[ODE] Networking

[ruben-gerlach@web.de]

Hi,

thank you. The problem is, how can I allow the player to interact with the environment if its only like a replay? In this approach simulating the scene on the client side would be problematic. I already tried something similar and it looks sort of crappy.. Imagine 2 boxes that have to go trough each other because they are on opposite positions on the server, they will behave chaotic when intersecting. If I dont simulate I have no possibility to move the player, have I?

Maybe there is anyone else with experiences in this field?

Regards,
Ruben


> -----Ursprüngliche Nachricht-----
> Von: "Bruno Sanches" <bcsanches at gmail.com>
> Gesendet: 01.11.07 21:03:19
> An: "Ruben Gerlach" <ruben-gerlach at web.de>
> Betreff: Re: [ODE] Networking


> I worked on a game two years ago where we created a replay system (I know, you are talking about network), but I believe the issue is similar.
> 
>  
> 
> What we did is that we recorded all physics activity (storing positions and quaternions transformations) for each frame (I dont remeber if it was for every frame, or every 100 ms). And when playing the replay we just updated the objects interpolating this data. It needed some extra info like sound when collisions ocurred, this was included on the data too (not at the same rate).
> 
>  
> 
> This worked pretty well for us, maybe you can use a aproach like this on network.
> 
>  
> 
> Regards
> 
> Bruno Sanches
>  
> 
> On 11/1/07, Ruben Gerlach <ruben-gerlach at web.de> wrote:
> Hey everyone,
> 
> I just subscribed to the mailing list, hope this will give some new input :)
> I use Ogre and the ode wrapper OgreOde for a 3rd person beat 'em up (no beating yet ;)) I've read a lot about lag compensation, client prediction, dead reckoning, cubic splines etc. in the last days, but still I have no clue how to synchronize my ode scene over the network. I face the following problems:
> 
> 
> -How to predict behavior of specific objects, e.g. if i know my object was at a given position at a given time I should move it forward regarding the velocity and the passed time so that its synchronized with the server, but what to do if it collides on that way? In general, should I try to have exact the same scene on client and server using prediction or should I accept a time gap on the clients side, making things easier to handle but maybe worse looking/reacting to the player.?
> 
> 
> -How to move the clients player on server and client in sync, so that they do not end up on different positions (thats what happens now due to differences in the scene)? If I send the absolute position of the player to the server, not a "now moving forward" signal, I will have different moving simulation of the players on client and server apps, ending up in totally different behavior of e.g. boxes that cross the players way, besides this would allow no-clipping cheats, etc.
> 
> - Should I simulate the scene with ode on the client side or just use target positions sent by the server? If so, how to realize client movement?
> 
> 
> I would be interested in any experiences in this field, even if not matching my problems.
> 
> Best regards,
> Ruben
> 
> PS: I'm from Berlin, Germany, so please excuse any language barriers.
> ______________________________________________________________________________
> 
> Jetzt neu! Im riesigen WEB.DE Club SmartDrive Dateien freigeben und mit
> Freunden teilen! http://www.freemail.web.de/club/smartdrive_ttc.htm/?mc=021134
> 
> 
> _______________________________________________
> ODE mailing list
> ODE at ode.org
> http://ode.org/mailman/listinfo/ode
> 
> 


______________________________________________________________________________
Jetzt neu! Im riesigen WEB.DE Club SmartDrive Dateien freigeben und mit 
Freunden teilen! http://www.freemail.web.de/club/smartdrive_ttc.htm/?mc=021134