Skip to main content

Online edition for BLOGE 0.9.8-RC1 · facts verified 2026-09-15 · 中文

Foreword {-}

Head First BLOGE is a guided learning track for the BLOGE orchestration engine. It starts with graphs, nodes, dependencies, and data flow, then works toward durability, observability, and production wiring through real examples drawn from this repository.

BLOGE in one sentence: BLOGE is a Java-first orchestration engine: you describe business steps as a graph, and the engine takes care of ordering, parallelism, retry, timeout, branching, and recovery.

BLOGE ships as a small core plus opt-in extensions: bloge-agent-ext (LLM-driven agent orchestration with tool dispatch), bloge-session-ext + bloge-session-durable (crash-safe multi-turn sessions), and bloge-state-ext + bloge-state-durable (crash-safe state machines). These are covered in later chapters.

Three core properties

  • Visible workflow — the shape of the process is explicit instead of hidden across service methods.
  • Dependency-driven execution — you declare who depends on whom; the engine decides what can run now.
  • Built-in resilience — retry, timeout, fallback, waiting, and durable progress are part of the model, not hand-written glue code.

Who this book is for

  • Primary audience — Java developers new to orchestration who want to understand how BLOGE models, executes, and recovers business workflows.
  • Secondary audience — developers evaluating BLOGE against Temporal, Camunda, Spring Batch, or hand-written orchestration.

How to read this book

  1. Go in order. Each chapter builds on the previous one.
  2. Run the examples. The code comes from real files in this repository — open them, execute them, break them.
  3. Use the reference docs for exact syntax. This series teaches how to think; the specs define what is legal.

The seven learning arcs

  • Arc 1 — Introductory: build the mental model for graphs, nodes, dependencies, and flowing data.
  • Arc 2 — Maintainable authoring: turn DSL changes into readable, testable, reviewable engineering work.
  • Arc 3 — Composition and time: practice reuse, iteration, suspension, and crash recovery.
  • Arc 4 — Lifecycle models: model sessions, state machines, nested ownership, and bounded agent loops.
  • Arc 5 — Technical confidence: match tests, telemetry, Spring wiring, and capacity evidence to precise claims.
  • Arc 6 — Business promises: verify owner-approved contracts with controlled inputs, source-bound evidence, governance, and attribution.
  • Arc 7 — Selection and capstone: decide when BLOGE fits, then build the smallest justified graduation project.