Xyzzymposium 2014: Joey Jones on Best Implementation

Joey Jones is a writer of text games including Andromeda Dreaming and Danse Nocturne. Co-author of the weird-fiction puzzler Sub Rosa in IF Comp 2015, he is currently working on a long-form ChoiceScript game set in the 18th century underworld.

The Best Implementation finalists were Hunger Daemon, With Those We Love Alive and Hadean Lands.

Hunger Daemon (Sean M. Shore)

Sean M. Shore’s Hunger Daemon was clearly nominated for the Implementation category for three of its virtues: its robustness, its reactivity, and its strong maintenance of game-flow. While it is not an overly long game, it has a solidity to it that is a result of some considerable polish.

Robustness is the lowest bar a game must climb to be considered competently implemented. By robustness I mean, the game must not be obviously buggy, exits should be clearly listed, there should be no walking deaths (this is not 1990, dead ends should be obvious and reversible in a turn or two). A reasonable range of synonyms are accepted. As ever, there are some formulations of commands that probably should work but don’t, but most often the obvious command is the correct one.

There are different kinds of reactivity that we appreciate in a parser game, and Hunger Daemon satisfies all of them to different degrees. The promise of the parser is that it might respond intelligently to whatever we type: but most of all we want it to react to the things we think the game is telling us should work. The actions required along the core path often give a custom response. There are plenty of amusing things that are implemented that lend extra depth. Crucially, these additions don’t require wholly unmotivated acts, but rather flow out of natural interactions the player might come up with, with the items to hand whether that be looking up Lovecraftian lore in the library or playing around with the rubber glove.

Most impressively, is the maintenance of flow in the game. At every point in the game, the player knows what they are currently trying to achieve and has enough information to begin formulating a plan towards that goal. The goals start out small and easily manageable (like getting a sandwich) before growing in complexity. The progression of the plot and the actions it requires of you proceed logically if not sensibly. The pizza-smear puzzle for instance was fair and clearly hinted but very silly. This is all fitting for a comedy game. It is fine (and often more fun) for a game to require the player to do bizarre acts, so long as acts are properly motivated and consistent with the tone of the rest of the game.

Granularity of action is also competently handled: the player rarely has to take many small steps unnecessarily. Driving the car just requires entering it and going. The threaded conversation is well prompted. Whenever a scene requires a change of location, the player is automatically moved along (such as when they go down to the basement with Jeff or into the synagogue after giving the ticket). The design principle here is that whenever there is only one action the player should perform and the action is not dramatically significant, it should be done automatically.

With Those We Love Alive (Porpentine & Brenda Neotenomie)

In With Those We Love Alive, Porpentine is keen to stress at the beginning that there are no wrong choices. This is a promise that is kept through the implementation. The game presents a consistent world model that the player can explore over numerous game days. On top of this cyclical model, there are plot scenes that trigger when the player sleeps enough times or does some fairly obvious action (usually smithing). There is no branching: the permanent choices the player can make are perceptual rather than mechanical. This works most strongly in the beginning part of the game where there are more unique city and palace scenes, as in the second half the game progress can be made almost exclusively by sleeping.

Exploration of the palace and small city area is a core part of the game experience, with the player often rewarded with a range of new descriptions upon revisiting old locations. This is in contrast to most hyper-text games, which are arranged chronologically. In a typical twine, the player will have their full range of current choices in front of them on the same page at any given moment. The game is more like a traditional text adventure only in the respect that player options are spread out over a series of locations and some require specific action sequences to proceed.

The range of locations is broad enough to give a sense of exploration, but narrow enough that it does not become too great a chore to regularly check each place for something new. The guided meditation (with customisable breathing times) was a particularly nice touch, encouraging a slower, more contemplative play style. Indeed, as with most Porpentine games, the delivery of the plot is abstract enough that a slow reading (or a rereading) is necessary to have any hope of forming a strong conception of what might be taking place. Moreover, the player is encouraged to strongly relate or self-identify with the protagonist role they embody in the game: the act of physically drawing on yourself works to blur the player/character distinction.

Porpentine has a distinct vision in mind, and she deftly twists the Twine engine to her needs. Brenda Neotenomie’s soundscape is evocative but non-intrusive and competently looped. The screen effects complement the Porpentinean slime femme aesthetic (a lot of hot pink as usual). The prompts to draw sigils (the notable mechanic of game) are made clearly on a fresh page each time, ensuring the instructions are not overlooked. The sigils themselves also have a role to play in reinforcing the structure of the game: distinct episodes in the game are clearly marked out on the player’s skin, giving a much needed sense of progress for a game formed out of repeating similar days.

Hadean Lands  (Andrew Plotkin)

Andrew Plotkin set out to create a big system-driven puzzle game and in this aim he succeeded.

Hadean Lands is big: there’s a lot of puzzle content implemented, with hours of gameplay. You wouldn’t praise a novel or a film for its length, but in commercial games there’s an expectation that some aspect of its value for money is linked to the length of the experience. Length is only a merit insofar as the content justifies it, and fortunately the puzzles remain compelling throughout.

The same structure underlies all the puzzles in the game, making it a very focused experience. There’s expanding realm of alchemical recipes which have to be learned and tweaked and improvised on and chained together to open locked areas which unlock new recipes and new areas to unlock. Despite what you might initially expect, you don’t actually learn a magic system itself: the principles of alchemy that underlie the rituals in the fiction as presented are purposefully obscure. The puzzles are about collecting and altering recipes. The entire design of the game is relentlessly focused around this core system, sometimes to the expense of mimesis. For instance, there is no compelling in-fiction reason why there should be singular sheets of paper describing the exact ritual component you need to learn lying in the corridors or inside locked cabinets. This kind of artificiality is incredibly common in other forms of video game: finding journal pages littered around the place is commonplace in computer roleplaying games, survival horror, and even hidden object games. In contemporary interactive fiction we usually expect a consistent fictional world (rather than just a straightforward puzzle-scape). In other respects, Hadean Lands upholds a figleaf primacy of setting over puzzles: the ingredients in the game are usually in reasonably sensible places and the alchemy itself is thoroughly embedded in the fictional world.

The reason why Hadean Lands won the nomination is not due its impressive size or the totality of its puzzle mechanic. It won because it recognises it is a vast game and makes every reasonable accommodation to the player to make sure the core experience (solving puzzles) is a smooth as possible. The recall command is indispensable and saves time wasted pouring over previously examined papers. The reset mechanic frees the player up to experiment and neatly avoids unwinnable situations in a game built around finite ingredients and (mostly) one-way processes. There are further accommodations like a go to command for heading to previously experienced rooms and items. These various easements are well communicated to the player. The jewel in the crown is of course the ability to perform any previous multi-step ritual in a single command once it’s been performed once. This effectively allows the player to shift the granularity of their actions from performing individual parts of a ritual to performing multiple rituals.

One aspect of it being so large a game sets a clear contrast with minutely crafted small games like Hunger Daemon. So much effort has clearly been put into making the core systems repeatable and robust. This scalability of game responses comes at the expense of bespoke responses. To give an example: when you first take the draught preventing hot things from burning your hands, and you take an item from the furnace for the first time, the game still uses the default response ‘Taken’. The descriptions in the game are short and functional, the heavy lifting is done by the player’s imagination. In shorter games, the act of plucking something from a fireplace would be significant and would warrant its own custom first time response. But here, there are so many extraordinary things you are doing, action success or failure is mostly described briefly and consistently. Contextual, specific or one time only responses are mostly forgone to enable greater flow. In a game with so many parts (and with the undoubted difficulty in testing so huge a project) this is perhaps a necessary sacrifice.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>