Skip to content

Diagrams

This section contains published visual interpretations and planned models for Because Drift Happens. Diagrams should clarify doctrine, vocabulary, and architecture patterns without turning the framework into a product diagram set.

The canonical doctrine graphics should use the language in Doctrine, especially autonomous drift as the master term and agentic drift as an AI-agent subterm.

Canonical Doctrine Views

The current doctrine uses three visual views:

  • Scope view: nested concentric layers showing federated drift containing autonomous drift, and autonomous drift containing drift.
  • Escalation view: a vertical stack showing drift escalating from local deviation to autonomous divergence to federated systemic risk.
  • Governance view: a paired response model showing awareness, alignment, and coordination as the capabilities that reveal, reduce, and contain drift.

Current Doctrine Graphics

These generated graphics are published working interpretations of the current doctrine, including graphics used on the homepage and in social previews. The six-line doctrine remains the reference; publication does not make every visual choice final. These raster assets are not editable diagram source files.

The Drift Doctrine

The Drift Doctrine reference view

The Drift Doctrine social graphic

Expected Diagram Types

Future diagrams may include:

  • Drift paths across intent, context, execution, and outcome.
  • Control-plane relationships above agents and workflows.
  • Governance checkpoint flow.
  • Veto, hold, escalation, and rollback paths.
  • Coherence boundaries in federated systems.
  • Audit trail spine across tools, agents, and state changes.

Diagram Guidelines

Diagrams should:

  • Use plain labels.
  • Use Autonomous Drift in primary doctrine graphics.
  • Use Agentic Drift only when the graphic is specifically about AI agents or intentionally uses the term as a subterm or audience hook.
  • Use the tagline Small deviations. Amplified. Compounded. where a tagline is needed.
  • Avoid vendor-specific architecture unless clearly marked as an example.
  • Show authority, feedback, and intervention paths where relevant.
  • Avoid implying that the framework is complete or tied to one implementation.
  • Be understandable without a long explanation.

See Planned Diagrams for the diagram backlog. Planned models are separate from the published graphics above.