P1
State before events
Business reality is represented by state. Events are evidence of change, not the primary source of truth. A system that treats its event log as the truth has confused the record of what it observed with what is the case.
Layer 1
Foundational principles
Four claims about how business systems behave. Everything above depends on them, and a finding that contradicts one of them is a finding about the foundation rather than about an application. Each is a position that could be wrong, and the full statement of what would show it wrong is in Core Principles.
P2
Obligations before transactions
Commitments exist independently of payment activity. An obligation is created, modified, settled and terminated on its own lifecycle, and the transactions that satisfy it are a separate class of object with a separate lifecycle.
P3
Convergence over completeness
Systems can move toward correctness without complete information. Correctness is reached by mechanism rather than assumed at the point of input, so a system may be temporarily wrong and still be correctly designed, provided the convergence mechanism is defined.
P4
Observability is limited
What is observable and what is true are not necessarily the same. The gap between them is a design constraint to be modelled explicitly, not a data quality problem to be resolved upstream.
Layer 2
Formal models
The abstract machinery: states, time, obligations and convergence.
M1
State Space Theory
Finite states, transitions, boundaries and invariants. Supplies the MECE requirement and the notion of a structural guarantee preserved across transitions.
M2
Temporal State Theory
Due dates, ageing, deadlines and elapsed time. Time as a transition trigger rather than a timestamp attached to a record.
M3
Obligation Algebra
Creation, composition, decomposition, settlement and termination. The operations under which an obligation remains a well-formed obligation.
M4
Convergence Theory
State recovery, reconciliation and correction. The formal treatment of how a system that is temporarily wrong becomes right.
Layer 3
Computational models
How the formal models are computed under real observability.
C1
Snapshot Computation
Computing from observations rather than events. The balance-delta family of techniques and the conditions under which they are sound.
C2
Partial Observability
Computation under uncertainty and missing information, including what a system may legitimately conclude from an incomplete signal.
C3
Observational Computing
Observation-driven systems and evidence-based computation. A candidate discipline rather than a settled one.
Layer 4
Obligation systems
The first application domain, and the proving ground named in the strategic plan.
O1
Instalment Systems
Systems in which a single obligation is satisfied through a defined schedule of partial fulfilments, each with its own due condition.
O2
Arrangement Management
Systems governing the lifecycle of an agreement to pay, including its breach, its recovery and its renegotiation.
O3
Obligation Lifecycle Systems
Systems that own an obligation from creation to termination independently of the account or customer it attaches to.
Layer 5
Decision systems
Systems that take decisions over obligation state. The distinction that matters here is between a decision (a function of state, and therefore modellable) and an action, which is an effect and is not.
D1
Collections Systems
Systems that decide what to do about an obligation in arrears, treated as decision architecture rather than as workflow.
D2
Customer Commitment Systems
Systems that record, track and evaluate commitments made by a customer, including promises that are not yet obligations.
D3
Recovery Systems
Systems concerned with returning an obligation to a satisfiable state, and with measuring whether that has happened.
Layer 6
Enterprise systems
The model at organisational scope, where the object of design is no longer a single system but the relationship between many.
E1
Executable Architecture
Architecture expressed so that it can be executed and tested rather than only described. The enterprise expression of derivation integrity.
E2
Enterprise State Models
State treated as a strategic asset held at organisational level, with defined ownership and invariants across system boundaries.
E3
Operational Resilience
The capacity of a state model to survive outage, delay and partial failure without losing the ability to compute a correct outcome.
Layer 7
Research streams
Open positions the body of work is arguing for. A stream is not a topic; it is a claim that could turn out to be wrong.
R1
Obligation-Centric Systems
That obligation, not transaction or account, is the correct primary object for a class of business systems.
R2
Event Independence
That correct financial computation does not require a complete event history.
R3
Observational State Machines
That a state machine can be driven by observation rather than by event, and remain deterministic.
R4
Convergent Business Systems
That convergence is an architectural property that can be designed for and demonstrated.
R5
Enterprise State Theory
That enterprise state can be treated formally rather than as an emergent property of integration.
Layer 8
Canonical works
Four syntheses. A canonical work draws together published items and cannot precede them: none of these may be drafted until every item it depends on is published.
K1
From Events to Obligations
The shift from event-centric to obligation-centric architecture.
K2
Computation Without Complete Knowledge
The incomplete-information work drawn together as a single argument.
K3
Enterprise Systems as Formal Models
The state-based and enterprise work, synthesised.
K4
Toward a Science of Business Systems
The capstone. Draws on the three canonical works above.
Complex systems, made legible.
Where this fits
The map defines what the body of work is about. The mission explains why it exists. The topics on this site are where Layer 4 (obligation systems) is being worked out in practice.