You guys do NOT make this easy -- MMP Part 2: Technical

(Part 1 and Part 2 are related. I'm only doing this at JCL's request.)

Hello Conitec,

I've put a lot of thought and time into this post... some might say 5 years worth of time.

I asked the community what they thought of your advertising this engine as MMOG capable and in general what is their opinion of GS Networking. Here are some of their replies:


Quote:
But , I agree , that we should all ask...or maby , demand (dunno if we have that right) that Conitec finishes 3DGS's MP and MMO capabilities. This should , in theory , be their main focus , because the future of gaming isnt single player...Even cell phones already have MMO games. :|
Why shouldnt we be able to create MMO games ?





Quote:
If I sold something and said it could do (INSERT EXCITING THING HERE), and really within a 10 year period nobody could actually do the (INSERT EXCITING THING HERE), then I'm sure people would be mad.



Quote:
If I could I'd set up a law suit demanding either:
A) Conitec to prove it's MP is capable of 1000's of players
-or-
B) Remove all claims off their site



And my personal quotes:

-After 5 years of dedicating myself to MMOG, I am still unable to create a product that is stable and publishable. You can blame my code all you want, but the things like "incompatible version" cropping up from who knows which server all the time is not our fault. Things like anyone being able to create a dummy program, anyone being able to connect to our system without us preventing it, and then injecting false data into our system and crashing our servers is not our fault... as a matter of fact it's not Gliders fault either and that was years ago and it's still not resolved!

-Several other members have attempted to create MP and MMP solutions with the engine and failed. Or rather, the engine failed them...

-I am not alone in my dissatisfaction with networking. At one point, network fixes on the forecast had an actual priority. Now they are lumped together with other junk (EDIT: I notice that on Aug 15, encryption and encapsultioan are RDY? I am SO confused as to what you guys are doing with Networking and it's still piecemeal changes on a sunset (DPLAY) engine! Do you think throwing us a random bone our of the blue every fews years qualifies as supporting your network engine? I don't. I suspect others won't either.)

The bottom line to this thread is the MMOG promised made on your page:


Quote:
Multizone/multiserver support for MMOG (Pro Edition)


MMOG AND GS DO NOT BELONG TOGETHER UNLESS YOU, CONITEC, STEP UP TO THE PLATE AND GIVE IT SOME DEDICATED CONCERN.

Dplay is NOT MMOG friendly. If you keep trying to squeeze water out of that stone, you will merely be wasting your time for you KNOW that DPLAY cannot and will not stand the test of time. Furthermore, it's building on a house of cards... there is a reason why MS sunset DPLAY and why DPLAY has not been used on any commercial MMOG's that I know of and precious little (if any) modern MPs today.

So what am I asking for here in "ask conitec"? Well a few things:

1) Can you drop Dplay and switch to Raknet? Should not be difficult if your engine design is modular and OO.

2) Can you drop Dplay and switch to XXX? There are many other great network engines out there and in your research, you may find better than RakNet.

3) Can you drop Dplay period? MP and MMP in GS is a failure practically and commercially. You can't argue otherwise. To keep it in the engine is a disservice to those of us that bought it "thinking" that you had testing it for MMOG, believing that when you said "1000's of players", you had actually done it... two things which I and others have found out is not the case.

On todays beta list:
[P] dplay_encrypt activates encryption and compression of data packets sent over the network

Just to inform you.

- insert here what I've written the last time -

jcl, would you be able to answer if I'd post a thread about what I would do to redesign the network system?

A lot of work on encryption.
Too bad it's all for naught since encryption is the LEAST of our net problems.



Do it anyways. JCL or not, it will be very informative to me and others who want to do it right. But put it here so it gets noticed and is less a "theoretical" discussion on "multiplayer" and more a "call to action" on "Ask Conitec".

what MMOG is not possible? im wasting my time? and all my work for nothing?

I would prefer that Conitec drops the MP part. Because everybody can use RakeNet/Enet/... All he has to do is packing the thing into a plugin.
And I think someone who really can finish a MMOG, is able to do this as well

And the positive effect if you drop the network engine: You can concentrate on more important things (Ingame Mapeditor, Shadows for terrains, ...).

Plus there are already good solutions for free (GSTNet) so I don't understand the logic to implent a new network engine which will cost a lot of important time.

Dark_Samurai

Count yourself lucky. Took me 5 years to reach this conclusion.

no no, there must be some way, i mean plugins and things like that, this network engine must be updated, because limiting clients is a bad side of this. and my game is based on network player more then 16 (MMOG) and this means no project done?

THIs must be a nightmare...
Has conitec a solution?

An MMOG requires a lot funding and knowledge, but aside from that it's the same as with all other game genres: Developing it is the responsibility of the user. All we can do is providing functions in the engine.

All network functions are well tested and guaranteed to work. When you have a certain technical problem with an engine function, ask us and we can help. But we can't guarantee that you can finish your game or that it will be successful. That's outside our possibilities.

