> This would make it easy for players to predict who would be the tank > and thus make things less interesting from the angle of game play. The way > aggr_update() polls now, it randomly chooses who the attacked player will be > by polling and then going through the whole list of players in the room.
That's not such a big deal.. just keep a list of rooms that people move into during the tick, and do an aggression check on them at the end of the tick (after everyone following has moved into the room), targeting a random person in it. Unless you have an event queue, then it's even easier. > We're talking about legacy code which was load tested back on a 386 > with like 8 megs of ram. I think that optimizing that section of the code > when we're dealing with P2-P4 generation processors is moot. Optimization is never moot. Damn kids with their fast machines... :) --Palrich.
