Other/Dijkshtra.java has some problems:
- Hard-coded magic numbers (1000000l)
- initialization of a very large 2-dimensional array when it is not needed when using another implementation. - No description was given on the input format the algorithm was expecting.
- Poor commenting.
- Name of the algorithm is wrong. Should be "Dijkstra"
I have created a new pull request #559 that has new code for the algorithm. It has none of the above mentioned problems from what I can tell.
If the previously mentioned pull gets accepted, then I would recommend that the old version "Dijkshtra.java" should be removed.