# Drawing the map of knowledge

- kind: problem
- status: open
- year: 1964

**Statement.** Is there a layout of a knowledge graph that keeps time, dependency and similarity faithful at every zoom level, and can it be computed efficiently?

Every map of knowledge chooses what its geometry means and gives up the rest. Garfield's 1964 citation historiographs put time on one axis; later maps of science place fields by citation similarity and lose chronology. Even one clean criterion is hard: minimizing the edge crossings of a drawing is NP-hard, shown by Garey and Johnson in 1983. This atlas has two drawings with different trade-offs: the Map keeps containment and loses time, and the Chart keeps time and field and loses similarity. A layout that keeps all three at every scale is an open problem.

## Sources

- [Wikipedia: Graph drawing](https://en.wikipedia.org/wiki/Graph_drawing)
- [Wikipedia: Crossing number (graph theory)](https://en.wikipedia.org/wiki/Crossing_number_(graph_theory))
- [Wikipedia: Citation analysis](https://en.wikipedia.org/wiki/Citation_analysis)

[Algorithms](/wiki/Space:algorithms.md) · [Computation](/wiki/Continent:computation.md)

---
Revision 69, 2026-09-30 05:09 UTC, by atlas import (import, claude-opus-5-5, Claude Code 2.1.280): Imported from content/02-computation.json, curated for the atlas by Claude Opus 5.5 at the direction of Danielle Fong (September 2026).. History: /wiki/Problem:map-of-knowledge?action=history. Edit: POST /api/wiki/page/Problem:map-of-knowledge {source, summary, base_rev: 69}.