Can I design use UTscript creative MMORPG / RAC / ACT / FIG ... game ?

[Brexer]

We weren't disputing that Brexer. What we're saying is, is that for high performance networking ... TCPLink probably isn't ideal.

Oh nono, i know thats not what we where disputing, i just ment incase it really is so low performance, that we should do some test and if so, push Epic to increase performance atleast on this non source UDK.

And i dont plan on using TCPLink for "high" performance networking, its just a interaction layer between "real" client network and UDK/US

[Mode7James]

I understand this thread is a few months old, but does UDK ship with a way to make a connection through TPC, upgrade to SSL, then open a UDP port for sending/receiving packets to an external server? Is UDPLink removed from Unreal Engine 3? I'm still researching UDK for a multiplatform MMO project we've been building.

If not, is there a plugin architecture to build our own client-side libs in C++ and interface with UDK somehow? I'm surprised this isn't an issue that's been raised more, as I've seen many MMO's (Eve Online's upcoming Incarna) starting to use Unreal Engine. Thanks!!

[Phopojijo]

DLLBind lets you attach libraries to Unrealscript.

[Mode7James]

Thanks Phopojijo! Great to know, I'll look into it! Is there adequate docs for how the engine expects to receive packets, and how the packets are organized? I know the hybrid client/server nature of the engine, which limits concurrent players, but I would like to incorporate some of the predictive strategies built in, with a large-scale distributed main server we are building. Obviously we could do it with full source, but hopefully we can make it work with the udk. That would be the ideal solution anyway.

[Phopojijo]

Yeah -- just because you can attach libraries to unrealscript doesn't mean it's an effective way to use your time.

[JonTerp]

The ideal solution here is to contact Epic Games China. They have built Atlas which is designed for MMO scale games using Unreal Engine 3.

And we've certainly seen how well THAT works. Mortal Online is an unholy mess of a game, and so is any other game that used Atlas. Not saying that's Epic China's fault though.


In the end, if you have to ask these questions you have no business trying to hobble together an MMO. I've ranted about this in the many other "I wanna make a super cool MMO" threads, MMOs are incredibly complex, lumbering, scary, beasts. They cost more time (and therein money) in addition to years of software engineering skills and network knowledge to produce than you could possibly imagine.

For the sake of argument lets say you actually manage to spend 100% of your life for the next 4 years getting some sort of compelling game that allows over 64 players inside a single environment. Would you ever get 64 players to play it at once? Let alone play it for a long period of time? I'll cite LOVE as an example of this.

LOVE is an MMO by one very talented, very experienced, individual. He spent probably around 2 or 3 years making it and released it relatively recently. Have you ever heard of it? You see, even though the concept was interesting and unique it never really caught on or got a sizable player base. Despite all the work that was done on it, the game is just not that fun to play to the masses. (PS: LOVE is awesome, go play it. It's only 10 euros for 120 days.)

To avoid rambling on and on, I'll close out with this: Go make a single player game, or a small, very simple, multiplayer title. It doesn't have to be ground breaking or genre shattering, it just needs to work and it needs be fun to play. If or when you and your team manage that, make a slightly bigger more balls out game, rinse repeat. Then the sky is the limit.

[Mode7James]

I really don't think this is the proper thread to be arguing if a MMO should, or should not be attempted using Unreal Engine. We can all point to failures as a reason to not even try. How can one succeed something they don't even attempt?

Getting back to the topic at hand, I have read about Atlas but their site still has info for GDC 2009 demos. Is it a completely separate tech from the Unreal Engine? I like the scalability and multi-platform nature of Unreal Engine 3, which is why we are researching if it's even possible to use UDK with our custom MMO engine. Mostly the client-side parts of the engine should work for us, but without an ability to send/receive UDP packets for game state information, then we'll need to build our own engine. We aren't a large team and licensing the complete source isn't feasible at this time.

It would be really nice to somehow put a stripped-down game onto each server node, to handle predictive physics / movement.

[Mode7James]

After some research - I don't think DLLBind will work for what we want to do, because it's limited to windows-based dll, not native classes that compile with the final build. Unfortunately I don't think the basic UDK would work for what we want to accomplish. Full licensees have access to the source - but unfortunately the scope is very limited with UDK.

What does a full license cost these days? Still $300k? Do they have a royalty-based license? I emailed for info but never got a response...

[CrystalCore]

Try closer to $1,000,000
Your best bet? I know UDK might seem the dream engine. But go with something else. If you had source access it'd be different. UDK just isn't designed for the MMO developers.

[Mode7James]

Yeah, that's what I discovered a few months ago - I figured I'd do some more research. We could probably build our own engine, but after the GDC tech demo - and the support for mobile platforms ... CryEngine, Unigen, Id Tech 5, Frostbite 2 - all in-house and something we can't even consider at this point. Ahh well - thanks guys for the input. We may do a simple UDK game just to get funding for the full source.