Maps and Discovery

The knowledge graph

Content is a hierarchy plus a graph: worlds → topics → boxes, with each box naming the boxes it builds on. Validation guarantees the graph has no cycles.

flowchart TB
  W[World: Graphs] --> T1[Topic: Traversal]
  W --> T2[Topic: Shortest paths]
  T1 --> B1[BFS]
  T1 --> B2[DFS]
  T2 --> B3[Dijkstra]
  T2 --> B4[Bellman-Ford]
  B1 -. builds on .-> B3
  B3 -. builds on .-> B4

Illustrative example of the structure.

Maps

A topic map lays out a topic's boxes by depth in the build-on graph, coloured by the learner's progress: not opened yet, started or done, with markers for kept boxes and boxes due for review. A world map shows its topics with mastery. Maps are drawn on the device from the cached taxonomy and local history, so they work offline and update live as the learner keeps or reviews boxes.

Opening a box whose prerequisites aren't met shows a "This builds on" sheet once, as the box opens (never mid-read), linking the earlier box with an option to read on anyway.

Search and explore

• Search runs on the device over the cached catalog. Queries are logged, scrubbed, to search_quality and content_requests so missing content shows up in analytics.
• Explore lists worlds and topics; worlds not open yet show a waitlist button whose sign-ups go to world_waitlist, which helps decide what to write next.

Interests

Onboarding asks for a few interests. The behaviour-interests job adds interests learned from behaviour, and the planner uses both, with stated interests weighted higher at first.

${sch.nav()}