Physics, MMoKIT , TSE and other questions
by Darren Howes · in Torque Game Engine · 06/20/2006 (3:53 am) · 5 replies
Lo folks, I am currently shortlisting TGE/TSE against other engines for use in a proof of concept for a new type of mmo enviroment.
Was just wondering if folk could help answer a few questions to help me on my evaluation before I start shelling out bucks for licenses.
1. Physics
I have read on a lot of forums that TGE and TSE's current drawback is the limited physics capabilities. I am guessing that the ragdoll kit was brought in to help strengthen some of this.
What success have developers had incorporating TGE and TSE into other physics engines (ODE and Novodex for example) ideally I am looking for something that is not too expensive and preferably did not require folk to buy extra hardware to get the most out of the title.
2. MMoKit
Had a look at mydreamrpg.com for helping to give a leg up in the back end development of this mmo and was wondering what folks experience of it had been. Anyone used it on titles ported to TSE ?
Had a look at the online demo and found it clunky at best. Let me know your experiences :)
3. TSE
I know that this is still in some lvl's of refining, is it safe to start utilising things such as the atlas terrain generation for trialling out some proof of concepts on lvl design ? What is the general consensus of how good TSE is especially when combined with GG's lighting kit ?
4. TSE MMO viability
What is the general consensus in the community of TSE's viability as a foundation to build an mmo on ?
Currently I am evaluating Garage games products against TV3D so any points for and against TGE/TSE against it's rival would also be very much welcomed.
Thanks in advance for any help/insite you can provide guys(and gals) :)
Darren
Was just wondering if folk could help answer a few questions to help me on my evaluation before I start shelling out bucks for licenses.
1. Physics
I have read on a lot of forums that TGE and TSE's current drawback is the limited physics capabilities. I am guessing that the ragdoll kit was brought in to help strengthen some of this.
What success have developers had incorporating TGE and TSE into other physics engines (ODE and Novodex for example) ideally I am looking for something that is not too expensive and preferably did not require folk to buy extra hardware to get the most out of the title.
2. MMoKit
Had a look at mydreamrpg.com for helping to give a leg up in the back end development of this mmo and was wondering what folks experience of it had been. Anyone used it on titles ported to TSE ?
Had a look at the online demo and found it clunky at best. Let me know your experiences :)
3. TSE
I know that this is still in some lvl's of refining, is it safe to start utilising things such as the atlas terrain generation for trialling out some proof of concepts on lvl design ? What is the general consensus of how good TSE is especially when combined with GG's lighting kit ?
4. TSE MMO viability
What is the general consensus in the community of TSE's viability as a foundation to build an mmo on ?
Currently I am evaluating Garage games products against TV3D so any points for and against TGE/TSE against it's rival would also be very much welcomed.
Thanks in advance for any help/insite you can provide guys(and gals) :)
Darren
#2
Ian "Xest" Winter pretty much took the words out of my mouth. Just read what he said, I share the same opinions.
06/20/2006 (6:48 am)
Hi Darren Howes,Ian "Xest" Winter pretty much took the words out of my mouth. Just read what he said, I share the same opinions.
#3
The MMOKit is still new in terms of developement time (as is MoM). Proper zoning and setup of the backend will require another huge chunk of dev time as most boxed solutions are based on a number of assumptions which may or may not work for your gametype (and MMO is not a gametype, it is a collective title to a large number of gametypes).
TSE is usable, but I do not know if it is usable in the nature you need it. It has massive terrain sizes and is paged, but it is not "zoned" in the manner of MMO's. It would take some time and programming dedication (see a trend here) to do it.
Of course, comparing it to TV3D, networking is still a strong point since TVNet is still unavailable. You will have to implement a third-party networking package to get up and running before TVNet is released.
TV is nice and easy to work with, though you will have to do some major tinkering to make it netsafe for a MMO. C4 as well. Actually, you will have to tweak any engine, even a tried-and-true MMO engine for your specific needs.
