MMO kit question
by Brian Irvin · in General Discussion · 10/06/2006 (6:41 pm) · 115 replies
Http://www.mydreamrpg.com/ is this realy worth geting? I am wanting input i may full around it but im not sure since I am currently in the process of designing a game
About the author
#62
If you want to make your game right now, use the MMOKit and later port it to Titas.
And just so you know, I have years, and years of experience in Linux and use it as my primary OS even for the desktop.
I try to make sure that all things coming from us support Linux as quick as possible, since it means I can play with it too :)
Regards,
Dreamer
10/10/2006 (3:50 pm)
Well honestly if your just starting out, your going have an easier time with the MMOKit than Titas since Titas is not yet easy to get running and going just yet.If you want to make your game right now, use the MMOKit and later port it to Titas.
And just so you know, I have years, and years of experience in Linux and use it as my primary OS even for the desktop.
I try to make sure that all things coming from us support Linux as quick as possible, since it means I can play with it too :)
Regards,
Dreamer
#63
10/10/2006 (3:59 pm)
I hear that. thats one thing i want to do as i make mroe and more games is make sure that all os's can support them im still using windows because its the onyl thing that supports games thats being published. I just installed linux on my server.
#64
Such things have to be judged by a combination of software and hardware, not just Dreamer's software. If you really want to pull it apart, the game's architecture apart from Dreamer's code is one element on its own. Then the MMOtoolkit, then TGE/TSE's code, and finally the hardware of both client and server.
Surely with all that in play you cant be trying to hang everything on Dreamer's architecture? That's silly.
I believe that a stress test was done (I follow RC, MMOtoolkit, Kaneva, Multiverse and other such mmo-generation systems) with 600 players and not only were they held stable, but a solution for securing even more players was developed as the result of any professionally run test would.
You seem to contradict yourself where you skip from "because there is always a limit to how much one piece of hardware can do" to "It's not a hardware question."
They are working on a demo MMO and offer more than enough information (comparable to other solutions) to make a decision IMO.
10/10/2006 (4:03 pm)
I dont think that fancy papers or little graphs show much of anything. Tim, your link to multiverse makes your statements a little suspect if only at the outset. Im not affiliated with the MMOtoolkit, but you seem to be attacking Torque on one of its strongest features...the netcode...and thats a bit of a mistake.Such things have to be judged by a combination of software and hardware, not just Dreamer's software. If you really want to pull it apart, the game's architecture apart from Dreamer's code is one element on its own. Then the MMOtoolkit, then TGE/TSE's code, and finally the hardware of both client and server.
Surely with all that in play you cant be trying to hang everything on Dreamer's architecture? That's silly.
I believe that a stress test was done (I follow RC, MMOtoolkit, Kaneva, Multiverse and other such mmo-generation systems) with 600 players and not only were they held stable, but a solution for securing even more players was developed as the result of any professionally run test would.
You seem to contradict yourself where you skip from "because there is always a limit to how much one piece of hardware can do" to "It's not a hardware question."
They are working on a demo MMO and offer more than enough information (comparable to other solutions) to make a decision IMO.
#65
10/10/2006 (4:07 pm)
Realm crafter sucks i had to re activate the stupid thing every so oftner
#66
Their architecture is buggy as heck. Their load balancing is not in place. Ambient and directional lights are just being added; shader support and shadows have really not been, yet. Their art pipe is very difficult to use (a Torque flaw too). Their tools are limited, and still buggy. I know they are anticipating an 'open beta' soon, but frankly - my guess is they are at least a year from being complete enough to make a game from MV without massive alterations to the source code for the team. Featurewise, they are about where Realmcrafter was this time last year, maybe a little worse.
