Removing Netcode

#1 You could replace it with your own ... although Torque has some of the best netcode ever and it's completely intertwined with the engine. If you want to make a networked game Torque probably already does stuff better than just about anything around. If you've got some really cool new features perhaps you could just integrate them into Torque's current algorithms though.

#2 Well, it's more a whole rework. It doesn't work on a server/client system. I'm working on a very specific project with some fantastical netcode ideas, but it requires a whole redisgn of network. It's more a p2p sytem then server/client. Do you have any recommendations as I really /have/ to use the netcode that I have.
Thanks alot for you reply :D

#3 The Torque Game Engine is entirely built around the netcode and the cleint server architechture. If you have the source code you can change it in anyway you want. Although I have to think that after the modifications required to remove and replace all the netcode as well as the client server aspect of the system and you will have pretty much built your own game engine.

It can be done yes, I don't know if anyone here will be able to say if they can make a case for it being worth your time and effort.

#4 It would be a very bad idea to rip out the Torque net code and put in your own. The Torque net code is so efficient, and insanely well designed that you can't really beat it. No matter what kind game you are making, its the optimal solution. You might have to tweak the settings a little, or adjust the scoping code to get something optimal for your game. But you shouldn't need to replace it.

It would be like buying a Porshe and putting a lawn mower motor in it.

#5 Unless Torque supports p2p, then I can't keep its netcode :(.

#6 So you don't need any form of security to prevent cheating? And you don't mind uses having major problems, with closed ports (Windows XP has the firewall enabled by default).

#7 Actually, Torque's netcode isn't very far off from being p2p, depending on what you're doing. If you just want to do distributed message routing, you could probably get an implementation up (using NetEvents) in about an hour. If you want ghosting to work in a distributed way (ie, bidirectionally), the system is actually designed to allow this, though it'd take a bit of work to make it run (not something anyone's ever needed to do).

There's a lot of value in the existing code and you're likely better off just modifying it/extending it to do what you want than completely replacing it, but if you want to, go ahead. :)

#8 For security, we have a very very good idea for it ;). So good in fact that we've tested it and it works very well, so good that we're going to license it afterwards ;). For windows firewall, that's a problem with EVERY online game, if they don't have access to open up ports on the windows firewall, they can't play. The netcode also has some pretty good NAT traversal.

The point is, we already HAVE netcode that we are going to put in, we just need an engine to do everything else, after all, we don't want to go write a whole engine by ourselves.

#9 Good luck replacing the network solutions :) Will most likely take a couple of months.

#10 Adil, you might want to check out OGRE 3D. It is componetized in a way to allow you to pick and choose the different pieces you need (you can do that some in Torque but not really for the networking...it is too fundamental). From your requirements, it sounds like OGRE is a better fit for this project.

#11 Thanks alot! I'll check it out ;)