Re: [Adonthell-devel] How client/server deals with maps

[Alexandre Courbot]

> Just quickly about this server/client stuff - DOn't you think the risk
> of players cheating is something we don't relly need to worry about.
> Since most people playing in one player mode will hve server and client
> and all the data means they could in theory look at everything anyway.
> Since it's open-source and all the stuff will be available for download
> in one form or another determined cheats will always find a way.
> Most normal players on the other hand probably will never bother looking
> at the code/maps etc anyway. Basically, what I'm saying is I wouldn't
> worry much about what makes cheating harder but rather what will work
> most efficiently.

Here is my opinion about cheating. I mostly agree with James here, there
are several points that must be taken into account:
-What importance do we give to cheating during single player game? I'd
say none. If the player wants to cheat, he can do. He will only waste
his pleasure and we can put all the mechanisms we want, we can't do
anything against it because the sources are available and both client
and server are running on the player's machine.

What we would rather want to avoid is cheating during multiplayer games,
i.e. we must prevent cheats that allows a player to do what others
can't. This is much more serious. Still, as James mentionned, it is
important to keep things efficient. What we discussed about (only
sending to the client what it is supposed to see) has all the
advantages: preventing cheating and reducing the amount of data sent by
the network. Kai and Lakin discussed about the door/key stuff: as Kai
said, doors can be "items" instead of mere mapobjects that the client
knows about. And even if they would be mapobjects, there are stuff that
would prevent the player to know the most important about it (he could
know there is a door, but how to open it is another run).

Of course, we haven't gone deep enough into the new mapengine to know
how exact this is, but the implementation seems viable, and anyway
you'll be able to see and test stuff in progress! :)

Alex.

-- 
http://www.gnurou.org