On top of all that, they take a 50% fee out of gross revenues earned by any project which uses their engine. They do not host, in return for that, or do anything but simply provide their code. On a tiny (3000 players) game like ATITD, this results in licensing costs of as much as $270,000 a year. Ouch! Their fee scales down as "low" as "only" 30%, but even that is $180,000 a year for even an ATITD-sized/priced MMO. Again - you get nothing in return for this besides the engine code. Comparable cost for Titas: one time commercial TSE license, plus $150 for the Titas code.
No contest.
Perhaps worst off, every MV game must have all players connect through the central MV server before they connect to the individual game servers. Because of the way this service is set up, if MV ever goes out of business without releasing the code for their login server, your business is over. Ended. You cannot run the game anymore. While I certainly hope they would release that code (and they've said they would, if it came down to that), I do NOT feel like gambling my entire business on a promise made on a forum with no contract backing it up.
Compare Titas with Realmcrafter2 (upcoming in a few months) if you want. Kaneva and Multiverse are IMHO not viable cantidates for game dev because of their licenses, if nothing else.
10/10/2006 (4:28 pm)
Tim, one thing to keep in mind about Multiverse: it's all 'pie in the sky' til they get it working. Which, so far, they have not. Multiverse makes Realmcrafter games look really, really good right now. (Yes, I'm in the MV beta, and no, there is no NDA.)Their architecture is buggy as heck. Their load balancing is not in place. Ambient and directional lights are just being added; shader support and shadows have really not been, yet. Their art pipe is very difficult to use (a Torque flaw too). Their tools are limited, and still buggy. I know they are anticipating an 'open beta' soon, but frankly - my guess is they are at least a year from being complete enough to make a game from MV without massive alterations to the source code for the team. Featurewise, they are about where Realmcrafter was this time last year, maybe a little worse.
On top of all that, they take a 50% fee out of gross revenues earned by any project which uses their engine. They do not host, in return for that, or do anything but simply provide their code. On a tiny (3000 players) game like ATITD, this results in licensing costs of as much as $270,000 a year. Ouch! Their fee scales down as "low" as "only" 30%, but even that is $180,000 a year for even an ATITD-sized/priced MMO. Again - you get nothing in return for this besides the engine code. Comparable cost for Titas: one time commercial TSE license, plus $150 for the Titas code.
No contest.
Perhaps worst off, every MV game must have all players connect through the central MV server before they connect to the individual game servers. Because of the way this service is set up, if MV ever goes out of business without releasing the code for their login server, your business is over. Ended. You cannot run the game anymore. While I certainly hope they would release that code (and they've said they would, if it came down to that), I do NOT feel like gambling my entire business on a promise made on a forum with no contract backing it up.
Compare Titas with Realmcrafter2 (upcoming in a few months) if you want. Kaneva and Multiverse are IMHO not viable cantidates for game dev because of their licenses, if nothing else.
#67
10/10/2006 (4:32 pm)
Unless they fix that crap of having to reactivate realm crafter im not going back to it
#68
Such things have to be judged by a combination of software and hardware, not just Dreamer's software. If you really want to pull it apart, the game's architecture apart from Dreamer's code is one element on its own. Then the MMOtoolkit, then TGE/TSE's code, and finally the hardware of both client and server.
Surely with all that in play you cant be trying to hang everything on Dreamer's architecture? That's silly.
I believe that a stress test was done (I follow RC, MMOtoolkit, Kaneva, Multiverse and other such mmo-generation systems) with 600 players and not only were they held stable, but a solution for securing even more players was developed as the result of any professionally run test would.
You seem to contradict yourself where you skip from "because there is always a limit to how much one piece of hardware can do" to "It's not a hardware question."
They are working on a demo MMO and offer more than enough information (comparable to other solutions) to make a decision IMO.
10/10/2006 (5:39 pm)
I dont think that fancy papers or little graphs show much of anything. Tim, your link to multiverse makes your statements a little suspect if only at the outset. Im not affiliated with the MMOtoolkit, but you seem to be attacking Torque on one of its strongest features...the netcode...and thats a bit of a mistake.Such things have to be judged by a combination of software and hardware, not just Dreamer's software. If you really want to pull it apart, the game's architecture apart from Dreamer's code is one element on its own. Then the MMOtoolkit, then TGE/TSE's code, and finally the hardware of both client and server.
