|
Shannon V. OKeets -> RE: Comparision (8/8/2008 6:46:00 PM)
|
quote:
ORIGINAL: SewerStarFish
quote:
ORIGINAL: Shannon V. OKeets
The message transfers are going to be short strings most of the time. The only place where we have been worrying about needing to send a lot of data at once is for players who disconnect from a game in progress (or arrive late to a session) and need to be brought up to date on what has transpired in their absence. I still think this is not a lot of data (even at 9600 baud).
This is not so far fetched a scenario as might first be supposed, since the NetPlay design enables one side to play while the other isn't present. For instance, if it is time for the Axis to make all their land moves in the middle of the summer of 1942, the Allied players might just say "see you next time", instead of waiting for the hour or so for all those decisions to be made. When the game resumes (a week later?), the Allied players would receive Game Record Logs containing all the Axis players' moves.
Great point, connectivity could be an issue. I'd hate hours of game play to be lost because a player lost his connection for whatever reason. Heck even losing a big chunk of a turn could change quite a few things.
This brings to mind a question: Civilization IV has an option that "locks" the random numbers so that a player playing the AI can not restart and replay an event to get a better combat result; how will MWiF's random number generator behave on multiple loads of the same situation. I ask, because if it is locked that would require a player to follow the exact same sequence should some event crash a net play game and the players need to restart from a previous load. I presume the PBM will prevent knowledge of the result of a roll until the other player has seen it --so that shouldn't be an issue there. And who really cares if a player chooses to cheat the AI on a solitaire game. [:-]
It's just computers can be quirky, they crash [perhaps as a programmer you didn't know [;)]] and thus there will be problems with continuity when one goes BSOD during a game.
The program assigns the Team Leader of the Allied side as Master MWIF. Usually this is the Commonwealth, unless the CW is not in the game. In reality the players aren't really aware of this; it is just how the program knows which computer rolls the dice (i.e., draws random numbers). When the sequence of play requires the program to affect the simulated game world (e.g., draw a random number, advance to the next phase of the game), only the Master MWIF computer does so. It then passes that information to all the other computers/players as Game Record Log events.
I don't know of any way to circumvent the player just going back to a previously saved game without involving a second computer. So, before Master MWIF will do anything that affects the simulated world, the Team Leader for the Axis has to be connected. This way, when the GRL event occurs it can be immediately transmitted to the players on the other side. Then the Team Leaders' computers transmit the GRLs to the other players in the game.
To get back to the example I gave in a previous post, if the Axis players are making their moves while the Allied players are off-line, then all the Axis moves are sent as GRL entries to the Axis Team Leader, who both accumulates them and distributres them to the players on his side. That way allies can coordinate what they are doing. But the Axis player can not "advance to the next phase" until the Allied Team Leader is on-line. Once the Team Leader computers are connected, the Axis Team Leader's computer will transmit all the Axis moves to Master MWIF (Allied Team Leader), including his command to "end the phase". At that point Master MWIF will advance to the next phase.
Clear as mud?
|
|
|
|
