Terminal inspector / lesson 07
Project execution without inventing structure
The viewer derives from the graph and completed trace instead of maintaining a second diagram.
Know what the primitive owns.
Problem
A separately maintained architecture diagram drifts from the code that actually executes.
System boundary
Terminal renderers return deterministic plain text. They do not write to stdout, add ANSI styling, mutate execution, or promote state fields into nodes.
State changes become visible.
Inspectable: Lifecycle boundaries, declared edges, trace state changes, evaluation, feedback, comparison, decoded paths, sensitivity, and useful flow.
Switch between graph and run views, Unicode and ASCII.
Static specimen / JavaScript adds controls without changing the lesson.
Same idea, honest surfaces.
import {
renderGraph,
renderRun,
} from "@composable-model-graph/terminal";
console.log(renderGraph(graph));
console.log(renderRun(graph, await graph.run({})));from composable_model_graph.terminal import (
render_graph,
render_run,
)
print(render_graph(graph))
print(render_run(graph, graph.run({})))What this does not imply.
Orientation and character set may change. Nodes, edges, fields, evaluation, and feedback meaning may not.