I can understand your frustration and anger when a huge project fails. Sadly this can happen and does happen with MMOGs as well as with other projects. It's easy to blame the engine then. But I don't think you can tell us seriously that you're developing an MMOG for 5 years, and have not noticed all the time that the engine you're using doesn't work.

SOOOOOOOO not the point.

When I bought the engine, Conitec had advertised connecting 1000's of users to the engine. This was the case all up until maybe a few months ago and A7.

Q1: Can you guarantee that 1000's of clients can connect to the engine?
And if you can, can you guarantee 1000's of USERS (which implies that GS is doing some logic and not just DPLAY shuffling packets) can connect to the engine?

On your current page, you have multizone for MMOG as well as stating that GS is a commercial game engine. Putting both together and we come up with the promise that GS can create a commercial MMOG.

Q2:Can you guarantee that the engine is up to a commercial MMOG?



Again, not the point.

The current network functions work fine. The problem is that they are not enough for MMOG and they haven't been for years!

Consider:
Back in the A6 days, session_connect was broken. It is an obviously critical part of any MMOG. However, as you attached bug fixes with updates AND THAT UPDATE TOOK 6 MONTHS (or more?), GS Networking was on standstill for 6 months, my project was on stand still for 6 months, and honestly, there has been little improvement since! How is this MMOG friendly? For that matter, how is that Conitec supporting it's MP and MMP comunity?

Consider:
Since Glider what was it 3 years ago? -- there has been the problem of anyone creating a dummy GS program and connecting to any GS server, injecting false data on any network index, and this crashes the server.

As there is no way to prevent someone from sending data once they connect AND there is no way to prevent someone from connecting to the server, how is this MMOG friendly?


Consider:
For at least four years, I have stated that intra-server communication is vital for a mmog network. Right now, I have to route all my player AND server farm traffic through a single server. This means that one application has to take ALL the traffic for the game and puts critical data dangerously close to the internet cloud. Now then, given the current single server bottleneck for the engine, how is this MMOG friendly?

Consider:
For the same amount of time, people have been begging for a server list so that they can create lobby games. This is critical for MP games, especially FPS and RTS games. Yet for years of people asking, nothing has been in this direction. How is this MP friendly?

Q3: So how are we supposed to complete a MMOG if a) there is no guarantee from the engine makers that it's even possible, b)the tools are not there for us and c) when we ask for the tools or changes necessary, they aren't given?



Q4: Finally, are there any MMOG's made with DPLAY? If not, then how can you guarantee that anyone can make one with GS if it's based on DPLAY? If not, then why keep putting time and money and risking our projects to keep supporting a network solution THAT NOBODY IS USING.

Do you have ANY idea how hard it is to have someone take you seriously about an MMOG when you tell them it's DPLAY based?

And if you won't listen to me and others, then perhaps a legend can help you see the light:

[gratuitous name dropping]
When I consulted RICHARD BARTLE about DPLAY and MMOGs on a phone conference, he could not think of a single one.

Furthermore, he expressed reservation at my using it because as he said it "there is a good reason why MS sunset DPLAY and a good reason why there are no MMOGs with it". What that reason is I did not ask because the point is that if Bartle expresses concern over DPLAY, well, I listen...
[/gratuitous name dropping]

... and if you don't listen to me, perhaps you'll listen to him.

(PS: I'd hate for it to come to this, but if you want me to ask the co-father of MMOGs why DPLAY is a bad choice, if this will finally convince you that Conitec's continued support of it is a bad idea, I will be MORE than happy to do so!)

Four questions. One answer:
REPLACE DPLAY AT ONCE.
GIVE GS MP PRIORITY FOR ONCE.

Have you ever kept up on Fastlanes work? His many attempts at success? Is this your suttle way to blow away the rising issues coming to your attention?

I'm kind of stunned. JCL go out and make us a demo displaying how great your engine is - prove to us (THE USERS THAT KEEP YOUR COMPANY GOING) that we have reason to believe you.

That's nonsense. You can connect a million users to the engine, but you can only connect several hundred, at maximum, to a single server.

We have advertised 50 users per server on the FAQ page until last year. We have meanwhile removed this recommendation because with A7 and current server technology you can have a lot more than 50, but certainly not "1000's".

Am I the only one that remembers, quite clearly on the network features page, that it said "connect 1000's of users"? Just want to check before we see who's nonsense is more nonsensical.

Have you? I'm not going to believe that someone has been working hard for 5 years and is just now figuring out that the engine isnt going to work out for him. It would appear to me that his 5 years of labor has not been very focused in that case.

Yes actually I have kept a close eye on Fastlane. He may be quick and whitty to comment about someones project - but he is a very dedicated man in his project. Please go look at a few of even his most recent posts then come back here...I'm not going to spoil Fastlanes attempt to get to Conitec.

@Fastlane
http://web.archive.org/web/*/http://www.3dgamestudio.com

To be clear, 3 years of that was learning about networking. This falls under the "I didn't know better".

The last 2 years where "I know better, but I have too much time and money invested in GS already so I have to try to make it work".

And thus after 3 years of learning and 2 years of doing, here we are!

