The primacy of
behavior (services and processes)
This page is
published under the terms of the licence summarized in the footnote.
All free-to-read materials at http://avancier.website are paid for out of
income from Avancier’s training courses and methods licences.
If you find the web site helpful, please
spread the word and link to avancier.website in whichever social media you use.
With traditional building architecture in mind, you may assume the architect’s
aim is to define a form or structure.
General system theory encourages us to put behavioral views of a system
ahead of structural views.
This paper makes some observations about behavioral or dynamical views of a system.
Contents
The
primacy of behavior (recap)
The
structure/behavior dichotomy in systems
The primacy of structure in our mental
models
The
primacy of behavior in general system theory
Encapsulating
processes behind service contracts
Matching
the length of a process to the granularity of the system
Viewing
structural elements as long-running processes
How to differentiate "systems thinking" from
"thinking"?
What delimits the subject matter of systems thinking discussions?
Some discussions describe or classify discrete entity types.
A discrete entity is a thing separable from the rest of the universe.
Does every entity count as a system?
Some discussions are about politics, economics or sociology.
Does every discussion of those topics count as systems thinking?
Does every group of people count as a system?
The concern of general system theory (GST) is activity systems that operate in the real world, displaying behaviour of some kind.
A system can be characterised as parts that interact in regular, repeated or repeatable behaviors.
“A set of elements or parts that is coherently organized and interconnected in a pattern or structure that produces a characteristic set of behaviors." Meadows
The systems of interest here are activity systems, which have both structural elements and behavioral elements.
The table below shows many different word pairs that express the structure-behavior duality of a system.
|
The duality of activity
system elements |
|
|
Structural view |
Behavioral view |
|
Actors |
Activities |
|
Bodies |
Behavior |
|
Components |
Co-operations |
|
Entities |
Events |
|
Form |
Functions |
|
Items |
Interactions |
|
Objects |
Operations |
|
People |
Processes |
|
Roles |
Rules |
|
Stocks or Stores |
Streams or
Flows |
|
Workers |
Work |
Finding so many
ways to express the same dichotomy suggests there is a general truth.
The
structure-behavior dichotomy can be found in many systems theory sources.
At the time of writing, Wikipedia says a system is usually expected to have:
“Structure: parts (or components) that are directly or indirectly related to each other;
“Behavior: processes that transform inputs into outputs (material, energy or data);
“Interconnectivity: parts and processes are connected by structural and/or behavioral relationships.”
Chilean biologist Maturana defined a biological entity as system of components and processes in line with general system theory.
The original
definition (in Autopoiesis and Cognition:
the Realization of the Living - 1st edition 1973, 2nd 1980) is very
slightly simplified below.
An autopoietic
machine is: “organized as a network of processes of production
(transformation and destruction) of
components which through their interactions and
transformations continuously regenerate and realize the network of
processes (relations) that produced them; and
constitute it (the machine) as a concrete
unity in space in which they
[the components] exist by specifying the topological domain of
its realization as such a network.”
Boulding (1956) outlined two approaches to system theory.
First, definition of common system properties - identification of features common to systems in different disciplines
Second, classification of system types – e.g. arrangement of system types into a hierarchy of complexity.
Taking the first approach, Boulding interpreted the features of general system theory with respect to “higher level” social organisations.
This table compares a generalised description of a system (distilled from analysis of many different sources) with Boulding’s view and Meadow’s view.
|
General System Theory |
Boulding’s social organisation |
In Donella
Meadows’s terms |
|
A bounded structure
of |
A bounded organisation
of |
A bounded and coherent
organization/structure of |
|
actors/organs
that interact by playing |
individuals that
interact by playing |
elements/parts
that interconnect dynamically by
performing |
|
roles in |
roles in |
|
|
behaviors
to meet |
repeatable processes
according to |
behaviors
to meet |
|
goals, by
maintaining |
|
functions/purposes
by exchanging |
|
system state
and exchanging |
the state of their remembered mental images |
|
|
inputs/outputs
with each other and with |
and exchange messages. |
inputs/outputs
with each other and with |
|
entities outside
the boundary, using |
entities outside the boundary. |
|
|
system resources. |
|
We learn to describe the world using nouns and verbs.
