Solipsis - a Decentralized Open-Source MMORPG

Anonymous Reader writes "Calling it an MMORPG is a bit of a misnomer because at this point there aren't any players, much less hit points, monsters, or flaming swords. Solipsis is an open-source project that aims to create a decentralized multi-user virtual world. It's still very much in its infancy, and as such the visuals are a bit lacking, but the aim is to create an endlessly scalable user-contributed world and it seems it's a nice platform to play with."

Comments

Glorified Decentralized Chat

[bhive01]

This program is no more than a glorified decentralized chat proggy. Anybody figured out where to find people yet? The Hive

Re:Glorified Decentralized Chat

[ZephyrXero]

Yeah, kind of reminds me of "The Palace" from back in the day....I was hoping for something a little more akin to current MMORPGs....perhaps 3d even?

most MMOs

[HTL2001]

most MMOs forbid client modification... this makes it easy!

and since its decentralized, server modification also. unless they store user data securly on a central server, cheating is gonna be BAD.

Re:most MMOs

[Elwood P Dowd]

unless they store user data securly on a central server, cheating is gonna be BAD.

No, cheating is going to be redefined. While they're a billion miles from Metaverse-like, Neal Stephenson already worked this out pretty well. While you're walking along the street, the computers that run the street place limits on the appearance of your avatar. The sunbeams shooting out of your hairdo do not extend across everything else on the street.

When you get into a fight on one server (in a bar, for example :), they can track your stats. If several servers agree that they trust each other, then they could share stats. Everyone knows there's no cheating in the Black Sun.

If you're worried that the <Lord Pants; Level 60> floating above your head won't mean anything because anyone will be able to do that... then yeah, you're exactly right. Some servers will follow conventions and some won't and that's fine. Hang out in the areas where you like the rules.

Croquet

[Elwood P Dowd]

Also vaguely interesting and along similar lines is Alan Kay's Croquet project.

It's not particularly mind blowing, but it has potential.

Cool... Let's see where this leads to.

[obi]

I'm quite interested in such a system. However, for a true decentralized system you need to put trust metrics at the core of the system, because cheating would just be too easy otherwise.

With a client/server model, you can just say: "everyone trusts what the server says, what the server says goes". With a P2P model you have no such easy way out.

Anyway, I'll be very closely watching this - the only distributed system that comes close is opencroquet, but that's not really suitale for a real-time environment.

While they might not necessarily succeed, it'll be very interesting to see their experience and conclusions once their prototypes start being used.

Re:Cool... Let's see where this leads to.

[rvw14]

Here is an interesting article on cheating in MMORPGs.

http://www.arena.net/news/articles/mikearticle0408 02.html

Re:well, in theory.

[ArsonSmith]

> n00bs making whorehouses and warez trading

If we're lucky!

As an oldtimer, let me say...

[Anonymous Coward]

*Waves cane*

We called them MUDs in our day! And people played with the concept of decentralizing them back then as well. Nothing ever came of it, AFAIK. As other posters have said, trust is a huge issue. There are other problems with latency, bandwidth, synchronization, etc.

Hacks

[CrazyJim1]

When you do that client side stuff, you need to put checks on hackers. I won't get into much detail because maybe .01% of people who read this care, but you can pull off anti-hack tricks. Its important not to allow players to know the anti-hack tricks because they'd work around them. But if you make people pay for the game, and ban them, the overhead of loss will prevent most hackers.

For example: Have every client connect to the main server to track stats. If a stat gets modified faster than it could be changed in game, then an alarm goes off.

If you set up tons of trip wires like this, and institute a player based police system such as Halo 2 has, then you're set. If you don't protect a client side game against hacks, then if it gets popular, it will be hacked into the ground.

Re:Hacks

[cbr2702]

Or you can make a fully decentralized system and pay no heed to cheating. You don't need to have a global idea of stats to have a good game. Each player can run a server component. Different servers correspond to different parts of the world, with a registration/transfer system. Sort of like the web with hyperlinks. Let players decide which servers they like. A server can keep stats, protect names, restrict avatars, or not. Servers with silly rules won't get used. Let econ do the work for you.

Re:Hacks

[Jahf]

For example: Have every client connect to the main server to track stats. If a stat gets modified faster than it could be changed in game, then an alarm goes off.


Under that assumption you have to at least allow 20 hours if not 24 hours of change. Sorry, but given the rate of casual players this would still screw them. You could still advance your character (characters with multiple computers or program instances) as if you were playing all day every day. Not nearly as fast as an insta-cheat, but still far more than a regular "real" player can keep up with.

