Game Design Dialectic: Dwarf Fortress and Goblin Camp

This is only the beginning of a story, but it could prove to be a very interesting story if it bears out. I think it already contains instructive lessons for game development and design.

On the left, Dwarf Fortress. On the right, Goblin Camp.

I hope you know about Dwarf Fortress, the very complex roguelike-lookinglike fantasy world sim / citybuilder. From a development perspective,  DF is a very long-running obsessive project coded by one guy, Tarn Adams, who makes more money than I do (not difficult) entirely by donations from his fans. I admire Tarn’s goals and his creative freedom which lets him indulge his whims – I wish I could do that. I even had fun playing some Dwarf Fortress until I explored most of what there was to explore. It was sweet while it lasted, but I grew tired with the tedium of a very rough user interface and tedious gameplay.

This brings me to a common criticism of Dwarf Fortress: development over the last year, year and a half has focused on revising very low level details about the game’s simulation of how materials interact, particularly how creatures bodies are built with layers of bone and muscle and hair, what properties these each have, etc. It is true that part of the charm of Dwarf Fortress is about the ridiculous level of detail. But it has been over a year and I still have to press a series of awkward keyboard shortcuts to build things, I need to hand-designate every square of ore to be mined, I need to tell each workshop exactly what to make. Frankly, the user interface achieves mind-blowing levels of confusion and  unnecessarily repeated actions which lead to a sense of tedium and frustration that overwhelms my interest in continuing the play the game – so I don’t. Many people don’t even manage to fully learn to use the interface (or don’t want to)  due to – I’ll say it – how bad it is. And further, most new players are overcome by the sheer detail and volume of information that needs to be processed and managed by hand: To speak for my own experience (to those of the DF community that hold in high regard their ability to manage an extraordinarily complex game), it’s not that I am incapable of running a complex Dwarf Fortress game, it’s just that I don’t want to because it’s boring to have to hand-tweak every little thing to keep the place running, and worse still, it actually hurts my hands to press the same keys again and again and again as is required.

I love what this game could be and reading the development page it is full of admirable, sky-high aspirations. But I can’t bring myself to play it. It’s a beautiful idea but an extremely flawed game.

To get the Dwarf Fortress Experience, you’re better off reading the stories people write about their games in forums, eg. the classic Boatmurdered. This removes the frustration of playing the game itself and gives you the high points of amusement at the absurd details and situations which arise during gameplay (which inspired a good deal of the silliness we have in Dungeons of Dredmor, I should add).

And then, if the post’s timestamp is correct, just two days ago on July 14, Ilkka Halila announced Goblin Camp in the SomethingAwful forums.

Now things are getting interesting.

Goblin Camp looks like Dwarf Fortress, uses the same ASCII-graphics, and starts from a foundation of the same sort of gameplay built upon semi-autonomous agents collecting and processing resources in a world build of tiles, but it makes several important departures in terms of project development and design philosophy.

1. The code of Goblin Camp is released open source. In interviews, Tarn Adams has expressed concern about releasing DF’s code because he could lose control over the focus of the project, lose financial support, and he is not interested in supporting other people modifying (and breaking) the code. Goblin Camp has already been extended by coders other than Ilkka – if the initial interest maintains its present momentum, the game could develop at an extraordinarily rapid pace. I must observe though that GC’s appeal is somewhat cannibalistic on DF’s – It is frustrated DF fans that are excited about CG.
2. Goblin Camp streamlines play. For example, there is a central depository of craft goods in GC which the player gives orders like “Maintain a stock of 500 wood planks”. Workers are automatically assigned tasks to fulfill this requests, they are sent out in the woods to cut logs which are returned to a carpenters shop to be processed. In DF, one would have to designate a single worker as a lumberjack, scroll out into the map, designate an area of trees for chopping (using the keyboard, by the way), then queue tasks in a carpenters shop by-hand. When designated trees run out, the player has to re-designate more trees – and the player is not told when they run out of designated trees. GC handles the minutiae for the player, reducing the hand-interaction required from perhaps 10 actions to one action, at least. To be frank, this design ethos of streamlining interaction blows DF out of the water in terms of playability already (Dwarf Fortress was released four years ago, by the way).
3. Goblin Camp abstracts details. While DF has spent a year of development time simulating the material properties required to properly model the penetration of an iron bolt through leather armor, flesh, and bone, as appropriate to the details of a given creature’s anatomy, GC was coded in its entirety in two months and uses simple die rolls for attack skill and damage. The resulting playability of each game’s combat is not a radically different experience: guys swarm each other and people get chopped to bits. The idea of abstracting small details to implement fun features more quickly appears to lay behind every aspect of GC’s development. Further, the mod-friendly and open source nature of the game allows other people to fill in small details at their whim while the primary developer concentrates on the more general framework of gameplay.

Goblin Camp was made, to paraphrase Ilkka, because he loved the sort of game that is Dwarf Fortress but he is impatient and wants to play the game DF could be, that he wants DF to be, now rather than waiting for Tarn to add certain features to Dwarf Fortress – if he ever does. A game like Goblin Camp was bound to happen in response to Dwarf Fortress, and I think Tarn and many others knew it would come. I’m pretty sure similar attempts have been made (there was an Elf Forest joke-game, I believe), but none seem to have really taken off. Maybe Goblin Camp will.

Goblin Camp, as a game and an approach to development, is a critical response to weaknesses in the game and development of Dwarf Fortress. Maybe Tarn will have to react to Goblin Camp out of a need to save his source of income, maybe he will re-focus on making Dwarf Fortress a playable game rather than a complex simulation lost in it’s own obsessive detail, accessible only to an extremely dedicated few. It’s like Josh Petrie’s advice to beginning game developers: “Write Games, Not Engines” mixed with the ethos and methodology of Open Source software, Wiki-style content, and the absurd power of the internet.

I can’t wait to see what happens next.

Tags: Dwarf Fortress, game design, Goblin Camp, open source