Dynamic Orchestration Graph Language is a library-first platform for describing orchestration definitions in plain text.
The current architectural direction is:
- a human-readable
.doglsource format; - a syntax-facing language front end with source fidelity and diagnostics;
- a BPMN-aligned semantic model;
- validation, machine-readable interchange, and BPMN adapters around a stable core;
- embeddable language, render, and editor surfaces as the platform grows.
The project is in early development (0.0.0).
The current priority is architectural convergence:
- repository and crate structure;
- syntax and semantic boundaries;
- BPMN-aligned AST and semantic-domain structures;
- stable contracts for later parsing, validation, rendering, and import/export work.
Runtime, execution, hyperautomation, and broad integration concerns remain future extension areas rather than the current architectural center.
At the surface-syntax level, a simple DOGL file can look like this:
collab HelloProcess
() Start
=> Review
[] Review
=> End
() End
This is syntax, not the canonical semantic model.
Inside the platform, syntax-facing structures are lowered into a BPMN-aligned semantic shape built around concepts such as:
DoglFileCollaborationParticipantProcessFlowNodeSequenceFlowMessageFlow
See notation/DSL_syntax.md for the notation guide.
This README describes the current platform direction, layering, and BPMN-aligned semantic concepts at a repository level.
- Readable source: plain-text process definitions that are easier to review and diff than large XML artifacts.
- Library-first architecture: the primary product is an embeddable platform, not a monolithic application.
- Clear semantic layering: syntax-facing structures, semantic lowering, validation, and adapters are treated as separate concerns.
- BPMN-aligned core: the semantic model uses BPMN-valid concepts where BPMN provides the right meaning.
- Extensible platform direction: rendering, editor capabilities, bindings, import/export, and future runtime work can grow around the same core.
The intended project structure is converging toward:
dogl-languagefor the embeddable language core;dogl-renderfor the embeddable render core;dogl-editorfor the embeddable editor core;dogl-adaptersfor BPMN and other external integrations;- additional delivery surfaces such as CLI, WASM, and bindings.
Not every host application will need every part. The architecture is designed so consumers can embed only the layers they need.
DOGL may eventually support broader workflow and automation scenarios, but the current documentation and implementation work should be read through this rule:
- the authoritative current scope is the library platform and its semantic foundations;
- future runtime, orchestration, and hyperautomation capabilities are extension tracks, not proof that those parts are already implemented.
Dual-licensed under MIT or Apache-2.0 at your option.