I readily admit that I stuck with this engine longer than anyone else did and try to make it work. I am after all the original GS Net FanBoy. So the point is not that I'm not focused, but rather that IN SPITE of that focus, my advice as to what is needed has never been heeded.

Trito, we have VIRTUALLY THE ENTIRE MULTIPLAYER COMMUNITY saying the same thing. It's not just me. If it took me 5 years to reach this conclusion, then at least Conitec should learn from me instead of dismising what we are ALL saying under the pretense that my project is a failure

Well I saw that he told a lot of people what he thought he was going to accomplish, but without being absolutely sure of his technology. That is not only unwise, but its unscientific. I am sorry for his projects failure, but it is just as much a failure of his own preliminary research.

I think this is where you went wrong, getting upfront money, and then trying to make something work which you were not 100% convinced about.

From 2002:

"In a large scale online game however where the players are mainly walking or chatting, and the game world is split into several zones, much more players - several 1000 - can be handled."

IF GS networking is single server,
AND a single server can only hand 50 people,
AND a single server is a single zone,
THEN how can "serveral 1000" players be handled in a MMOG fashion?

In fact they advertised not 1000 but even unlimited players on the feature page.

But I think you need to apply some common sense, even with only little knowledge of MMOGs it should be clear that you cant literally connect unlimited users to the same PC. For more users you need more servers. I think the FAQ also states this clearly.

Ok, since when did this become about me and my project and not about the engine?

By your words Trito, you too admit that GS Networking is a failure. THAT is what this thread is about.

Don't you and JCL try to make this a personal issue as a way to not address the issue at hand, which is the current state of teh GS network engine as it relates to it's MMOG capabilities.

Exactly. And an unlimited number of servers, each with only 50 players in a separate zone, is NOT MMOG. That's my point. That's a bunch of MP worlds, each with only 50 players, not a single MMP world with hundreds of players.

I don't understand what the problem is here, JCL.

Every person who has used the GS Network engine says it can't do MMOG. EVERYONE.

Every person who has used the GS Network engine says it can't even do MP properly. EVERYONE.

What is it going to take for our message to get through?
When will you admit that DPLAY is the wrong network to support so we can start thinking about the right network?

Or put another way, I and others have been trying to do MP and MMP applications for 5 years... exactly how long have you and Conitec been trying to create a MMOG that your experience should trump the entire MP communities experience?

I think the solution is simple, when there really is something wrong with the MP functions why dont you post a bug report?

Project failure is always sad but there can be many reasons for that and if they are not technical you cant blame Gamestudio. You say only some guy told you this and another guy told you that, but when you worked 5 years with multiplayer you should know yourself very well which of the MP functions does not work. When all functions work and your project failed nevertheless its not Conitecs fault. But when some MP function really does not work, put the problem on the bug forum and it will be fixed and we all will be happy.

Ok, let's get this out of the way right now for it started with JCL and it's snowballed: I don't see WHERE, anywhere, I have stated that my project is a failure.

I have created a viable MMOG solution which right now suffers from a "incompatible engine" error that pops up for no reason and is unpublishible (works in development but not when we publish). Thus, by using only session_connect and the send functions, I have created what is a few bugs short of an MMOG.

And in order to do this, I had to turn off all GS Dead Reckoning AND entity update routines... the former because it was useless as anyone will tell you and the latter because it update all entities in a server (that's really MMOG friendly there!).

As well, we had to create a dll so I could store a players handle in a DB because of GS's notorious "last digit non-precision" when outputing vars outside the GS engine which meant that we could not store a Handle accurately using the native GS code.

As well, I had to create a series of lookup tables and other DLL to handle the fact that in my architecture, what I believe is the ONLY MMOG Architecture with native GS, clients and the "server farm" are all bottlenecked at the one true server, thus making a single failure point at the server (which anyone will tell you is bad network architecture design as there is no fail-back option).

So you see my friends, it is not because my project is a failure that I ask for these changes but rather because it is a success IN SPITE OF GS NETWORKING that prompts me to this thread.

So please, don't read between the lines and please stay on topic; you too JCL: this thread is not about my project, it's about the network engine, it's past, present, and future functionality. There is no need to bring my project into this "as a failure" to discuss this point though I am more than willing to share any aspects of my architecture and code to help you understand my point.

And furthermore Spirit, it's not that the functions are buggy. Again, please stay on topic. It's that the entire GS network architecture is flawed, barely MP capable, and not conducive towards MMOGs.

If you don't believe me, consider that using GS Networking, with all functions working perfectly, would have the server updating every entity about every other entities position every client_update tick. Does that sound like a MMOG friendly architecture?

I have not done much with multiplayer and maybe I'm wrong but I can tell you that you can store player handles in a var.

I store handles in a var all the time. Handles are 32 bit and a var is 32 bit also. There is no "last digit non-precision" because handles have no digits. Technically, handles are long integers. For converting handles to char you just use itoa() and lose no digit.