Surely with all that in play you cant be trying to hang everything on Dreamer's architecture? That's silly.
I believe that a stress test was done (I follow RC, MMOtoolkit, Kaneva, Multiverse and other such mmo-generation systems) with 600 players and not only were they held stable, but a solution for securing even more players was developed as the result of any professionally run test would.
You seem to contradict yourself where you skip from "because there is always a limit to how much one piece of hardware can do" to "It's not a hardware question."
They are working on a demo MMO and offer more than enough information (comparable to other solutions) to make a decision IMO.
#69
10/10/2006 (5:52 pm)
Well i purchased mmo kit and waiting to download it
#70
When the question is "how many per server" then a good answer is "we did 600 on a Pentium 4 3 GHz with 1 GB of RAM." A bad answer is "it depends on the machine." Alfred's message helps, but still has a few problems:
1) It doesn't come from the mydreamrpg people, so they can't be held accountable to it.
2) It doesn't state what hardware was used (might have been less, might have been more).
3) It doesn't state whether this was all in a single zone (level), or players spread out over multiple zones (levels).
These should be very simple questions to answer for anyone selling a toolkit, and the fact that Dreamer hasn't, seems off-putting to me.
All the talk about Mosix, however, made me very skeptical, becaue Dreamer stated that Torque with MMOKit is already threaded (which it isn't), and that MMOKit would automatically scale on a Mosix cluster (which it won't, as far as I can tell). This implies to me that Dreamer is walking on very thin ice, without the decency to caveat his statements with what he actually knows, and what is just conjecture (and, if so, on what basis).
10/10/2006 (7:42 pm)
What surprises me are the smoke and mirrors, and outright un-truths, that Dreamer have added to this thread, with no caveats whatsoever. That shows, to me, either a lack of understanding of the subject matter, or a lack of regard for the people he's talking to, or both. Neither is a good sign.When the question is "how many per server" then a good answer is "we did 600 on a Pentium 4 3 GHz with 1 GB of RAM." A bad answer is "it depends on the machine." Alfred's message helps, but still has a few problems:
1) It doesn't come from the mydreamrpg people, so they can't be held accountable to it.
2) It doesn't state what hardware was used (might have been less, might have been more).
3) It doesn't state whether this was all in a single zone (level), or players spread out over multiple zones (levels).
These should be very simple questions to answer for anyone selling a toolkit, and the fact that Dreamer hasn't, seems off-putting to me.
All the talk about Mosix, however, made me very skeptical, becaue Dreamer stated that Torque with MMOKit is already threaded (which it isn't), and that MMOKit would automatically scale on a Mosix cluster (which it won't, as far as I can tell). This implies to me that Dreamer is walking on very thin ice, without the decency to caveat his statements with what he actually knows, and what is just conjecture (and, if so, on what basis).
#71
10/10/2006 (7:47 pm)
Well as soon as someone gives me the download link ill download it and play with it some hopefuly it doesnt havee that activation every so often like realm crafter. I allready posted on the license thread please give me the link please
#72
Mike
10/10/2006 (8:02 pm)
@Brian - Sorry for the delay Brian you never signed up for the website which made it harder to find and complete your transaction. However, I am delighted to inform you that everything has been verified and you may now proceed to the client area of our store to download your kit. Thanks and hope to see you in IRC.Mike
#73
10/10/2006 (8:07 pm)
What is the .rar files for?
#74
You need to be more careful talking about decency and un-truths. I believe as Dreamer has said that it does not behoove anyone to give out hard figures, because no 2 people are going to see the exact same results, or even similar results. Even given the same hardware there are software differences such as the OS and what other processes are running that will change the number of achievable users.