There is almost no way that a true P2P game would be able to prevent hacking, even with a background checking server (which wouldn't be true P2P anyway). We've already seen a few cases of P2P hashes being hacked without changing their sums recently.

A police system is going to be far more effective than the alternative, but then you have to deal with the question of "10 people flagged this account as cheating but 20 people flagged him as being ok". Its less of an issue in a multiplayer game like Halo 2 (it isn't "massive" and you don't care nearly as much if player A is cheating because he only affects players B through Z, not players B through ZZZZZZ).

There is another plausible idea ... have the P2P network randomly change various binaries used in the game and force all clients to update to the new binaries to continue playing. The clients would have to download the binary -and- propogate it and each client would perform checks back to the server (perhaps even sharing it back to the server in random intervals) so that the server can confirm that not only is the hash the same, but the bits are not different. If the server finds altered binaries, it can force a traceback through the clients that that slice was shared from until it finds the "right" slice. The slice -after- that is the one causing the problems. Eventually the entire client would be refreshed and any impurities wiped out.

It wouldn't prevent cheating 100% of the time, but it would remove 100% of the cheats -over- time.

And if it gets used by a software company, consider this prior art.

Nice... but unfortunately.

[AzraelKans]

Unfortunately this is nothing "new" a good bunch of indie developers have tried to keep down the costs of creating a indie mmorpg by not having a central server (include myself to that list), and keeping it peer to peer instead.

However so far is a lot more of theory (and some mixed bag tech demos) than actual results. Lets face it, if a super MMORPG (like WOW) is having trouble to keep a lag free (more or less) environment by using centralized state of the art equipment and systems with lots of bandwidth to spare. What chances does a run of the mill client in a home PC have? (which is usually connected to a bandwidth sucker proxy which is connected to a bunch of dumb users with a lot of spyware installed) a: none. The lag would be completely unbareable is hardly noticeable for web surfing but for a system sending an update of several dozens of users each 2-3 seconds is a killer.

In the future we are going to see more systems like "guild wars" in which areas are instantiated for a limited number of users (including user based servers I think) instead of one server farm trying to handle all the users all the time. Now thats an idea that actually works! (although it takes the "massive" mostly out of the equation.) and it should be interesting for small developers.

Missing the point

[istewart]

I think the guy who said that it's basically a glorified decentralized chat system hit it right on the head. I read this and thought "Metaverse," and their webpage/wiki says as much. I don't think it's meant to be any sort of a coherent game, although doubtless someone will use it as such.

I think world boundaries and "streets" and other such metaphors for the physical world can be set up by using connection forwarding through other servers. For instance, if your Solipsis server is hosting a structure that's down the "street" from your buddy's server, then you would only accept incoming connections from your buddy's server. You would also block connection spoofing and maintain the illusion by checking back with his server to ask, "is XXX.XXX.XXX.XXX connected to you, and is it headed my way (trying to connect to me)?" Of course, lag issues would have to be worked out, but I certainly think it's something to work from.

I think goofy hacks will run wild, just like in Snow Crash, but server security can be set up to maintain a coherent world and keep out people you don't want around.

Better in theory than in practice

[petrus4]

I had a look at this a week or so back, since there was a link to it on Terra Nova.
I really wasn't particularly impressed, to be honest, although I'm willing to give it the benefit of the doubt and say that it is still very much early days as far as the project is concerned.

There would also be a couple of major obstacles to this in the real world, sadly.

a) With regards to content in particular, Sturgeon's Law would probably apply with a brutal vengeance.

b) With client-side character files and (worse yet) individual control of bandwidth from peers, you'd see 14 year old Neo wannabes swarming out of the woodwork everywhere, with things like the recent Blizzard speed hack, item duping, and so forth.

c) Although most people might, not everybody has broadband yet, sadly...and for this, everyone would need to. (I'm still on a 56k modem myself)

At least in terms of its level of progress, Croquet is far more interesting. I downloaded it and had a mess around with it...and although there are some issues which could be majorly improved, (texture size needs to be made uniform, for one thing) it's coming along well. It will be a while I think before a sufficient portion of the online population will have the processing capacity or bandwidth for a networked version of Croquet to be large-scale viable...but when we get to that point it could be very interesting. It essentially looks like an ancestor of the sort of completely 3D, networked virtual environment that Gibson and others wrote about.