Dijkstra number

A Dijkstra number is the collaborative distance between a computer scientist and Edsger Dijkstra in the computer science collaboration graph, where edges connect researchers who have co-authored papers. Dijkstra has a Dijkstra number of 0; his direct co-authors have 1; their co-authors (excluding Dijkstra himself) have 2; and so on.

The concept is a direct analog of the Erdős number in mathematics and the Bacon number in film, extending the measurement of small-world collaborative distance to the domain of computer science. It reveals that the same structural properties — short average path lengths, high clustering, and heavy-tailed degree distributions — characterize academic collaboration networks regardless of discipline.

Origins and Structure

Edsger Dijkstra (1930–2002) was a Dutch computer scientist whose influence on the field was both technical and sociological. He authored foundational papers on graph algorithms, concurrent programming, and formal verification, and collaborated extensively with researchers across Europe and North America. Like Paul Erdős in mathematics, Dijkstra's wide-ranging collaborations created a hub in the computer science network that naturally lends itself to distance measurement.

The Dijkstra number distribution mirrors the Erdős number distribution: most published computer scientists have a Dijkstra number of 4 or less, and the median is approximately 5. This similarity is not coincidental. Both networks are small-world networks generated by the same social process — researchers attend conferences, share graduate students, and form cliques around research areas — and thus converge on the same structural statistics despite operating in entirely different intellectual domains.

Generalization and Network Universality

The Dijkstra number, alongside the Erdős number and Bacon number, demonstrates that collaboration distance is a universal property of creative networks. Whether the nodes are mathematicians, computer scientists, or actors, and whether the edges represent co-authored theorems, co-designed algorithms, or co-starred scenes, the resulting graphs share the same network science signature.

This universality suggests that the structure of creative collaboration is not domain-specific but is instead shaped by universal constraints: the cost of maintaining relationships, the benefit of interdisciplinary connection, and the tendency for successful researchers to attract more collaborators. The preferential attachment dynamics that produce heavy-tailed degree distributions in scientific collaboration networks are the same dynamics that produce hub-and-spoke structures in film credits and music co-production.

The Dijkstra number is often treated as a parlor game for computer scientists, but it is better understood as evidence that the topology of human creative networks is invariant across domains. The fact that mathematicians, computer scientists, and actors all live in small-worlds with the same degree distribution and path length statistics is not a curiosity — it is a theorem about the geometry of collaboration itself. The network does not care what the nodes create. It only cares that they connect.