Definition of Action

Definition Last Updated 17-Dec-2015 12:51

An action is a unit of activity for which no internal description is provided, i.e. it is an event (potentially of extended duration), characterised entirely by a specified change in state. Defined thus, in terms of changes of state, an action may be seen as the occurrence of an exercise of a capability; actions may be performed by a performer – but they might also simply be reactions, i.e. an effect arising from a particular cause.

It may be possible, and indeed it usually is, to divide an action into smaller actions; whilst this may appear to convert an action into an activity (given that the description of a thing should be fixed unless it is acted upon), activity is the fundamental concept: an activity may be described as an action only for the actor who is unable to interrupt it.


Action is a defined term of Enterprise Architecture. Action is a defined term of Business Analysis.


Article Last Updated 17-Dec-2015 12:51

The characterisation of action applies equally to what is done intentionally, such as picking up a pen, and what just happens, such as two raindrops coalescing on the window on a rainy November day in England.

As far as picking up my pen is concerned, the action is reaching out and grasping (considered as a whole); in the case of the raindrops, the action is touching, which then causes them to coalesce.

Whilst it might be thought that an action should have a single, primary actor who does something that causes something else to happen, it is easy to construct counterexamples.

Consider two atoms of hydrogen (H) of constant velocities that collide to form a molecule (H2). If either were stationary, it might be natural to call the other the primary actor since it does the “colliding”. However, (inertial) motion being relative, there is no reason to prefer either as the actor, so there is none in this case.

Contrariwise, when there is a genuine asymmetry there may be case for identifying some actor as the primary actor; in general, however, an action need not have a primary actor.

An action is minimally specified as a transformation by its initial and final states; whilst every transformation is an effect of something, and therefore has a cause, the primary actor may not be identifiable.


Definition of Activity

Definition Last Updated 17-Dec-2015 12:51

Activity is continuous change in one or more observables during a specified period of time, which, if the activity is current, is open-ended to the future and may be open-ended in the past; an entity engaged in activity has that activity as a behaviour is an performer.

An activity is recognised pattern of change, i.e. a mapping from particular initial states to particular outcomes, which may or may not be goals.


Activity is a defined term of Enterprise Architecture. Activity is a defined term of Business Analysis.


Article Last Updated 17-Dec-2015 12:51

Activity vs. Behaviour

Section Last Updated 17-Dec-2015 12:51

All behaviour is activity, but not all activities are behaviours: behaviour requires the identification of an entity to which the activity can be attributed; activity does not.


  • “The lioness is creeping through the long grass.” describes a behaviour: the lioness is attributed the activity
  • “Some of the long grass is waving.” describes an activity: some grass is not an entity to which the activity can be attributed
  • “There is movement in the long grass.” is ambiguous: “in” suggests that only part of the long grass is moving, but it is not well-identified enough to be an entity; but, if the emphasis were on “the long grass”, suggesting a defined border separating long grass from shorter grass, the long grass might be in fact be an entity in its own right1If this type of ambiguity (not some actual threatening movement in long grass somewhere) turns out to be a significant problem for anybody I would a) be surprised b) like to know about it. .

An Activity

Note that the performer engaged in an activity need not be aware of or even capable of recognising the initial states or outcomes that define an activity: activity may just happen.

Is Waiting an Activity – or the Absence of Activity?

Waiting is an activity because waiting for something entails continuous observation in order to determine when that something (the state of interest) occurs. Even if an observer only observes intermittently, it must still monitor the passage of time in order to determine when to make an observation. Waiting for something also implies an expectation or belief – whether that justified or not – that that that something will occur.

According to the definition given here, when a performer does not engage in a particular activity, or perform a particular action until caused to do so by some external event, the performer is simply not doing a particular thing, the performer is not necessarily inactive: the performer may be doing something else.

