Layers and Their Questions

Layers make a large model understandable by separating engineering questions. They are not sequential project phases.

Layer	Question	Typical elements
Context	Who uses the device, where, and with what external systems?	Actors, intended use, use context, use errors
Operational	What work and outcomes matter to stakeholders?	Activities, capabilities, scenarios
System analysis	What must the system accomplish across a scenario?	Capabilities, functional chains, functions
Requirements	What measurable claims must be true?	Needs, system/software/hardware requirements
Logical	Which implementation-neutral responsibilities exist?	Logical components, functions, exchanges
Software/hardware/physical	What realizes the logical design?	Software items, firmware, processing nodes, assemblies, ports
Risk and cybersecurity	What could cause harm or compromise, and how is it controlled?	Hazards, threats, vulnerabilities, controls
Assurance	What test and evidence supports each claim?	Verification cases, validation cases, evidence

How to use layers in Architect

1. Begin with a review question.
2. Choose the viewpoint that includes the relevant layers.
3. Select an element and follow its typed relationships across layers.
4. Switch to a neighboring viewpoint only when the question changes.

For example, a reviewer investigating over-infusion might move from:

flowchart LR
    Hazard[Risk layer:<br/>over-infusion] --> Control[Risk layer:<br/>independent monitor]
    Control --> Requirement[Requirements layer:<br/>detect excessive flow]
    Requirement --> Function[Functional layer:<br/>monitor flow]
    Function --> Software[Software layer:<br/>safety monitor]
    Requirement --> Test[Assurance layer:<br/>shutoff verification]

Organizing files

Folders may mirror layers, but the semantic kind—not the folder—determines where an element belongs. Use files to support ownership and review; use MEMO kinds and relationships to preserve meaning.