One thing I've been noticing with my on-again-off-again development on my BYOND roguelike is an exceedingly larger and larger spiral of complexity. Before I've even gone ahead and fully implemented a system, I'm already writing notes on yet another layer of complexity in the system. (People familiar with me will know that I have a tendency to do that in all of my games, however.) In the roguelike genre, there is generally a rule: you have to think of everything. With such monsters as Nethack, ADOM, and Angband gracing the internet, people look down on your roguelike if it doesn't have everything, even the kitchen sink (Nethack spoiler)!

However, the dungeon crawl genre doesn't have to be about complexity. It has to be about dungeons. The idea is to abstract the gameplay down to its foundations. I'll attempt to showcase them:


Element of Extreme Risk
A dungeon crawl is not a dungeon crawl if there is a chance of never dying for a player of only reasonable skill. Action games and adventure games like The Legend of Zelda are about the story; the player is usually able to quest through the whole game without losing a life, if he's of average skill or better. Not so for dungeon crawls, where only the very best survive.

Dungeon crawls are inherently unrealistic, because the basic premise of a "dungeon" is a labyrinthine, progressive, or otherwise restrictive area in which you have limited room for maneuver and must memorise the layout of the region in order to succeed. I doubt you know the last time you've been in a place that was completely labyrinthine in reality (though some suburbs might count). Thus, the inherent lack of realism of a dungeon crawl allows you to make creative exceptions with other types of realism, like the permanence of death. Though roguelikes themselves almost always mandate permanent death, the broader dungeon crawl genre doesn't need to. Diablo II is a dungeon crawl, yet when you die, all that happens is your corpse slumps to the ground, and a fresh clone of your person appears in town. The only explanation you need? Magic.

The ideal point of balance should be where a player can beat the game, eventually, after investing several hours learning the subtle nuances. But before the player learns those nuances, it's feeding time for the dungeons.


Random Content
A dungeon crawl is more interesting if it has random content, because a hard-coded dungeon is rarely entertaining once you memorise it. Though you have to be sure to include sufficient randomness to engage the user, such that the user can play at least a few days of real time in any given dungeon before the experience becomes complete ennuis, randomness is just one possibility. It is completely possible to make a dungeon crawl with no random content... as long as the experience surrounds dungeon crawling and the maps are detailed enough that a player could never fully explore them in one play through, and the actual treasures earned and monsters encountered at least vary from room to room every play through.

The venerable Gauntlet did not feature a single iota of random content; monsters were purely deterministic.


Exploration and Reward
One of the biggest qualities that define a dungeon crawl is the art of exploring -- venturing into unknown territory and mapping it while discovering fantastic treasures and devices of great power. One great way of doing this is to encourage the independence of certain characters and eliminate the possibility that a single character type can accomplish everything effectively. For instance, if you're designing your dungeons well enough, perhaps there are secret doors that a warrior would always miss but a keen-eyed elf would spot, leading to alternate routes providing treasure or shortcuts around severe danger... but there are also rooms where an elf would stand absolutely no chance of victory due to the sheer number and volume of enemies; the elf would have to pass these areas and their treasure by.

If you can make the player wonder what sorts of areas he could explore if he could start over as another character, you've done your job well.


Whacking
Perhaps the final and most important factor of a dungeon crawl... you need to kill things. Repeatedly. Without prejudice or moral consideration. Whenever the player is not exposed to enemies, he is not gaining equipment and all of his expensive toys are going to waste. Though craft skills and other skills can supplant the character's power, a dungeon crawl absolutely must involve combat... a game where a player can peacefully make a living crafting swords and selling them to NPC adventurers is a medieval life simulation, not a dungeon crawl.


Drive to Continue
Perhaps your character needs to eat. Perhaps your character is being chased by some evil zombie who fears the artifact that your character is attempting to retrieve, but gets emboldened enough to venture down another level towards your character in case your character might never make it. Perhaps you are competing against other adventurers who are trying to take the prize first. Or perhaps there is a looming cloud of doom and an ultimate evil rising over the horizon, and the only way to stop it is to gather up the parts of an artifact of great power before the pestilence arrives.

Whatever the underlying cause, a dungeon crawl must actually encourage the player to proceed. Conservative gameplay eliminates the risk of a dungeon crawl, as the player can simply sit, train, and meditate and bide his time to increase his power instead of whacking everything in sight.


How difficult is it to integrate these qualities into a dungeon crawl? I would say not difficult at all! On that note, I would like to see what people can come up with: I would like to call it the "Pick-Me-Up Dungeon Crawl Challenge". The rules are simple:

1) You have to make a dungeon crawl. Popular sentiment has to agree that your game is actually definable as a "dungeon crawl", "roguelike", or "dungeon romp". RPGs do not count.

2) The player should be able to defeat the first "quest" within fifteen minutes of downloading the game. Gameplay should thus amount to a simple set of options; I would suggest using the arrow keys to move, with bumping into an enemy identifying intent to attack, but an entirely mouse-based system is also possible. If you have access to Diablo II, I would refer you to the Den of Evil, where the player could simply point-and-click his way to victory.

3) The game's source code should be no larger than 20 KB. Note: Code compaction beyond trimming of extraneous whitespace is not legal; code should be human-readable. Source code does not need to be made available; this is just a ballpark figure for your own benefit. It's a soft limit... you can make the game as big as you'd like, but if you find yourself making it a lot larger than 20 KB, you have to wonder what exactly you're doing.


This isn't actually a contest, so there are no prizes, deadlines, or hard restrictions on these rules. If anyone is willing to go the extra mile to make it a contest, I'd be willing to chip in a BYOND membership, but the point of this is simply to encourage a simple type of dungeon crawl game and to see what the fruits of this endeavour are.

Any takers?