About LWN.net

The Worldforge Project: A Gamer's Perspective

An open source "Ultima Online" with the artificial intelligence (AI) depth of "The Sims"? That's the promise (or at least the goal) of The WorldForge Project. Headed by Bryce Harrington, a NASA programmer "in his spare time", this group project is generating huge amounts of documentation and copious quantities of code.

Their FAQ describes this as a "very ambitious project" and warns that we will "... probably have to wait until 2001 at least before [we] can expect a '1.0' release of the game engine". They're looking for testers of their game engine and games right now, but they warn that, "... anything considered 'stable' at this point in the project probably doesn't do very much".

As with most massively multiplayer games, The WorldForge Project will use a Client/Server architecture. The server will manage player character (characters directly controlled by players or PCs) interaction, world events, and the actions of characters not run by players (non-player characters or NPCs). Each player's computer will run a client application, which will handle graphics and sound generation for that player, based upon data from the server.

At present, they only intend to produce server software for the Linux/UNIX environment, requiring a Pentium II with adequate memory. They do intend to produce client software for other operating systems, with an apparent focus on the Windows environment first.

Their intentions for client software are just as ambitious as the rest of the project. They plan to allow the use of multiple forms of clients. This way, each player can link to the available games in a way that suits his own interests. Currently planned client versions include:

1. An overhead-view, two-dimensional, isometric client, similar to that used in Ultima VII. There will be a nod at three-dimensionality, but no perspective or distance effects (hence isometric). This will be targeted initially at a low-end Pentium system, but they plan to drop the system requirements down to "a 486-ish platform."

2. A fancier isometric client, "... as close to 3D as possible without actually being 3D". This will be a state-of-the-art front end with foreground and background animations, time of day, weather, shadows, lighting, and so forth. They expect this to require a Pentium166 with a PCI video card at minimum.

3. If the server supports it, they may be able to build a true three-dimensional client, with all objects in the world rendered in the round using polygons (think Myth II, from Bungie Software or Tomb Raider 2, from Eidos Games).

4. Also mentioned as possibilities are text-only, audio-only (that should be interesting), VR goggle, and tactile interfaces. No mention of a direct-to-brain interface, but can that be far behind?

5. The screen shots that they show so far are variable in quality, but given the state of the project, they seem really quite nice on average. They seem to have some talented artists contributing to their project, but whether they will be able to maintain a high level of graphical quality in a project this large remains to be seen.

   The area with the most interest, to me at least, is their plan to implement extensive artificial intelligence for seemingly every intelligent being in each world. "When completed, NPCs in Cyphesis will do all sorts of interesting things like attempt to accomplish ever-changing goals that NPCs set for themselves, gossip to PCs and other NPCs, live, die, and raise children. Cyphesis aims to make NPCs act just like you and me. Every time you play WF, Cyphesis will make the game different and interesting because PCs and NPCs will all be individuals with different aspirations, goals and their own methods for reaching those goals".

   This is exactly the sort of AI that serious game players have been wanting for a very long time, but only in a limited-environment game like The Sims has this goal even come close to being realized. The Ultima series games are perhaps the closest to an actual implementation of this sort of AI in a large-environment role-playing game to date, and they don't have the depth alluded to above. In order to make a world large enough to be interesting in the long term, and still have this type of depth throughout, someone has a huge coding job ahead. All I can say is good luck; if you can do it, I'll play it.

   This game system, as might be expected from an open source project, has generated several active game world design projects. Listed on their website are:

   1. Dural: The fantasy realm that provides the first full-scale test of the WorldForge system.

   2. Relia: Dural-esque world with one major difference: it will be built from scratch by the players themselves.

   3. Dead Center: The first science-fiction implementation of the WorldForge System.

   4. Dodecahedron: An educational application of the WorldForge system.

   5. Project Phoenix: A science-fiction world based on the Star Trek universe.

   6. Underground: A Relia-esque fantasy world involving the exploration, creation and conquest of a new world by players. The game closest to completion is one called Acorn, a simple, proof-of-concept game of pig herding. (No, that's not a mistake, at least on my part.) According to the design notes: "I'm going to resist the urge to suggest a combat-oriented game, since it's too obvious and Hack And Slash is looked down on in this project. ;-) Plus, it's more challenging making the game fun this way." Players buy pigs, fatten them up, then sell them to the butcher (apparently hacking and slashing is acceptable if done by an NPC.) Along the way, they have to contend with Skeletons (which will eat big pigs), Crabs (which will eat small pigs, and Wolves (which prefer to eat sausages, but will eat Crabs if they bump into them, or pigs if they are too hungry). Acorn is scheduled to debut early in 2000.

      The Worldforge Project looks like a fascinating concept, for which I have nothing but the best wishes. However, the sheer scope makes me wonder whether the development team can actually finish a game engine and at least one solid client, plus implement one or more games which achieve WorldForge's stated goals. Game players can only hope that this project will succeed.

      March 14th, 2000
      Douglas Sundseth