0011 — Mind maps and flowcharts are separate tools
Decision
Both document the work, but they answer different questions, so they are two tools with their own operations, not one tool with a type switch:
| Mind map | Flowchart | |
|---|---|---|
| Answers | what is known about a topic: a map | how something flows from one point to another |
| Shape | a tree around one topic, with cross links | a directed graph from start to end |
| Nodes | topic, goal, decision, fact, hypothesis, question, task, idea, dead end, note; open / done / dropped | ISO 5807 symbols: start, end, process, decision, input/output, subprocess, document, data, manual input, preparation, connector, delay, note; logic gates AND, OR, NOT, NAND, NOR, XOR, XNOR |
| Edges | parent → child branches; labelled cross links ("contradicts") | arrows with conditions ("yes", "no"), loops allowed |
| Typical use | working memory of a task, research, the project's architecture | a request path, an algorithm, a release procedure, the logic of a check |
| Operations | mindmap_new, mindmap_add --parent, mindmap_link, mindmap_flow <node|branch|link> <flowchart>, mindmap_done / drop, mindmap_show | flowchart_new --direction down|right, flowchart_add --after <node> --label <cond>, flowchart_link, flowchart_show |
A mind map can point at flowcharts. Any node, branch or cross link may link to one or more flowcharts: the map says what, the linked flowchart says how that part works. The link is shown as ▸ flow #N on the node, and under the tree for a branch or a link.
Both tools:
- live in
Memory(IF-7); - point their nodes at journal entries;
- are drawn by frontends in text and as SVG (journal#140);
- show in the transcript as their own events (MINDMAP, FLOWCHART).
Why
A flowchart forced into a mind map loses its direction and its symbols' meaning. A mind map forced into a flowchart becomes a false process.
Separate tools let the agent pick by the question it is answering. They also let each tool validate its own structure:
- a flowchart has a start, every arrow leads somewhere, and a decision has labelled exits;
- a mind map is a tree.
Context
The memory backend carried over has flowcharts as a --type of its mind maps: a --type flowchart on its command for a new map, ISO 5807 symbols and gates, direction down or right.
The harness keeps that model's content and splits the interface. The backend adapter maps the two tools onto it until the backend splits too.
Rejected
- One
maptool with a type flag (the backend's current shape): the agent has to know the switch exists and the operations mix (--parentmeans nothing in a flowchart,--afternothing in a mind map).
Roles — clarified by the owner, 2026-11-01
-
A mind map is a document: the map of the project's decisions, questions, constraints, invariants and specifications. When a developer asks why the system is made this way, the map of decisions, constraints and specifications must answer it.
-
A flowchart is both a document of a flow and the live view of step-debug mode (ADR 0013): the chart a pipeline's run walks through, step by step. That is what sets it apart from a mind map.
-
Both render to SVG files (not built yet). A map or chart can be saved and shared as a picture outside the harness, as the carried-over memory backend already does. It is a good practice, and the harness keeps it.