Slashdot Mirror
Full GPL Game Company - Nevrax
Loic Dachary writes "Nevrax is quietly building a 100% GPL'ed game client and server that contains a framework, a 3D engine, an AI engine and a Network engine aimed at running massively multi-user entertainment in a 3D environment over the Internet. Since this is a company based in Paris I follow their progress with great enthusiasm, although I'm not a game developer myself. Their business model makes a lot of sense: they won't sell CDs, they will sell access to the massively multiplayer server. What I also like is that they don't plan to release the client under GPL and the server under a non-free license. They release both under the GPL. The proprietary part will be the data files for the world (graphics, maps etc) but that was to be expected. They released a demo game for people to play with so that external contributors won't have to build a whole game of their own before testing their first patch."
Appart from the fact that people will just run free servers like they do now for Ultima Online, wouldn't it make more sense to charge for those graphics that are costing you a fortune to create? Sure, keep the source open, you get all those great benefits and don't have to assign developers to fixing bugs, but you've paid for the graphics and the sounds and the like to be created, doesn't it make sense to charge people for this? If not, put that under a GNU license too!
How we know is more important than what we know.
...the open source game Xconq. That one really huge map of the world is way, way impressive.
("Way impressive" is a technical term they learnt me in college...)
You cannot apply a technological solution to a sociological problem. (Edwards' Law)
I've heard a lot of comments that a GPL's game system will be prone to cheats.
....er ...I mean monsters you kill. Rather a good virtual environment will be about the interactivity between players and their environment. Thus reducing the impulse to cheat, because you truly would be cheating yourself of the experience.
This is true, up to a point, but soon after cheats are discovered then patches will be put to work to stop them. That is the GPL way
The real challenge will be to make a virtual world which isn't about how long you play or how many scripts
What causes 'cheating' and other 'anti-social behavior' in current MMP games is a series of fundamental design flaws, deeply ingrained into the culture of the game developer community, exacerbated by a misguided belief that human society is as amenable to technological manipulation as AI code. It has nothing to do with exposing the client.
The flaws:
- Anonymity without authentication and persistence of identity (I'm not just talking from the developer/server security perspective, which is all that gets attention in the industry; I'm talking about the importance of those three features for healthy interaction among and between human members of an online community)
- Malthusian, (faux-)closed-system game mechanics, where the optimal strategy (and often explicit purpose) is to gain personal power by killing others and stealing their stuff, and where there is no device to recognize nor quantify the synergetic creation of community value;
- Abstract, symbolic, stat-based (read: easy to quantify and shove in a database table) substitutions for the breadth and depth of human experience. Stats are an anchronistic legacy ill-suited for a networked environment. Online Game developers are still trapped in man-vs-machine dynamics. In an online world, populated with real people, 'reputation' can (and should) be more than a database entry, 'skill' more than a roll of the die, and 'power' more than an abstract product of macros and having no actual life.
Frankly, my peers in the industry all sincerely believe that they can literally *engineer* better societies into being, that all it takes is tighter code to solve anger and greed and hate and venality (and that, on the other hand, they can write subroutines for trust and honor and humor and epiphany). They expend tremendous energy in a wasteful search for responsive AI and fool-proof rules, while squandering the resource of the collective eons of human experience embodied in their player community.
The more they fail and the more disfunctional their player communities, the more code they write, database tables they build, and rules, limitations and barriers they erect to players' freedom and self-expression and self-government. "Never trust the client" is a mantra that invisibly and inevitably expands in the mind of developers into "never trust the player". The name of the development game has morphed from 'serving the players' into surviving them.
That is one of the reasons that, despite the incredible immersive power of these technologies, they reach but a tiny portion of their potential audience.
If you build worlds where the political, economic and social fabric is woven by a human network, where the developers are to players as experienced directors are to powerful actors, rather than parents are to wayward children, where the lessons learned from functional RL societies are adapted to the circumstances of online community, where people enjoy anonymity and the freedom to live their fictions, but are still accountable and held responsible for their interactions in a community, and where the collective efforts of participants builds new wealth, rather than necessitating squabling over finite resources;
If you build world where, in other words, the ethos of the network prevails over the ethos of the oilwell;
Well, then, an open source engine is a source of delight and inspiration, rather than a subject of fear and loathing.
As someone working hard to build such worlds, I certainly wish them success.
Flout 'em and scout 'em,
and scout 'em and flout 'em;
Flout 'em and scout 'em,
and scout 'em and flout 'em;
Thought is free. - Shakespeare [The Tempest]
On the other hand, easy scriptability is a game defect. If a game can be "mastered" by writing a script, then the game is uninteresting -- deterministic -- a solved problem, and playing the game is going to get boring after a while. There's plenty of games that get boring fast for this very reason.
The real challenge is twofold:
1) Writing a game that cannot be mastered by computer automation. Something that really, truly requires a human level of creativity, intuition, and interaction. Generally, having humans play against each other introduces a lot of strategy into the game.
2) Designing the interface between the client and server so that the server implements the "laws of nature", and the client implements the "human intelligence."