When, for example, a piece of hardware or software is said to be waiting for input it may be:

  • Polling (repeatedly observing) the input channel as fast as possible (e.g. in a tight causal loop) or it may be
  • Polling a timer prior to checking the input channel at particular intervals
  • Inactive until triggered into activity by a specific signal such as an interrupt

Notes   [ + ]

1.If this type of ambiguity (not some actual threatening movement in long grass somewhere) turns out to be a significant problem for anybody I would a) be surprised b) like to know about it.


Definition of Agreement

Definition Last Updated 17-Dec-2015 12:51

Agreement is the mutual commitment of two or more parties (individual or organisational entities including political states) to adopt specific policies (normative specifications), typically subject to the provisions of implicit or explicit higher-level policies, such as may govern failures to meet the commitments made.

An agreement is a specific policy.


Contract, treaty, memorandum of understanding.

Note that a law is not an agreement: it is a policy of a state (or supra-national) organisation whose norms apply to its members (citizens or subjects) irrespective of explicit commitment to them.


Agreement is a defined term of Enterprise Architecture. Agreement is a defined term of Business Analysis


Article Last Updated 17-Dec-2015 12:51

Types of Agreement

Recognised types of agreement include

  • Agreement, e.g. Non-Disclosure Agreement (NDA), Service Level Agreement (SLA)
  • Contract, e.g. Contract of Employment, Contract of Sale, etc. where the policies typically include the provision of something and financial remuneration for it, and IT Service Contracts where the agreement may be simply that provided certain conditions are met a certain capability will be exercised on behalf of the service requester
  • Memorandum of Understanding, may be, but typically is not legally binding1For further insight into MoUs see Wikipedia 

Treaty, sometimes legally binding

Notes   [ + ]

1.For further insight into MoUs see Wikipedia

Architecture Model

Definition of Architecture Model

Definition Last Updated 17-Dec-2015 12:51

An Architecture Model is a description of one or more entities in terms of other entities (contained components and components of the environment) and their attributes, specifically including innate properties and behaviours as well as the relationships between entities.

A formal architecture model is such a description codified in accordance with an (architecture) standard (which may reference subsidiary standards), such as an architectural Framework or Metamodel.

A formal model should identify the specific standard (by reference to dates and times, version numbers, standard document status, etc. as necessary) so that the compatibility of two or more models is determinable.


Architecture Model is a defined term of Enterprise Architecture.


Article Last Updated 17-Dec-2015 12:51

No discussion yet.


Definition of Attribute

Definition Last Updated 17-Dec-2015 12:51

An attribute is a state comprising one or more characteristics, i.e. observables1Though not necessarily under all circumstances, it must be observable in principle. or tags, that has been attributed2i.e. given to, from the latin attribuere (to give to) formed by composing ad (to) and tribuere (to give), typically implying intrinsicality or innateness; the appropriate verb for connecting characteristics to places and times in the absence of identified entities would be allocate, ad (to) and locare (to place). to a specific entity (i.e. some thing that has been identified3See the article on Entity for information on the identification of entities.). If the characteristic is named4Here name means a simple designator, such as mass; the characteristic whose name is mass could also be described as, “the intrinsic constant of proportionality between force and acceleration”. In general, anything that designates may be called a name, but some things are more naturally names than others., it may also be referred to as a property.

Behaviours and relationships are also attributes: a state may effectively describe an activity, which, when ascribed to an entity, may be referred to as a behaviour, and tags encode relationships, .

Distinct entities may have attributes bearing the same name: the assignment of characteristics to things is achieved via the specification of observations; this is discussed in detail in the main article.


[Attribute] is [state] attributed to [entity]

[State] attributed to [entity] is [attribute]

[Entity] has [state] as [attribute]


Attribute is a defined term of Enterprise Architecture. Attribute is a defined term of Business Analysis.


Article Last Updated 17-Dec-2015 12:51

Note that although entities may have attributes, entities must first be identified before anything can be attributed to them. Identification is performed by the application of rules (containing criteria that constitute a reference description or standard for the identification of something) to the values of observables. Attributes are therefore ontologically secondary to observables – they are only attributes because they belong to an entity.