I also did not understand what you said about the architecture, as far as I've learned, client-server systems control the entities on the servers and send the speed updates to the clients. That's how MMOGs work. The book I've read was "Multiplayer Game Programming" by LaMothe. Other engines I know do it the same way.

Have you tried storing them outside of the engine?

There is a chance that this has been fixed, but store several entity handles in a text file and then look at the handles... you will find that several handles repeat themselves in the last (third) decimal digit because of the lack of precision in the last (third) decimal place. And since Handles use all digits, including decimal, for information, this means that if you do this you can't retrieve the right entity. Hence, Handles cannot be stored in a text file nor a Database with native GS. There are several workarounds but the point is they are workarounds to this faulty system and every workaround gets us further and further away from native GS. Does this seem MMOG friendly to you?



Correct. However, the native GS Architecture sends a "speed update" to EVERY client on the server, no matter if those clients can "see" each other in the world. To give you an example, imagine a MMOG Zone based on planet Earth. If you are in the US, do you really need to know the speed of every person in Japan? Of course not and yet that is exactly what native GS Architecture does.

Sure you can store entity handles outside the engine. I use an array of handles in my C++ app that calls the engine DLL. As I said handles are nothing but long integers, so I dont know for what reason you are not able to store them or what you expect Conitec to fix.

If you think theres something wrong with handles, then Spirit is right you should post a bug report, but if storing handles is really one of the things that prevent you from finishing your MMOG, then it seems to me that you have not had much chances anyway. I'm no expert but if you want I can write you a handle storing function in 5 minutes.

Also, sending of entity updates only to close clients is certainly nice but I dont think that this prevents you from doing a MMOG either. When the world is divided in zones, the servers only send updates to entities in that zone anyway, and in no MMOG you would place the US and Japan in the same zone.

This is not just "nice" for MMP, but "necessary" and native GS's inability to deal with this to date is very much preventing ANYONE from creating a MMOG.

If you can't see why, then please, respectfully, as someone who admits to not having much working experience with MP, please stay clear of this conversation.



Under Lite-C this is true. But for those of us that have developed under C-Script (still supported by Conitec last time I checked), handles are Vars and Vars do not "export" well.



Please read the entire thread before commenting as there are a TON of other issues (most listed above) that affect the ability to achieve a native GS MMOG of which the handle is but one and not even the most important one.

Here you go... "C++ App"
This is not native GS anymore.

What fastlane intends to say and obviously nobody understands: in order to make an mmog, you have to a lot of workarounds, many of them as DLL, because the standard network support does not support these things.
Selective updating of entities is just one of those things. Imagine how cool it would be if the engine took care of that. Bt instead, you have to turn off *any* automatic sending issued by the engine and do it all on your own.

Not what I'd expect from a MMOG capable engine.

Then again I've never tried MMOG, but I've had those thoughts when I built some network framework for 4 players ("what would I have to do to keep traffic down when having more players?").

Thanks Firo. It feels like I'm talking to a wall here, especially with everyone who doesn't do MP putting their two cents in (no offense guys, I know you are just trying to help ), and I appreciate those few of you that "get it" .

I totally understand the issue, I just don't get why to make such a big thing of it. Many companies announce with big words.

Or do you really "enjoy" drinking a Coke? Is the beer you are drinking really the "best premium lager beer"? Is the PC you most recently bought really "the best PC"?

It's marketing, everyone does it. And I'm plenty sure that they still will in the future.

Is it "possible" to make an 1000-player-MMOG with 3DGS using the Acknex Network Engine? It certainly is, but it is never stated that the server would be extremely slow and that you probably have a lag of like 20 seconds. But it "is" possible.

Is it "practical" to make an 1000-player-MMOG with 3DGS using the Acknex Network Engine? No.

Infact, it is ridicously impractical to do so, but still possible.

Use DLL's, period. Even companies using the U2/3 engine need to use plugins. Or even have a look at the Quake 3 engine, many modern games are stillmade with i, but the developers need a lot of plugins in order to include today's features like shaders and the such.

It isn't that big of a deal.

With all due respect, I mentioned the C++ app only because Fastlane told that he could not store handles outside the engine. Of course you can also store handles natively inside 3DGS.

But when you store the handles in a database you need a DLL anyway because Gamestudio has no built in database. But thats not Gamestudios fault, most engines have no own database and need an external database.

I am not experienced with MMOG so I cant in depth comment other issues, but the thing I know well is handles because I use them, and I can tell that there is no problem and no workaround required for storing handles.

Tobias, I'm not going to correct your latest post as I've already stated that handles are a minor issue compared to everything else so it's not worth it.

And I know that you are trying to help but please, as a favor to me, if you don't have experience with MP or MMOG, please don't comment on a MP/MMOG thread, ok?

How do you know Michael?
What Network experience do you bring to this thread?
What do you know that we don't?

I'm not being facetious here or trying to make a fool out of you, but if you claim that this is possible with NO PROOF, if you are saying this with no experience using the GS net engine, then you are blindly touting the GS line (as I did for far too long) without any real practical knowledge of what the engine is or isn't capable of.