"Neither is a good sign." What is a good sign is the fact that Dreamer has 24/7 customer service both with a webpage live chat and IRC channel. The response is damn near instantaneous. There's a long list of completed features on his website. If you like those, buy it. If you dont, then dont.
As to the "not multi-threaded" I think you really, really need to re-read the source code, a simple grep shows threading code in many, many files. And while I'm not an expert at threaded code design, I'd have to say that it doesn't look like it has a thing to do with Theora. This sounds to me like your going on about something you know nothing about, like your "Script has no grabage collection" post.
Ive found a few of your posts www.garagegames.com/mg/forums/result.thread.php?qt=51121 and www.garagegames.com/mg/forums/result.thread.php?qt=50948 which dont make you sound like someone able to impeach Dreamer's credibility.
In otherwords, read the source and then come back if you still believe what your saying. Its not like he doesnt have a build working...1.4 of his toolkit is running!
Let's not get into a flame war though. Dreamer's got a viable, helpful solution for people who want an MMO system built on the beautiful Torque graphics/netcode.
10/10/2006 (8:32 pm)
You seem bent on slagging a good guy who lets people from other engine forums post on his boards with little to no moderation. You need to be more careful talking about decency and un-truths. I believe as Dreamer has said that it does not behoove anyone to give out hard figures, because no 2 people are going to see the exact same results, or even similar results. Even given the same hardware there are software differences such as the OS and what other processes are running that will change the number of achievable users.
"Neither is a good sign." What is a good sign is the fact that Dreamer has 24/7 customer service both with a webpage live chat and IRC channel. The response is damn near instantaneous. There's a long list of completed features on his website. If you like those, buy it. If you dont, then dont.
As to the "not multi-threaded" I think you really, really need to re-read the source code, a simple grep shows threading code in many, many files. And while I'm not an expert at threaded code design, I'd have to say that it doesn't look like it has a thing to do with Theora. This sounds to me like your going on about something you know nothing about, like your "Script has no grabage collection" post.
Ive found a few of your posts www.garagegames.com/mg/forums/result.thread.php?qt=51121 and www.garagegames.com/mg/forums/result.thread.php?qt=50948 which dont make you sound like someone able to impeach Dreamer's credibility.
In otherwords, read the source and then come back if you still believe what your saying. Its not like he doesnt have a build working...1.4 of his toolkit is running!
Let's not get into a flame war though. Dreamer's got a viable, helpful solution for people who want an MMO system built on the beautiful Torque graphics/netcode.
#75
10/10/2006 (8:45 pm)
@Brian you may want to come on over to our site for support issues since you'll get a much faster and better response.
#76
10/10/2006 (9:34 pm)
Alfred who was that directed at? 
#77
Ashtara
10/10/2006 (11:01 pm)
@hplus Wow talk about being transparent (Multiverse huh). I am a member of the staff over at Dream Games. In fact there are several of us Developers on this project. Good luck impeaching all of us. As for that you may as well try to discredit our entire community but that isn't going to happen. The MMORPG kit is a very good and well tested product.Ashtara
#78
/me begins looking for marshmallows
10/10/2006 (11:10 pm)
/me hides and puts on flameproof suit./me begins looking for marshmallows
#79
10/11/2006 (4:28 am)
It was directed at "hplus" of course - as I quoted him and pointed to threads he participated in.
#80
I wonder how those threads are relevant to the facts we discuss in this thread? In the first thread, I asked about debugging tools for TorqueScript memory management, because there was no documentation about it, and was told that there wasn't any -- i e, you have to be careful to add every object to a SimGroup (or go write in the support yourself). In the other, I pointed out a bug in the source code, and was told that the code as shipped by GarageGames has this bug, but if I apply a community bug fix, it will go away. (And, gee, here I thought that the vendor of the code would actually ship the latest available code -- silly me!)
