---
title: Block Structure (in Process Models)
type: concept
tags: [bpm, modelling, structuredness, soundness, gateways, design-by-construction]
sources:
  - "[[sources/2018-dumas-fundamentals-of-bpm]]"
  - "[[sources/2018-dumas-fundamentals-of-bpm-ch5-discovery]]"
  - "[[sources/2010-mendling-reijers-vanderaalst-7pmg]]"
  - "[[sources/1998-vanderaalst-verification-of-workflow-nets]]"
created: 2026-05-04
updated: 2026-05-04
---

# Block Structure

A **single-entry, single-exit (SESE) process-model fragment** whose entry and exit are gateways of the **same type** — one split, one matching join. Each path from the split leads to the matching join.

Definition ([[sources/2018-dumas-fundamentals-of-bpm]] §5.4.1):

> *A block structure is a single-entry single-exit process model fragment, such that the entry and exit points are two gateways (one split and one join) and each path from one gateway leads to the other gateway.*

Block-structured (a.k.a. **structured**) process models are nested compositions of such blocks plus sequential connections.

## Why block structure matters

### 1. Soundness by construction

If a model is built entirely from nested blocks of matching split-join type, it is **sound by construction** — no behavioural anomalies can arise within blocks. This is the easiest way to **achieve correctness without verification**.

This corresponds to [[sources/1998-vanderaalst-verification-of-workflow-nets]] §5 transformation rules T1–T4 (and T5 composites): each rule produces a structured fragment from a structured starting point, and Theorem 18 proves they preserve soundness.

### 2. Pragmatic quality

[[sources/2010-mendling-reijers-vanderaalst-7pmg]] **G4** ("model as structured as possible") is empirically supported: structured models are both **more understandable** and **less error-prone**.

[[sources/2018-dumas-fundamentals-of-bpm]] §5.4.3 (Pragmatic Quality) explicitly recommends block-structuring as a usability mechanism: *"a structured process model, besides being always sound, has been shown to often be easier to understand than its unstructured counterpart"* (Fig 5.15).

### 3. Tractable verification

Even when block-structuring isn't strictly enforced, structured models map cleanly onto **free-choice WF-nets**, for which [[concepts/soundness]] is decidable in polynomial time ([[sources/1998-vanderaalst-verification-of-workflow-nets]] Theorem 12).

## What block structure rules out

The four [[concepts/deadlock|behavioural anomalies]] addressed by Dumas §5.4.1 are anomalies that **cannot occur within** matched-type blocks:

| Anomaly | Mismatched block pattern that creates it |
|---|---|
| [[concepts/deadlock]] | XOR-split + AND-join |
| [[concepts/lack-of-synchronization]] | AND-split + XOR-join |
| [[concepts/livelock]] | Loop with mismatched entry/exit gateway types |
| [[concepts/dead-activity]] | Branch unreachable due to mismatched structure or unsatisfiable condition |

Inside a same-type block (AND-AND, XOR-XOR, OR-OR), none of these arise.

## Limited expressiveness

Block-structuring is **not free**. Some legal cyclic behaviours and arbitrary event-driven patterns cannot be expressed in block-structured form. [[sources/2018-dumas-fundamentals-of-bpm]] §4.1 (cycles) and §5.4.1 note the trade-off:

> *This type of model has limited expressiveness compared to unstructured models, as discussed in §4.1 in the context of cycles.*

For most administrative and transactional processes, block-structuring is sufficient. For knowledge-intensive, event-driven, or compliance-constraint-heavy processes, declarative models ([[frameworks/declare]]) or unstructured BPMN may be unavoidable — verify them with formal soundness checks instead.

## Construction patterns

The following patterns are block-structured and sound by construction:

- **Sequence**: A → B → C
- **XOR block**: XOR-split → {A | B | …} → XOR-join
- **AND block**: AND-split → {A ‖ B ‖ …} → AND-join
- **OR block**: OR-split → {active branches} → OR-join (with synchronising-merge semantics)
- **Structured loop**: XOR-merge ↔ {body} ↔ XOR-split (with explicit exit condition)
- **Nested blocks**: any pattern above with any of these as its branch content

Avoid:
- Goto-style cross-cut connections between branches.
- Joins matching splits of *different* types.
- Branches that re-enter the model from outside their block.
- Loops without explicit exit gateways.

## Block-structured vs free-choice

| Property | Block-structured | Free-choice WF-net |
|---|---|---|
| Definition | Nested matched-type blocks | Petri net with: ∀ p₁,p₂: •p₁ ∩ •p₂ = ∅ ∨ •p₁ = •p₂ |
| Soundness | By construction (no check needed) | Decidable in polynomial time |
| Expressiveness | Lower (some cyclic patterns excluded) | Higher (covers most BPMN with AND/XOR) |
| Practical applicability | Most administrative processes | Almost all WF-nets met in industry |

Free-choice is a *strictly larger* class than block-structured.

## Tool support

- **Camunda Modeler** — discourages but does not enforce block-structuring.
- **Inductive Miner** (process discovery) — produces block-structured models by construction.
- **Block-structured BPMN style guides** (e.g., Silver 2011 BPMN Method and Style — referenced-not-yet-ingested) — practitioner-level rules for block-structured BPMN.

## Related
[[concepts/soundness]] · [[concepts/workflow-net]] · [[concepts/7pmg]] (G4) · [[concepts/process-model-quality]] · [[concepts/deadlock]] · [[concepts/livelock]] · [[concepts/lack-of-synchronization]] · [[concepts/dead-activity]] · [[frameworks/bpmn]]

## Sources
[[sources/2018-dumas-fundamentals-of-bpm]] §5.4.1, §5.4.3, §4.1 · [[sources/2018-dumas-fundamentals-of-bpm-ch5-discovery]] · [[sources/2010-mendling-reijers-vanderaalst-7pmg]] (G4) · [[sources/1998-vanderaalst-verification-of-workflow-nets]] (transformation rules)