It's possible to create a car bridge made out of balsa wood... but it wouldn't be practical. And thus if not practical, then balsa wood should not be advertised as "bridge capable".


This philosophy is like buying a new car under the premise that you can drive it off the lot but then finding out that you have to augment or replace the engine yourself before it will even run! If a car company did this to you, would you shrug it off and say "It isn't that big of a deal."?

If GS states "MMOG" on their features page, then I expect that GS is capable of MMOG, not GS + DLL, not GS + my own network.. native GS. That's how it's presented, that's what I expected, and that's where I was wrong.

And to make matters worse, the only way that several of us working on MMP have found to MAKE GS work is buy turning off just about every high level Network feature GS offers... so what does that say about GS and it's advertisement as a viable MP/MMOG solution?

I just want truth in advertising or better yet, more focused attention on the GS MP community and what all of us are saying: GS Net Engine needs replaced!

The fact is, if you are not happy with the GS Networkengine use the other free solutions (GSTNet, an own plugin). Of course if there stands on the Conitec page that you should be alble to create a MMOG with the native GS Networkengine and you aren't because this is technically impossible than thats bad, but I would say you have to live with it and do the best with your situation. And the best is to move to an other networksolution
Just think about Conitec has to implent a new networkengine, than a lot time will pass until they are ready with that. All the other more necessary stuff where no good solutions are available for free (a ingame mapeditor) will take much longer. And what can you do in this months of time? Nothing because you have to wait for the new networkengine.
The fact is that there are much more people out there who don't need a MMOG network engine but a ingame mapeditor and it would be unlogical if conitec wouldn't do what the bigger group of people wants. Sadly that's the truth so I would think about another solution (like GSTNet, or an own plugin).
And by the way, writing an own plugin won't take long. Trust me and you will be absolutly happy with that thing. Another positive point is that if you find a bug, you can solve it imediatly and don't have to wait for others to fix it

Dark_Samurai

Well besides that things like "enjoying" something are subjective, are you aware that with this argumentation you undermine the credibility of Conitec and the 3DGS feature list?

In this business that's very important and I guess you know quite some examples of where some promised features didn't work at all or just not to the extent it was promised and how that affected to producers behind it. As 3DGS especially is made for beginners in game development such an advertisment is even more dangerous as many won't realise that the given MMO capabilities are unrealistic (well at least that insane number of supported clients was removed)...

When getting back to topic I really dislike the apathy that is showing up (once again). I mean there's no need for "panicing", stopping all the development and focus completely on this task but I have little understanding for the ignorance towards certain facts. I mean that any 3DGS owner with network capability can easily crash ANY server running ANY 3DGS game is a quite tremendous bug and it hasn't been discovered for just a couple of weeks. Still I see no sign of any plan to get this fixed...

Then there's the discussion about the networking engine itself. Fastlane makes a lot of good points as to where the current system lacks things and what would be a proper solution. Still there's no real reaction besides bashing him and Conitec (i.e. JCL) also didn't really show up yet with at least something like "We'll be thinking about what you proposed."...

To conclude this, my opinion is that a 3rd party networking engine really should be integrated. The reason simply is that they won't just be better than a homebrew solution (because Conitec can't spent the same amount of time for only one feature) but the integration of this also will be a thousand times faster than coding everything from ground up. The only disadvantage probably is the costs of all this but that's actually a discussion that could and imo should be made and I think that's actually what really is wanted here: An actual discussion about all this instead of just denying or ignoring certain facts...

This doesn't need to lead to what fastlane or I or someone else proposes but a real discussion has to take place in order to unveil the disadvantages of the current system and how these could be overcome...

Enjoy your meal
Toast

well first of all, I think it is rude to jump to conclusions before finish reading sentences, but thats not the point here, so lets move on.



I have lots of multiplayer experience, especially with 3DGS. I have even released the whole source code of a small shooter project I have been working on which supports up to 8 players in a good working environment. But you need to compile with the pro edition, as it uses Acknex' Network engine.

Anyway, over the course of the years I have developed several small MMOG approaches for testing ideas on how to realize them with 3DGS and figuring out how I do best make players move and update that information only to players in range, etc...

I thought you know me well enough around here to know that I shutup if I don't know what I'm talking about and only discuss when I know, that I know, what I know.

So, I should have "brought" enough experience to talk about this.

What know, is that MMOG's, ARE possible. Thoug for a real environment featuring many,many players with differenct actions(emotes, fighting, moving etc...) the 3DGS Network Engine is ridicously impractical, none the less, it is *possible*.



Ahh the good ol' comparison starts again No but certainly, you can't offer a comparison to a car bridge with an game engine. Let's take the "bridge build kit" for instance. Maybe the kit says, you can build a bridge that can hold up to 100 cars. Though, in the process of building the bridge you realize that it is way too impractical, BUT it is possible.



Once again it is a bad comparison. Cars drive, everyoneknows that for a fact. But a normal car certainly cant drive in some rally for instance. Even some of those high priced cars who are sold as being "as powerful as an F1 car" don't. People then replace the motor.