This embeds in our minds the idea that structures and behaviors are separate things.
Experiments suggest most animals can remember structures/entities, but struggle to remember behaviors/events.
It seems our minds are tuned to favour a structural view of the world.
We think of structure as what things are made of - what things are - persistent, real, concrete.
We think of behaviour as what happens over time - what things do - transient, ephemeral, abstract;
Yet the two are intimately linked.
Your human substance is the product of continuous bio-chemical reactions.
You, as structural instance, are a side effect of those behaviours.
In describing the world, we impose structural types on things we perceive to be discrete entities.
Biologists define a species in terms of characteristics shared by each entity collection of similar biological entities.
But these types are fuzzy, biologists often redefine what constitutes a species.
And they know the story of life is not a hierarchical classification of species as they look today.
It is a cladogram showing endless division of populations over time, a side effect of their behaviour.
In engineering the world, we design and deploy structures to perform required behaviors.
We design structure types (motor car models, human roles, OOP classes).
A designed structure is, in effect, a collection of activity types that can be performed in a run time system.
We deploy structures instances (motor cars, human actors, OOP objects) to perform required behaviors
At run time, a structure instance can be seen as sequence of activity instances occurring over time.
The essential property of a system is repeated or repeatable system behavior(s) that can be described.
“The principal heuristic innovation of the systems approach is what may be called ‘reduction to dynamics’ as contrasted with ‘reduction to components’ ” Laszlo and Krippner.
“Reduction to components” means decomposing a body's structure into organs and studying each organ separately.
“Reduction to dynamics” means decomposing a body's behavior into processes.
General system
theory encourages us to put behavioral views of a system ahead of structural
views.
Ashby wrote:
“Cybernetics
does not ask "what is this thing?" but ''what does it do?"
“It
is thus essentially functional and behavioristic.”
“[It]
depends in no essential way on the laws of physics or on the properties of
matter.”
“[It]
deals with all forms of behavior in so far as they are regular, or determinate,
or reproducible.” (Introduction to Cybernetics (1956) W. Ross Ashby)
Thus, general system theory and cybernetics shifted the focus from system structure to system behavior.
From individual actors/components to processes performed by cooperating actors/components
A process runs over time, from start to end, or cyclically.
In short, pointing at an entity and calling it a system means nothing.
An entity or social group becomes a system when it is assessed as having repeated or repeatable system behavior(s) or output(s).
In discussion of a system, the participants must share an understanding of what behavior(s) or output(s) characterise that system.
You may know the famous architect’s phrase “form ever
follows function”.
In business architecture, it means that structural roles and organisation units follow required services and processes.
The everyday purpose of a business is to deliver services to its customers.
The business is composed of processes that can be performed whenever behavior is requested.
The architects focus on supporting and enabling the processes that deliver those services.
A business does not hire or buy actors (employees, machines) for their own sake.
It hires and buys these actors for their ability to perform the required processes.
EA is
concerned with abstraction of system description from operational system
contents and workings.
It
encapsulates systems and subsystems; which means defining system behavior in the form of interfaces and service contracts.
And within any system of interest, EA prioritises taking an activity-centric view over an actor-centric view.
It formalises end-to-end behaviors (which may be called value streams, business scenarios or business processes).
The primacy of behavior over structure is seen in many approaches to business systems planning.
· In Six Sigma’s SIPOC the P stands for process.
· In Computer Sciences Cooperation’s POLDAT the P stands for process.
· Service-Oriented Architecture (SOA) defines a system by its external behavior.
· EA frameworks define functions/capabilities independently of the organisation's management structure.
· The Open Group’s portability principle encourages definitions of systems by their interfaces.
Table 1 presents two contrasting viewpoints that capture distinctions made in other papers.
|
Table 1: Theory viewpoints |
A set of |
Who act |
Form a |
|
Actor-centric |
Actors |
As they choose |
Social Group |
|
Activity-centric |
Actors playing
Roles |
According to rules |
Social System |
Architects
are taught to start by understanding or defining the products and services that
are requested from or supplied by a system.
