Declarative Process Modelling

A modelling paradigm that specifies a process by constraints on allowed behaviour rather than an explicit control-flow graph. Any execution that does not violate the constraints is legal; the control-flow is implicit (sources/2008-pesic-declare-manual, §1.1).

Imperative vs declarative

Dimension	Imperative (e.g. frameworks/bpmn, Petri nets, YAWL)	Declarative (e.g. frameworks/declare, ConDec, DecSerFlow)
Specification focus	HOW (ordered flow, gateways, begin/end)	WHAT (rules that must hold)
Default semantics	Only what is drawn is allowed	Everything not forbidden is allowed
Flexibility	Low — deviations = exceptions	High — broad permitted space
Best fit	Highly-structured, repetitive processes	Loosely-structured, knowledge-intensive, variable processes
Formal basis	Graph/token semantics	Temporal logic (LTL / LTLf / LDLf)

Why declarative

Imperative models over-specify flexible processes, forcing frequent deviations from the prescribed flow (sources/2008-pesic-declare-manual, §1.1). Declarative models encode only the genuine regulatory, compliance, or causal constraints, leaving the rest of the behaviour space open to the participants.

Formal grounding

Linear Temporal Logic (LTL) — original DECLARE semantics.
LTL over finite traces (LTLf) and Linear Dynamic Logic on finite traces (LDLf) — De Giacomo & Vardi, now standard for trace-finite process semantics.
Constraints compile to finite-state automata for verification (satisfiability, dead activities, conflicts) and for run-time monitoring.

Run-time constraint monitoring

Three-valued state commonly used (DECLARE traffic-light):

green — satisfied.
orange — temporarily violated, still repairable.
red — permanently violated.

Constraint typing: mandatory · optional · conditional.

Declarative process discovery

Automated mining of declarative models from event logs: Declare Miner (Maggi et al.), MINERful, DisCoveR. Contrast with imperative discovery (α-algorithm, Inductive Miner) — see methods/process-mining-basics and concepts/process-discovery.

Data-aware extensions

MP-Declare / data-aware Declare — constraints conditioned on data predicates.
Declarative data-centric models (e.g. DCDS, Montali) — integrate temporal constraints with persistent data.

frameworks/declare — canonical implementation (ConDec, DecSerFlow).
frameworks/cmmn — OMG case-management notation; shares declarative flavour but is GUI/marker-based rather than LTL-grounded.
frameworks/dmn — declarative decision logic; complementary, not a control-flow competitor.

Hub entities

entities/wil-van-der-aalst · entities/fabrizio-maggi · entities/marco-montali · Maja Pesic · Helen Schonenberg.