I wouldn't bother with unconnected nodes - for me it's simple, just do it bish-bang-bosh stylie - the only nodes you need are on walkable areas, therefore there are no unconnected nodes - and if you want to get technical and say there is a doorway between them, I'll still compute them as connected - coz the door may as well be a large piece of cheese or a badger - my nodes are still connected regardless of anything.
I don't use A* and I certainly wouldn't stick a node in a wall or any other impassable object. The only nodes I'd use are the ones you can walk on, then I don't have any unconnected ones.
Bish-bang-bosh, "puki-stylie".
I still don't get it though (the unconnected node bit) - surely a wall is static, so who cares, surely you can calculate most of them (if not all) as a 'look-up' external to any path-finding - on the basis that they are static you can precalc them - and secondly, surely this is map dependent - surely this relies on the data, not on the pathfinding algo?
I don't use A* and I certainly wouldn't stick a node in a wall or any other impassable object. The only nodes I'd use are the ones you can walk on, then I don't have any unconnected ones.
Bish-bang-bosh, "puki-stylie".
I still don't get it though (the unconnected node bit) - surely a wall is static, so who cares, surely you can calculate most of them (if not all) as a 'look-up' external to any path-finding - on the basis that they are static you can precalc them - and secondly, surely this is map dependent - surely this relies on the data, not on the pathfinding algo?