But of the engines listed, I would recommend TGE over the rest simply because it 1) actually has a network layer and 2) has a MMO kit which has made attempts to modify that layer accordingly. Not the most heartening of reasons or comparisons, but that's kind of where it is at in my eyes. You could look at Realmcrafter, kineva (I think that was how it is spelled) and the open source alternatives which are designed with MMO's in mind.
06/20/2006 (7:49 am)
You are going to have to make some rather hefty compromises with physics in a networked environment, especially if you need those physics to occur the same on all clients. Right now there is no system that does this effectively. Every system has a number of manual "best guess" tweaks which are not much different than TGE's datablocking and rigid body phyics. Most of the time these general features are pushed out and the client simulates them with tweaking in an attempt to assure that the end state is the same on all machines. A physics intensive MMO would be quite a feat using any boxed solution and would require major programming overhead to make sure that it was as fast and efficient as possible as well as being as accurate as possible, and those two things combined can take a lot of time out of your schedule.The MMOKit is still new in terms of developement time (as is MoM). Proper zoning and setup of the backend will require another huge chunk of dev time as most boxed solutions are based on a number of assumptions which may or may not work for your gametype (and MMO is not a gametype, it is a collective title to a large number of gametypes).
TSE is usable, but I do not know if it is usable in the nature you need it. It has massive terrain sizes and is paged, but it is not "zoned" in the manner of MMO's. It would take some time and programming dedication (see a trend here) to do it.
Of course, comparing it to TV3D, networking is still a strong point since TVNet is still unavailable. You will have to implement a third-party networking package to get up and running before TVNet is released.
TV is nice and easy to work with, though you will have to do some major tinkering to make it netsafe for a MMO. C4 as well. Actually, you will have to tweak any engine, even a tried-and-true MMO engine for your specific needs.
But of the engines listed, I would recommend TGE over the rest simply because it 1) actually has a network layer and 2) has a MMO kit which has made attempts to modify that layer accordingly. Not the most heartening of reasons or comparisons, but that's kind of where it is at in my eyes. You could look at Realmcrafter, kineva (I think that was how it is spelled) and the open source alternatives which are designed with MMO's in mind.
#4
Nothing personal, but good luck on this--mutually incompatible design requirements doesn't even begin to suggest how difficult this will be.
06/20/2006 (3:18 pm)
Just an aside, but you are taking the two most severely incompatible design requirements (MMO and strong physics) and trying to tie them together.Nothing personal, but good luck on this--mutually incompatible design requirements doesn't even begin to suggest how difficult this will be.
#5
EDIT: fixed links
06/20/2006 (3:45 pm)
Dreamers working on his D.R.E.A.M. mmorpg engine for release under the GPL. Theres also Titus for TSE and the MMOkit for TGE.EDIT: fixed links
Torque 3D Owner Ian Winter
The problem with the MMO Kit and well, TSE/TGE in general is that they're not designed for massive amounts of players (to me an MMOG should be capable of handling 1000+ clients). Josh Ritter is aiming to expand TGE's capabilities in this area for MoM but afaik it's not something that will be publicly released. You're going to need a pretty hefty overall of the server backend to make a truly massive MMOG, out of the box you can probably handle 200 players with a little tweaking however.
TSE will (hopefully) be nice, but it's still got some way to being finished and progress is rather slow. Atlas is extremely nice however and fantastic for a non-zone based MMOG ;) The lighting kit for TSE hasn't been released yet, hell TSE doesn't even have proper lighting, well, shadowing yet ;)
TSE is fine for an MMOG as a renderer but you're going to need to do a massive overhaul of the server backend as mentioned above if you want to support more than a couple of hundred players.
As for other engines, have you looked at C4? (www.terathon.com). It's a very clean, modern engine, whilst it lacks proper terrain/physics support right now I hear they're expected by the end of the year which is probably the same time scale for TSE to be finished also. It's much easier to learn with C4, if not only because TGE/TSE suffer the existence of a ton of hackish old legacy code lying around it's code base. C4 is also only $100, however TSE is $250 total right now.
All in all, neither engine is finished yet so you'll need a damn good coder anyhow to finish off unless you're willing to wait - I'd hold final judgement on both until they're in a more polished state if you can, I don't know anything about TV3D though so can't comment there.