[TUHS] RIP Claude Shannon

Doug McIlroy doug at cs.dartmouth.edu
Sun Feb 25 23:16:36 AEST 2018

> But a note on Dijkstra's algorithm: Moore and Dijsktra both published
> in 1959.

I was off by one on the year, but the sign of the error is debatable.

Moore's paper was presented in a conference held in early April, 1957,
proceedings from which were not issued until 1959. I learned about it
from Moore when I first I met him, in 1958. Then, he described the
algorithm in vivid, instantly understandable terms: imagine a flood
spreading at uniform speed through the network and record the
distance to nodes in order of wetting.

> But it is documented Dijkstra's algorithm has been invented and used
> by him in 1956.

Taking into account the lead time for conference submissions, one
can confidently say that Moore devised the algorithm before 1957.
I do not know, though, when it first ran on a Bell Labs computer.

That said, Moore's paper, which presented the algorithm essentially
by example, was not nearly as clear as the capsule summary he gave
me. It seems amateurish by comparison with Dijkstra's elegant treatment.
Dijkstra's name has been attached to the method with good reason.


More information about the TUHS mailing list