[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Game idea....





Mads Bondo Dydensborg wrote:

> One poster pointed out multiplayer issues. A single player that has no
> morale and gets a kick out of cheating, can completly ruin a dm game of
> Quake with 16 people in it. In Denmark, and particulary Sweden, these
> players exists. The real trouble here, is that they use bots/proxies, that
> is inserted between the client and the server, like this:
> quake client <--> bot/proxy <--> quake server

Yep, and the problem is that more and more games are moving to the
VM model (at least the big titles), so it looks like we'll end up with
the flexibility of the DLL/Proxy model, but with the ease of
"hack a few lines" QuakeC...

> And so on. I believe it is -really- hard to solve this effeciently. Unless
> you start encrypting the stream, someone will reverse engineer your
> network protocol, and there you go. Even if you encrypt it, you could
> probably have problems.

There has been a discussion about this on the Q2Java list, and somebody
said that the major reason the network protocol has been reversed,
is that people want to make demo editors.  So a trivial start of a solution
could be to decouple the way a demo is recorded from what's sent
between client and network.  But ofcourse the whole point of proper
C/S separation is that demo recoding becomes as trivial as recording
the packets and playing them back with a S stub..  But maybe they're
right, if the demo specs are available for nothing and have zero to do
with the network transmissions, there will be less incentive among the
"white" programmers to reverse netcode, and the "black" guys are
usually not smart enough for that ;)  (most botters/proxiers are just
the Q2 script kiddies)
  Good point though when it comes to OSS...  I have no idea  what
you can do about somebody who downloads the client source,
programs autoaiming in it, and recompiles...  CRC checks would
hamper platform independence (every binary would be different),
obfuscation sucks, and any protocol C/S negotiate to prove their
"realness" to eachother is futile with source available...

> Oh, and in case you did not know, let me tell you of a few client side
> model modifications that have been used troughout time;

Yeah, this is a problem; I know some people play with spikes
even in official wars...  No more sneaking around the corner in a 1:1.

Bert
--

-=<Short Controlled Bursts>=-