And also your example would be like comparing it to 3DGS in the sense of: You want to create a box and then see it as a 3D object in the engine, but It wontlet me compile until I apply a texture on it. I want the engine to show me what I have in the level editor, no matter what texture I applied to it.



But it IS, it just is very very impractical. They are not lying as you are stating, they are telling the truth. It just isnt practical.



I never saw anywhere "viable/pratical MP/MMOG solution", just "capable".

Oh boy. How I just want answers from JCL rather than others who are going to get this thread locked - and lead us with no answers.

Sad.

Thats a little tyrannical. Only those who agree with this thread are allowed to post...and jcl. OK.

i`m going to throw my few quid in (inflation don`ca know ) i purchased this engine a6 commercial, then i purchased a7 commercial. i had no idea that the multiplayer aspect of the engine was so terrible. i am a noob at coding and am baseing this on what fastlane has said, it would seem to me that multiplayer functionality should be a priority and all other features for a change should be put back for a bit. if the networking code needs to be fixed then it should be, it is as simple as that really.

I'm sure Fastlane69 can comment on this himself, but he said earlier that his project didn't actually fail.

I think that most of the technical points he made are very legit.

Also, and how ever minor this is, eventhough everything may be possible, this would still be different from being capable of doing/supporting something.

Being 'capable' sort of suggests the functions/features will work 'out-of-the-box' without using too much workarounds (you'll probably run into using workarounds with every engine though).

No doubt! But consider that with GS Net Engine, I have to turn the majority of it off and THEN create workarounds!

Using the car analogy above, it's like being told after you buy the car that in order to drive it home, you have to not use cruise control (ie: dead reckoning), can't rely on what the spedometer said (ie: network stat panel), can't use the turn signal at all (ie: entity updates) and have to find a workaround to using the gas and brake!

I'm not trying to downplay your efforts, Michael but 8 shooter DNE MMOG... and while experience with an 8 shooter can let you say "You can make a 8 shooter with GS" -- which I've never denied -- the architecture required for a MMOG is so radically differnet from that of an 8 shooter that you cannot infer that if you can do 8, you can do 100.

I urge you to TRY and do 100. Then let's see what your opinion of the engine is.



I'm not going to argue semantics with you Michael... call it impractical, call it possible, I don't care. No matter how you cut it, if you are advertising something that is impractical with your engine, that is misleading at best, false advertising at worst.

Michael, I'm just being "practical" with this thread, my friend. I'm trying to make the net engine better for all of us and if you keep interjecting that "all is well" with no experience or facts to back it up, then you are making it more difficult for those of us that NEED change to get it.

(PS: ironically enough Michael, I was you years ago! I too thought that if I can make 8 (which I did) I could do 50 (which I did, but by butchering the engine!). As a network Fanboy, I am probably responsible for keeping this failing solution going longer than needed and I am merely here to make up for my past Fanboy sins. )

And still you don't finish reading my sentences. I have already said that I have also done lots of simple MMOG prototypes (moving, simple clothing change, chat) which I first had to test on my commercial version and later ask a friend here from the forums to compile it on the Pro.

And like I said, it is possible, but very impractical, even for simple moving, lots of workarounds (in c-script) were needed, but it was possible, that is point.

I don't deny the fact that the 3dgs network engine is unusable, just that you accuse them of doing faulty advertising.

But let that be aside.. I am totally with you that the network engine needs help, but I don't think that this "crusade" will help much.

I'm sorry I missed this. I'm posting a mile a minute to keep the momentum going and I'm sorry if I glossed over your post in an attempt to get my main point across.

Note that the important point is not what you or I have done, it's that...



...which is all I'm trying to get across and change!!!!



Not in this thread. This thread is about updating the GS engine so it is MMOG usable. The other thread is about false advertising given the current network architecture of the engine.



Then what will? If EVERYONE agrees that the net engine is failing, if the ENTIRE MP community is saying this, then how can we get our voices heard and our demands for change met if not with a "crusade"?

I've just seen that this thread has grown very long meanwhile. So I apologize that I could not read it in all details. However, I understood that you worked 5 years on an MMOG project that wasn't successful, and you want to communicate that the reason was our engine, resp. some missing features. The main problems that I could extract from your posts were:

- You've heard that the Directplay network library is slow or unreliable, and want another library.
- You want the server to send entity updates only to a subset of clients in its zone.
- You got error messages in your game.
- People told you that you can't do a MMOG with Gamestudio.

Is that a correct summary of your main problems so far?

Please correct me if I'm wrong. Otherwise I'll think about how we can help in your case, and will post an answer in the next days. We take such a problem seriously.

The GS network is too weak to handle MMOG, though i think it is possible with lite-c? or plugins.. i mean, thats how others games may do it too...