Thus, when modelling, there is no fundamental type attribute: an attribute is just a set of characteristics that has been associated with an entity. An “attribute” that is not related to an entity5And so, by definition, not really an attribute. is just the concept of a characteristic that something might – or might not – actually have.

Note that characteristic in the definition is a noun; whilst a particular value of an attribute might be characteristic of (adjective) something (e.g. Colour: green, characteristic of plants), a characteristic defines the nature of the observation and its measurement or assessment.

Measurements are typically absolute and objective (though subject to errors, systematic and random) whereas assessments such as “good” or “bad”, “high” or “low” are typically subjective or relative, and prone to their own particular kinds of errors. Attributes such as “good”, “high”, etc. may be considered secondary or derived attributes as they are typically derived by relating some measured characteristic to one or more reference values and therefore have implied criteria.

It is also important to be clear about the distinction between parts of an entity and characteristics of an entity.

Parts vs Attributes

Section Last Updated 17-Dec-2015 12:51

An animal may have legs, a car may have wheels; are legs and wheels attributes or (component) parts?

Firstly, we can say that whilst legs and wheels can be observed, they are not observables: observables are intrinsic characteristics such as length or mass, whose values are independent of any entity identification. However, in order to be able to say, “There’s a leg,” one must have performed an identification whose result is “leg”. Therefore, since one cannot refer to (point at) specific legs or wheels without having previously circumscribed and identified them as entities, they are not independent of identification and so cannot be observables as defined.

Could legs or wheels be tags, the only other kind of characteristic, instead?

Again the answer is no. A tag says something about an entity; what does a leg say about an animal it belongs to other than, “I am the leg of…”, which is clearly a relationship – and relationships require the identification of entities. For “I am the leg of…” to make sense, the leg has to be an identifiable entity, and if it’s an identifiable entity then the most natural conclusion is that it’s a part.

Thus, anything that can be identified as an entity within another entity is a part and not an attribute of the latter.

That having been said, the number of legs possessed by an animal is a characteristic: legs can be counted and thus number of legs is an observable.

Note also that an attribute may itself be compound: the (physical) size of an object might be specified in terms of length, width and height and these might be individual attributes or they might be grouped together as a single size attribute with slots for the three individual values. When an attribute has parts, the parts are not attributes of the attribute, they are just parts of it – but they are attributes of the object, i.e. any part of a compound attribute of an entity is also an attribute of that entity.

Discussion of how one distinguishes an intrinsic, dependent part (such as my liver) from an independent part (the marble I swallowed for a dare) is deferred to another time.


Physical objects are typically coloured, i.e. have an attribute called “Colour”. In giving a value for the colour of something we might use simple colour words (red, purple, orange), or a formal colour-space measurement in terms of the brightness of red, green and blue colour components (RGB system). Whether we wish to use words or RGB pixel values only affects the data type of the attribute.

Relationships to Other Fundamental Concepts

The relationships between attributes and other the fundamental concepts (characteristics, properties, observables, states and tags) are shown in the figure below.


Figure – States, Characteristics, Properties, Attributes, Observables and Tags

Related Entries

Component, Property

Notes   [ + ]

1.Though not necessarily under all circumstances, it must be observable in principle.
2.i.e. given to, from the latin attribuere (to give to) formed by composing ad (to) and tribuere (to give), typically implying intrinsicality or innateness; the appropriate verb for connecting characteristics to places and times in the absence of identified entities would be allocate, ad (to) and locare (to place).
3.See the article on Entity for information on the identification of entities.
4.Here name means a simple designator, such as mass; the characteristic whose name is mass could also be described as, “the intrinsic constant of proportionality between force and acceleration”. In general, anything that designates may be called a name, but some things are more naturally names than others.
5.And so, by definition, not really an attribute.

Pin It on Pinterest