The problem is in what you consider to be "cheating." Think of Indiana Jones facing down the machete-swinging Ninja. Sure, pulling a gun on him and shooting him is "cheating", if you subscribe to the world-view that the Ninja, by challenging Indy, had defined the rules of the game as a sword-fight. On the other hand, you could interpret the Ninja as being the cheater -- after all, he was highly trained in sword combat, and, as far as he knew, all Indy had was a whip. The Ninja considered himself to be technologically superior to Indy, until Indy pulled his gun, and turned the tables.
If Indy had pulled a gun at a fencing match, he would be, at the very least, disqualified. However, in the wide world of the internet, it becomes impossible to determine who is "cheating" by modifying their client. Thus, the only way to guarantee that no one is cheating is to simply redefine what is fair. You have to start with the premise that you have a client/server interface, and anything on the client side is fair. All "unfairness" must be addressed on the server side and the server side alone. Otherwise, you're fooling yourself, and are going to be plagued with "cheaters."
The correct approach to solving the "cheating client" problem is NOT for the game designer to make a dumb client, and attempt to prevent people from "smartening up" the client. This will always fail, especially if the client is GPLed. The correct approach is to make the client as smart as possible -- so that improving the client becomes a genuine challenge. Then, if improvements are made to the client that create an "unfair" advantage, those improvements can be rolled into the next version of the client, not banned. Sure, it becomes an "arms race", but eventually everyone is going to run into the same walls, and the game remains fair.
The alternative is security-by-obscurity, which is really no security at all. After all, you wouldn't accept a "security system" that did password checking and authentication unprotected on the client side, would you? The common practice of releasing binary-only game clients has made game designers lazy -- a "hackable" client is just as defective as a "hackable" security system.
In a well-designed client/server game, a "smartened up" client with a human controlling it should, in general, be able to defeat a "smartened up" client that is running on autopilot. That "cheater" should wake up the next morning to find his player dead. If not, then your game suffers from a defect -- it doesn't require human-level intelligence to win.
So far I have yet to see a single game come out of the open source 3d engine projects... and of course they're all "massively multiplayer"
Beyond FreeCiv, what other projects have been completed?
We can at least have some hope they will do a good job in the organization of it. Young coders may be fine at the actually writing of a line of code, but a lot of the younger programmers seem to not have the level of organization needed to keep the code maintainable.
As far as all you people who say "security through obscurity is required to prevent cheating", I say that's just because the games weren't thought out properly and designed for it. I know that Blacknova has had a few problems with cheating in the past, but access to the source code wasn't needed. There were simply bugs in the program that could be exploited. Access to the code has only helped them, so that they can have assistance in figuring out how to plug the holes.
And sure people who know how to program well can alter their clients to make them more efficient, but people already do that (I recall external macroing to be very effective in UO). What needs to happen is that the server is secure and actually enforces the rules of the world. To often coders have put a lot of the rule checking in the client. That may let you run the server on less powerful machines, but sooner or later someone is going to figure out how to cheat and then you'll have to start checking it anyway. (eg: quake 2 proxys like speedbot) Better to start out doing the correct checks than just hope nobody will ever figure it out.
This is SO educational! -- Kintaro Oe
Instead of being a wimp and "saving" your character at the Dragon Inn, you would set parameters on his behavior while you are not in control. You might specify what activities he is to pursue while on "autopilot." You might specify how nice or naughty he is to be to other characters. The list is limited only by the imaginations of the server implementors. Choose your offline behavior wisely, and your character prospers...choose foolishly, and find him a ghost when you return!
Alternatively, clients with always-on internet connections could choose to script or program their autopilot behavior client side. Of course, there is the danger of a poorly coded script causing the character to run about the countryside shouting "Natalie Portman eats hot grits!!!" But such unfortunate incidents could be handled by game administrators.
Building up a character while you sleep is not necessarily a bad thing. It merely shifts the value from sitting in front of your terminal for 23 hours a day to choosing your character's activities wisely, both while playing him and while he is on his own.
"Rub her feet." -- L.L.
This might be a mistake. There are people who would prefer to buy the game, even a simple jewel-case with a paper insert and no box, because of the convenience of having a hard backup or not having to download the whole offline. And there are others who would gladly buy the game, if it was good enough, even after downloading it, just to show support.
They may as well offer the option, even if it's only over the Web instead of in retail stores. Why turn down more money?
6. Open Source Clients mean user built-in macros. If you're not a coder, you don't level as fast as the geek next cube who is.
5. Damnit! 'RMS_Troll' gets all the experience *and* loot!
4. Anti-cheat security often means 'Security through Obscurity', since game designers have to do things like encrypt game values in memory to keep them from being altered.
3. What's the 'Karma_Whore' guild, and why do they keep modding my exp points down?
2. Microsoft just steals the code and introduces 'ActiveMMORPG' components for Internet Explorer
1. Damn penguins pit for 9999 HP a peice
The next Slashdot story will be ready soon, but subscribers can beat the rush and slashdot the links early!