Missing / Teleporting units in PBEM

[DCWhitworth]

OK here are some thoughts that have occurred to me since Marshall mentioned that there are still unit vanishing / teleporting issues in PBEM that he hasn't been able to track down.

The fact that he says they *only* appear to be in PBEM suggests to me that most likely they are related to transfer of PBM files between players and possibly down to player error / misuse.

I know various odd effects can be achieved in PBEM games by not following the correct procedure, for instance -

1. You can load totally incorrect PBM files, even ones from another game by renaming them. OK why would you ? Well in one game at an early stage, one player sent out a file for the same move but in a different game. Most players renamed this file to what they thought it should be and it correctly loaded into the game, however it caused various odd things (like Turkish garrisons appearing in Austria). We were able to reverse and get out of this, but it shows odd effects can be achieved by odd actions.

2. I've seen this one several times. A backup goes out, a player plays his turn. Later he accidentally loads from the backup again, loads a few PBM files but comes to a shuddering halt when he can't load his own PBM file. So rather than ask for a backup, he replays his turn from memory (possibly inaccurately) then reloads later PBM files, which the game will let him do.

In these instances you have in case 1. A situation where the database can be messed up by importing invalid data and 2. A situation where one player's version of the game can differ from the other players in the game. I'm not implying these are the only way things can go wrong, just that they are definte instances I have seen,

Basically what I am saying is that I think if the validation of PBM files by the game is tightened up, some or perhaps all of the issues may disappear. The fact that there appears to be minimal validation of incoming PBM files is a potential weakness that needs to be plugged. It's the old IT proverb, garbage in, garbage out. The game *will* let garbage be imported into the game database, it needs some protection. There is already some, but not enough.

Thoughts ?

[DCWhitworth]

While I think about it I can recall another instance that could cause issues. Potential PBM file corruption in transit. I'm not sure that there is any internal checking or checksumming on the contents of a PBM file so you could potetially load a corrupted file.

In one PBEM game I was hosting, one player repeatedly was having problems with his game corrupting and needing to be sent backups. He swore he wasn't doing anything wrong. He mentioned eventually that he had an poor quality / unreliable internet link and he often had difficulty downloading the PBM files for the game.

It is now pretty much common practice in most games that players should zip their files before exchanging. Most people do this simply to save space but actually it has another beneficial side effect, zip programs checksum their files and if they get corrupted in transit then they won't open.

So as well as validation of *which* PBM file you are loading, I think the game needs validation of the contents, perhaps a simple hash total would probably do it.

[bresh]

Hi David, im not quite agreeing.
 
In my pbm game with Marshall as host, i moved Austrian forces into the minors i got control of when France Dowed them.
Fr-Au not at war. 1805, January.
I placed garrions on depots in those and moved corps there.
It happend in Genoa and Frankfurt.
Then after French move, the corps teleported random, one back to where he started, other to Luxenbourg through French controlled areas, while a Austian depot at Freiburg was just as close.
Worse happend to the garrisons, 1 factor ended from Frankfurt in Novrogod at Sk.Petersburg.  And the Genoa factor sailed to Corfu.
 
Marshall explained it with, its how is not allowing other MP's in the area when checking for control of a minor, who are not at war with the Owner. Something i think should be limited to the conquest month.
 
Beside the 2 dep-garrisons bugged jumps, game generated 2 new Luxenbourg-factors as garrisons in Transylvania.
 
Regards
Bresh
 

[pzgndr]

Basically what I am saying is that I think if the validation of PBM files by the game is tightened up, some or perhaps all of the issues may disappear.

May be some way to assign an internal "key" to a game when a host sets it up. Then all pbem files generated could have that key, and only files with that key could be loaded. That would at least resolve the potential problem of mismatched game files. Just a thought.

[DCWhitworth]

ORIGINAL: bresh

Hi David, im not quite agreeing.

In my pbm game with Marshall as host, i moved Austrian forces into the minors i got control of when France Dowed them.
Fr-Au not at war. 1805, January.
I placed garrions on depots in those and moved corps there.
It happend in Genoa and Frankfurt.
Then after French move, the corps teleported random, one back to where he started, other to Luxenbourg through French controlled areas, while a Austian depot at Freiburg was just as close.
Worse happend to the garrisons, 1 factor ended from Frankfurt in Novrogod at Sk.Petersburg.  And the Genoa factor sailed to Corfu.

Marshall explained it with, its how is not allowing other MP's in the area when checking for control of a minor, who are not at war with the Owner. Something i think should be limited to the conquest month.

Beside the 2 dep-garrisons bugged jumps, game generated 2 new Luxenbourg-factors as garrisons in Transylvania.

Regards
Bresh

I don't disagree with you Bresh, but also I wasn't saying that there aren't other reasons why units might move about, there very well might be. I was suggesting that game corruption could be introduced by careless operation of PBM files and that it might be wise to shut this off.

[DCWhitworth]

ORIGINAL: pzgndr

Basically what I am saying is that I think if the validation of PBM files by the game is tightened up, some or perhaps all of the issues may disappear.

May be some way to assign an internal "key" to a game when a host sets it up. Then all pbem files generated could have that key, and only files with that key could be loaded. That would at least resolve the potential problem of mismatched game files. Just a thought.

Also perhaps generate a unique number for each PBM file and the following PBM references this number. e.g. So a PBM file will effectively say 'I come after this PBM file' so you can't get someone putting a different PBM file in the sequence.

[bresh]

ORIGINAL: DCWhitworth



I don't disagree with you Bresh, but also I wasn't saying that there aren't other reasons why units might move about, there very well might be. I was suggesting that game corruption could be introduced by careless operation of PBM files and that it might be wise to shut this off.

All game files where zipped, so corruption is not always the reason.
But i agree zipping/arj/rar is a must, when players post unpacked files on googlegroup, i have to download them atleast 2 times, to compare bit size, and a 3rd time if these 2 dont match.

Regards
Bresh

[pzgndr]

Also, in anticipation of an editor, file exchanges should verify scenario data to ensure all players are using the same scenario. Believe it or not, in the early days of Strategic Command before this got fixed, some players would tamper with the scenario for their benefit without their opponent knowing. [:-] It may also be helpful to verify game version and provide a warning message if different versions are being used.

[MarshallEllis]

I've got the files tamper checked so we're good there.
I'm also verifying scenarios to make sure everybody is playing the same.
I have the version info BUT am not warning anybody yet. That sounds like a good idea!