Would be nice a little demo from conitec, just to show that MMOG is possible and pratical, because if its not pratical, then should not exist and start to think in a new network engine.

i gues that this wont help much if you go posting like this

i get that you are angry and want to try help use and yourself
and people that want to buy gs because of it easy working program white freedom

now may question is wat is going wrong, wat are we missing, wat do we want

what i miss is how much is incoming how much is realy going out on data trafic
-big data tranfser (5mb+ data transfer (for updating and that kind of stuff))
-sound transver
-local network servers finder(list of all sserver not just connect one)
-internet "" """(not realy necceray)

not for me
-mysql (dont know but maby handy)
-enz.


what do we want:
oke you want a different netwrok system so what are the options and what are the good side's and the bad side's
what do other player's think of it
(it will be a discusion like the pshycic enginge i gues)

i dont like talking that much so i will keep it short but hope that this help all is this info that knows every one

i hope i posted in the good tread of the 2 (:

flits

JCL, if you aren't reading the entire thread, then there is no sense in giving you a detailed reply that you won't read.

Instead, I will merely ask you one question:

If the entire MP community (barring Michael) is telling you the GS Net Engine is useless, then why won't you listen to us?

So why not work WITH us to get a better network for GS instead of AGAINST us in defense of a dead engine (DPLAY)?

Amen Flits.

But like a drug addict, we won't get anywhere until we get Conitec to admit there is a problem!

m sure there is a explanation about why keeping DPLAY, JCL sure does has a reason, that maybe its hard to workaround. But im optimist about someday it will must come another network engine, its a matter of time, however, this is a problem that must be quicky handled, alot of projects depending on this probably.

Hallelujah Mike!

If an MMOG is possible with DPLAY, then prove it.
Here is my proof that it's not:

Directplay games: see any MMOG or for th...browser based)?

If a 50 person MMOG is possible with native GS, then prove it.
The "global entity update" proves that it's not.

If a stable MMOG is possible with native GS, then prove it.
The inability to prevent someone from connecting and injecting data to crash our engine proves it's not (though I admit this may have been fixed in A7).

There is my proof of the inadequacy of DPLAY for the modern MP environment.
Can we work on solutions now or do you Conitec has counter-proof you would care to present?

Let's try this:

Fresh start: GS Networking discussion

the features list of game studio lite-c full says no multyplayr is supported?



HOW COME?? this means my game is not multyplayer able?

Apparently - I never knew that. Yes, you can still do network MP programming but you must use external DLLs - you cannot use Lite-C's native MP solution (Fastlane will be happy :P)

This does not make me happy, it makes me confused.
And it can't be right since A7 supports sending structs and other lite-c data structures. Look in the manual under "send_data_to".

EDIT: I do not think it is saying that lite-c doesn't do MP but rather if you only buy the Lite-C version of GS, you don't get MP. If you buy Comm or Pro, you do get MP, which Lite-C can do.

i can't even use the gstnet plugin for lite-c , it crashes engine, when creating client..


OH MY GOD THIS CANT BE!
I PROTEST.. "WE WANT TO MAKE MMOG's, LETS BRING A NEW DAY, AND FORGET DPLAY"

Our development team runs into the same issue with GSTNet with the Lite-C translation.

fastlane, as far as i understood, JCL offered you what you demand with this post:


why not take this offer and start being creativ?

Yep, that's it.

Oh come on, at least spent some hours trying GSTNet to work instead of bursting to tears here that it's gamestudio's fault at the first error message. I have GSTNet working perfectly fine. These post are just aimed to make a dramatic scene and are not helpful at all.

And there is nothing wrong with the lite-c translation nfs42 made.

He admitted to not reading the entire thread and 3 of his 4 points have nothing to do with the issue at hand.

Hence, I was creative as you say.

Ok this will be off topic but i have to answer because im not that noob your trying to make public think ok?







Jooz i did spent! yesterday you went bed, and i still tryed all night long, and i tryed for more 5 hours! the thing is that, when i use it to make a server it works, but when using as client, it crashes... but it has 2 different situations:

1: the ip is in a string pointer i.e:
STRING* s_ip="89.bla bla";~
and then call client start (e, 2300) and this wont crash but the connection will never be made..

or if i call direclty

client start("88.90.. ",2300) it will crash.. and im sure im doing everything like examples demand... Am not saying its GS fault! but i said, i cant even work with the lite-c version.of the gstnet one (as lite-c) ... dont get me wrong lol.

I used GSTNET for multyplayr once, with c-script and it worked perfeclty, so its not that i dont know how to work.

Lol, fastlane69 has finally lost his mind.

But you know what, he is right, the native 3DGS multiplayer networking doesn't work efficiently over the internet. Sorry Conetic, but that's the truth.

Make us a deathmatch demo that we can connect like 16-32 players to running around shooting and killing each other over the internet that runs smoothly and I might change my mind. Oh yes, in the real world not everyone has a LAN line speed connection over the internet, the native multiplayer networking should handle this, although the host should have best computer and connection speed.

I sit here thinking, how do you know it really works Conetic if you don't even have a demo made that shows it working for like 16-32 players over the internet? Surely you have a Shoot Em Up demo you have tested the networking with over the internet you can share with us. If not, then that leads me to believe you tested some simple multiplayer code over LAN and never really stress tested anything over the internet at all, which is very sad if true.

