change?
An entity may change by state change, or by mutation.
State change: a change to one or more state variable values, which may be caused or triggered by an input, a time event, or a prior state change.
Mutation: a change to the definition of an entity (its invariants, variable types or rules), which has the effect of replacing one version or generation of the entity by another.
What you perceive to be a stateful entity may change in both ways over time. For example, IBM has changed over decades from a hardware manufacturer to a service provider.
So, what determines the continuity of an entity? The continuity of
Generally, the continuity of an entity's existence is reflected in the continuity of the name or identity to given to it by observers. This may correspond to the continuity of the materials it is made from, the place where it is found, or its visual appearance.
How can an entity die?
An entity may evolve - may mutate so much that observers see it as a different entity.
An entity may be replaced by a newer version or generation.
An entity may be die or be destroyed from the exhaustion of a resource it depends on, from the decay or removal of its parts, from random destruction by an external force, or from deliberate intervention by an external agent.
Which comes first?
The systems of interest are dynamic, in which entities change state in response to events. Which comes first? Entities or events? That is a chicken and egg question.
Many think first of structural entities (actors or objects), which, once created, can perform behaviors (activities, or operations) or experience events.
Some take the opposite viewpoint. Physicists see structures as the results or side effects of processes. Material structures (like the moon) are created, changed, and ultimately destroyed, by events.
The philosopher Whitehead argued reality consists of processes rather than material objects. "Thus rejecting the theory that reality is constructed by bits of matter that exist independently of one another." Wikipedia 2024
Programmers may describe a software system as a set of objects that are related, as may be a represented in a class diagram. Some object-oriented programmers deprecate taking a procedural view; some even speak of discarding the concept of a process. Yet we are advised to start system design with the processes (use cases or user stories) that interacting objects are required to complete, or the services the system must offer to clients.
In short, both perspectives are needed in practical system design.
Related articles
This is one of several related articles. If you want to read these articles in the context of a book, watch this space.