Towards a metamodel of business architecture …

Link: http://feedproxy.google.com/~r/Soapbox/~3/BxVdYp75Zi4/towards-metamodel-of-business.html

From Richard Veryard on Architecture

#bizarch #entarch I have been challenged to explain the constructs underlying my Six Viewpoints of Business Architecture (blogpost, draft book), so here is a rough schema. Comments please.

Viewpoint
Key Elements
Structure / Behaviour / Function
Purpose View Goal, Influence (Cause-Effect), Stakeholder, Outcome, Value Goals and outcomes are linked by chains of influence or production. A goal aims at an outcome, one outcome influences another (positively or negatively), therefore one goal supports or opposes another goal.

From the Business Motivation Model (BMM), I include here Ends, Means and Influencers, but not Directives (which I include in the Cybernetic View)

Some people like to arrange goals and objectives as a hierarchy (“Aim Hierarchy”). Alternatively, influence chains can be drawn as a network, using a notation such as i*.

Activity View Activity, Event (Condition/Trigger), Interaction,
Outcome, Role, Work
Activities are linked by process chains
(sometimes called value chains/streams)
Capability View Activity, Capability, Component, Value, Work Components perform activities according to
some abstract capabilities. Capabilities and components
are linked by chains of dependency.
Knowledge View Concept, Event (Fact) Concepts (represented by entity types or
classes) are linked by facts (represented by relationships
and attributes). Simple and compound facts are linked by chains of inference or derivation.
Responsibility View Agent/Role, Outcome, Responsibility, Service,
Stakeholder
Agent/Role is responsible (accountable) to a
stakeholder (or stakeholder proxy) for some outcome.
Responsibilities and services are linked by chains of delegation.

Among other things, the Responsibility View allows us to view the Principal/Agent problem (which applies at every level of the management structure).

Cybernetic View Component, Control, Event (Signal),
Regulation/Rule (Directive).
Components are governed by regulations/rules,
and are linked by control loops or feedback
loops
.

Some people like to arrange policies and rules as a hierarchy (“Directive Hierarchy”).

Some important notes.

Firstly, note that some elements appear in more than one viewpoint, but they
don’t look the same. 

For example, in my schema an event appears in the activity
view as a condition or trigger for one or more activity, in the knowledge view as a fact
(e.g. a piece of information reporting a change of state), and in
the cybernetic view as a signal (allowing control and regulation). The knowledge view defines the semantics of the event (what it means in conceptual terms); the activity view and the cybernetic view define the pragmatics of the event (what it means in behavioural terms).
Secondly, I am not ready to claim any universal applicability for this schema. I am aware of various paradigms that have particular (and conflicting) definitions for these terms, and I am not trying to accommodate every possible meaning of every term.

For example, I am aware that some paradigms have a slightly different notion of event. I am also aware that some paradigms regard a signal as an output (from a management information system) rather than as a control flow (into some management process).There are also conflicting views whether we should regard capabilities as abstract or concrete. (For this reason I have removed the word “abstract” to which David’s comment refers.)

Thirdly, each viewpoint allows for both hierarchical and network structures. Although my own preference is to draw networks rather than hierarchies, I am not dogmatic about this.

Fourthly, there are some things I don’t regard as part of the Business Architecture but as part of some other domain. These include Location, Requirement, Solution, System (including Human Activity System or SocioTechnical System), Transformation and Use Case. I shall try to justify these exclusions in a separate post.

And finally, I am aware of how much work would be required to turn
this into a rigorous ontology or metamodel. If anyone is interested in
funding this work, please send me a large cheque.