I am not even going to discuss MMO's here. Just prove to us that 16-32 players in a multplayer demo that involves more that just walking around works first.

Not that it matters to me, we are working outside of 3DGS networking now, but the fact that you (Conetic) are not listening to your customers and actually putting them down for telling the truth and stressing that they "Failed" pisses me off to no end. Maybe they "Failed" because they tried to use your product as advertised and your product "Failed", which when it comes to 3DGS networking I think I have a pretty good idea of which "Failed".

Let's say I make a multiplayer game and advertise that 32 players can play it, maybe I should do months of test under different circumstances (connection speeds, etc) and make sure 32 players can actually play it over the internet before I release it and advertise it a such? If the game will only run over LAN or if everyone has to have a T2 internet connection speed or better, maybe I should put that on the game box label?

If you don't have the time to even read and take into account what problems are being presented to you in the forums and dismissing them without proving that the issue isn't true, then maybe we don't have the time to waste with your product.

Later,
Loco

I am working on a proof of concept with A6 pro that is entirely hinged on being a MMO. I have read carefully through this entire thread and every single post, because this kind of issue concerns me greatly.

It is my full understanding that in order to achieve an MMO I would need to create a game world that existed across xNumber of servers, each accomodating approx 64 players. With hand written enhancements such as only running things on clients and using the server to check that everyone is playing fair and in reasonably good synch with each other.

Limiting what is sent to and from the server to essential traffic.

If I understand what fastlane and the majority of the experienced MP developers who have posted here, including Mr.Schwarz, they are finding the network infrastructure to be somewhat over difficult.

The debate over what is advertised for the product, I believe, is sidetracking this issue, at least for now. The real message Conitec should be sending out, and maybe are trying too, is that any suggestions on ways to geniunly improve the product are welcome, and we will look into this seriosuly.

Frankly, a product based on a comunity of enthusiastic game developers needs to at least be seen to listen to its customers. So there is really no need to escalate the issue based on what is or isnt currently or previously advertised.

Is the issue even, d_play or not to d_play?

I'm not sure. I think the issue is that we don't have a conitec made demo for many of the core features and Multiplay, especially MMO's, are a very hot topic that we should definatly have a demo for.

TBH: If I was conitec, I would be making these demos, branding them with Conitec content and using them both as advertising for potential new customers and to prove core tech.

This would really help move gamestudio forward and greatly reduce any threads here that debate core functionality, and I would spend less time reading condesending responses about '5 year failed projects' and validity of the developers claim there in.

I am working as a developer for one of the largest games companies in Europe and our products can take 3-5 years to shelf. If we found out our core technology might not be capable of something huge like MP less than a year from shipping....

Chaos!

So please, lets try and figure out what the problem is and maybe someone here could make a 16-32 player demo, and we could all calm down a little

plus, I can continue getting my fumbled MP scripting a little less fuzzled

@fastlane: just spotted the 'False Advertising' thread, definatly good to seperate these issues. Sorry I missed it before this post

i gonna move fully to torque soon,i cann't wait till A10 released conitec will change their mind to listen and fully realize into market trend(mmo*),i predicted from beta pages in next 10 months we all need to renew our license for a8,a10? it means another 250$x3=750$+3 year waiting,iam not programmer guru,but if conitec really expecting us to really dependant on 3rd party plugin,i can imagine a7 will become so like .Net framework family,a full bloatware,even to run a hello word app you need 2 mb ram!! while with asm only 160kb used ram,and with c++ 250 kb used ram,can you see the different why us really insist all of the mmog infrastructure need to be supported natively by the engine not by the 3rd party?

As if conitec intention from beginning to only offer basic multiplayer networking support,i must agree with fastlane opinion regarding false advertisement,to be honest,i was ordered pro edition because of mmo capable advertisement,if i live in germany i will sue for this missdirection to me as customer.Conitec should change the feature in the pages into Networking Framework and eliminated the unlimited number of user in the pages as the supported user by the native network engine only around 20 user?if still unlimited ,conitec need to put a note with BIG black ITALIC note with "Need additional 3rd party addon"

We just want conitec to be more professional,and proof for whatever you listed in the feature pages! i just wasted a grand(a6 pro+a7 pro upgrade!) for this mmo network dream! :(,as i already said bought pro because of wanted to focus on mmo!

I does. I admit it was merely a way of getting Conitec's attention on this issue since we've been unable to get traction on these issues for years! However, once it accomplished it's goal, it got out of hand. However, in the "MMOLove" thread, I believe things are back in hand as we are discussion solutions to make the engine better... which is all I wanted to begin with!!!

As such, Indi and Loco, I encourage you to throw your two cents (or pesos or euros) into the "Make MMOLove not MMOWar" thread in the Future forum. Your posts would add great value to the thread and your insight will really help define what I hope will be the next evolution of the GS Net engine.


Can't lose what you never had.

Come on conitec at last write a plugin dll to give us, and everything would be okay. I think DLL is the best solution, istead of an update, since other players from other engines versions may not recieve it ( DLL plugin for MOG=)