“For external users, only this
external functionality, together with non-functional aspects such as the
quality of service, costs etc., are relevant.” ArchiMate
standard v2.
TOGAF offers
these examples of services: check customer credit, provide weather data, and
consolidate drilling reports.
ArchiMate
offers these examples: Policy Creation, Premium Payment, Claim Registration,
and Claim.
Each of these
services can be defined by a contract that encapsulates (hides the logical flow
of) any process that delivers service.
The essential elements of a service contract are found formal logic, in a form called a Hoare Triple, the statement {P} S {Q} means:
· If the precondition (P) is true when the event is detected and process is performed,
· and the event-triggered process (S) terminates successfully,
· then the post condition (Q) is true afterwards.
Service-oriented (stimulus-response) specification of business systems is discussed in other papers.
It is important to note that architects need logical service contracts to be extended with measurable non-functional requirements.
A process is – by definition - a sequence of steps that leads to a result.
Processes are composable and decomposable. E.g.
|
Level of process |
Description |
Length of process |
|
Business process |
A long-running process that results in delivery of
business services or products |
days, weeks, months or even years |
|
System use case |
A shortish process, a use case
or transaction |
seconds, minutes or hours. |
|
Atomic process steps |
Atomic executable process step |
micro-seconds |
A question arises here that goes back decades, at least.
How long should a process (or use case, or service) be?
At what level of granularity should we define a process?
The common answer is: “define a process to meet a goal, deliver a value or result”.
But this is no help on its own, for two reasons
First, because processes, goals, values and results are equally composable and decomposable.
And at every level of granularity, a process terminates in a result of value to the process invoker.
Second because, employees often have to complete processes that (aside from earning their salaries) unrelated to the employees goals (bar salary).
A way to make the question more helpful is to fix the system boundary before you start.
Fixing the system boundary means you separate what is inside and outside the system.
Then, you can reasonably say the goals/values/results of interest are those the external entities want.
Also you can distinguish the outputs of a system from outcomes in the wider environment.
Our process (use case or service) terminates by delivering an output that an external entity uses in the wider environment to achieve an outcome.
|
Process |
Output |
Outcome |
|
A rail ticket purchase ends by |
giving you a ticket, which you later use to |
get to the destination. |
|
A haircut ends with you |
having shorter hair, which you hope, later, |
will make a better impression on
others. |
“Dynamics" are about what happens over time.
There are three different time-oriented views of system behavior.
Process
The repeated or repeatable system processes discussed above are performed by actors.
The life history of
an individual actor/component
Ashby observed that system theory is mostly about the plural (class or role) rather than the singular (object or actor).
Individual actors may enter and leave the system, their roles remains the same.
Boulding suggested that the unit of a social organisation might be the role rather than the individual.
He pointed out that individual actors/components each have a life history within the system.
An individual’s life history runs over time from their entry into the system to their departure.
Enterprise architecture is focused on roles rather than actors; life histories appear in “road maps” and “entity life cycles”.
The life history of a
whole system
A biological species evolves/adapts when individual organisms are replaced by somewhat different, better-adapted, organisms.
An enterprise evolves/adapts when individual business systems are replaced by somewhat different, better-adapted, business systems.
Biological evolution can be seen as a meta system that reshapes the behavior of individual organisms in a species.
Enterprise architecture can be seen as a meta system that reshapes the operational systems of a business.
E.g. IBM as a system today differs from IBM as a system yesterday.
The named entity has adapted to circumstances by replacing and radically changing the business systems it uses (and sells).
Footnote: Creative Commons Attribution-No Derivative Works Licence
2.0 21/02/2017
17:13
Attribution: You may copy, distribute and display this copyrighted
work only if you clearly credit “Avancier Limited: http://avancier.co.uk” before the start and include this
footnote at the end.
No Derivative Works: You may copy, distribute, display
only complete and verbatim copies of this page, not derivative works based
upon it.
For more information about the licence, see http://creativecommons.org