Meanwhile, I have also contributed resources that fix the particle editor script generation, that fix the 3dsmax exporter system scale problem (so you can model with real scale), and an improvement to the telnet debugger API.
Anyway, let's stick to the facts of this thread, because I don't really feel the need to defend my person nor attack that of anyone else; I would honestly like to get answers to factual questions. Maybe I'm unique in this regard?
If you search for TORQUE_MULTITHREAD in the source code (the CVS HEAD version based off 1.4.2, that is), you find that the memory manager has been made thread safe, and the console code has been made to marshal any console function calls to the main thread using schedule callbacks. There is also an implementation of creating and joining threads for each platform. That's it. The actual game, simulation, networking or rendering code do not actually use threads.
All the other mentions of "thread" in the Torque SDK have to do with the concept of "thread" as a currently running animation, which is not a "thread" in the multi-threading, scaling-across-CPUs sense. I stand by my claim that Torque is not multi-threaded except for the Theora player part. If you can show me differently by referencing specific source files and functions, I will gladly accept the correction and apoligize for calling you a liar.
The fact that any TorqueScript execution marshals to the main thread makes threading across multiple nodes in a cluster pretty useless, as the context switch cost will drown out any possible gain from putting separate threads on separate nodes. However, even that is irrelevant, because OpenMOSIX actually does not spread individual threads of a single process across nodes -- it only migrates full processes.
So, from where I sit, Dreamer has not answered the question (for whatever reason), and has further actually advocated a technology that the MMOKit isn't using, to solve a problem that the technology can't solve. He may answer questions 24/7, but if the answers are of the caliber that I've seen in this forum thread, wouldn't you have to wonder how useful those answers would really be?
10/11/2006 (11:27 am)
Alfred,I wonder how those threads are relevant to the facts we discuss in this thread? In the first thread, I asked about debugging tools for TorqueScript memory management, because there was no documentation about it, and was told that there wasn't any -- i e, you have to be careful to add every object to a SimGroup (or go write in the support yourself). In the other, I pointed out a bug in the source code, and was told that the code as shipped by GarageGames has this bug, but if I apply a community bug fix, it will go away. (And, gee, here I thought that the vendor of the code would actually ship the latest available code -- silly me!)
Meanwhile, I have also contributed resources that fix the particle editor script generation, that fix the 3dsmax exporter system scale problem (so you can model with real scale), and an improvement to the telnet debugger API.
Anyway, let's stick to the facts of this thread, because I don't really feel the need to defend my person nor attack that of anyone else; I would honestly like to get answers to factual questions. Maybe I'm unique in this regard?
If you search for TORQUE_MULTITHREAD in the source code (the CVS HEAD version based off 1.4.2, that is), you find that the memory manager has been made thread safe, and the console code has been made to marshal any console function calls to the main thread using schedule callbacks. There is also an implementation of creating and joining threads for each platform. That's it. The actual game, simulation, networking or rendering code do not actually use threads.
All the other mentions of "thread" in the Torque SDK have to do with the concept of "thread" as a currently running animation, which is not a "thread" in the multi-threading, scaling-across-CPUs sense. I stand by my claim that Torque is not multi-threaded except for the Theora player part. If you can show me differently by referencing specific source files and functions, I will gladly accept the correction and apoligize for calling you a liar.
The fact that any TorqueScript execution marshals to the main thread makes threading across multiple nodes in a cluster pretty useless, as the context switch cost will drown out any possible gain from putting separate threads on separate nodes. However, even that is irrelevant, because OpenMOSIX actually does not spread individual threads of a single process across nodes -- it only migrates full processes.
So, from where I sit, Dreamer has not answered the question (for whatever reason), and has further actually advocated a technology that the MMOKit isn't using, to solve a problem that the technology can't solve. He may answer questions 24/7, but if the answers are of the caliber that I've seen in this forum thread, wouldn't you have to wonder how useful those answers would really be?
Torque 3D Owner Brian Irvin