Talk:I'm a software engineer, get me out of here: Difference between revisions
Talk:I'm a software engineer, get me out of here (view source)
Revision as of 13:49, 30 August 2018
, 5 years agono edit summary
No edit summary |
No edit summary |
||
Line 5:
::Reports of my demise have been greatly exaggerated, the firing squad were all cross-eyed and missed. Perhaps some sort of proof that using those specific algorithms are better in this case than some other alternative should be shown (in the output, dunno what though).
::The edges are not weighted (they all cost 1 day), so Dijkstra is an unnecessary overhead compared to breadth-first, albeit the relatively small one of maintaining and retrieving the lowest-cost node next (in my code the "next" variable only ever holds {cost}{cost+1}). Perhaps if the numbers on the map, instead of some magical "teleport" number were a terrain difficulty, with 1 being "straight fast motorways" and 9 being "dense undergrowth, steep difficult climbs, boggy marshes, minefields, and similar obstacles", so to move 1 square costs (this+dest) days travel, then that might justify using Dijkstra. It would of course mean there are no unreachable cells.
::As it stands, Part 2 can be completed with a simple breadth-first search (or two), calculating at most 820 routes, whereas using Floyd-Warshall creates 170,156 routes.
| |||