Guest Editorial:
Game Interface Design

By Harry "Het3" Teasley

hat is a game? A game is an environment where you make decisions that change the environment, and making these decisions is a Fun Thing to do. The game’s interface is the language you use to speak to the game. It should be obvious that the interface plays a role of monumental importance. It should surprise you to learn that the interface is often very much overlooked by game designers, but unfortunately, if you are an experienced game player, it likely doesn’t surprise you, as there are many regrettable game interfaces out there.

Why is this? Well, there are a number of reasons, but I would say the number one reason is that interface designs often seem to grow out of the game design, not in tandem with it. It is apparently easy to not outline the decisions the player has to make in a game, and then to not prioritize them appropriately. The most important decisions you as a player make should be the easiest ones in the interface, and by "important," I mean something different from "vital to the plot of the game." I mean "important" in the sense of "a decision that the player will make constantly, all the time, several times a minute under normal playing conditions."

Games are usually made up of several big, and hundreds of small decisions: the big ones are "rescue the princess" or "score the touchdown", and these sorts of decisions are achieved usually by making many little ones, like "train this infantry unit" or "open the door." The player makes a bunch of small decisions, and they make them over and over, hundreds of them perhaps, between the major decisions. The interface for making the small ones should be as invisible as possible.

Let me give examples, drawn from four different games, of a simple action: a character picking up an item.

In Quake, if you see an item, you walk over to it and pick it up automatically if you can (if you’re filled up on that particular resource, it stays on the ground.) Simple, but also not very involving.

In Grim Fandango you walk over to it and hit a generic "use" key, or, in LucasArts’ previous adventures, you would click on the item with your mouse. If the item can be picked up, you will pick it up. If it can be used, you will use it. Also simple, and allows a little more control over what your character does.

In Ultima Online, things lying on the ground can be clicked on and dragged to your character, either to the backpack (the default choice) or, if the character paperdoll is open, to a specific location on the character. Since it is quite relevant from a gameplay perspective what the item’s location is when in the player’s possession, their system makes placing the item on the player a deliberate action (the easiest decision to make is to put the item in your backpack, which is also the strategically neutral decision.) This is a more complicated method of resource acquisition, and takes more thought on the part of the player.

In Trespasser, if you see an item on the ground, you walk up to it, use the mouse to look at the item, hold the "extend arm" key down to extend your arm, hold the "pick up item" key to pick up any object near your hand, hold the "crouch" button down to bend down to get the item close enough to your hand, and then use the mouse to aim your hand at the item. Some items you may wish to pick up in a specific manner, so you would also want to aim you hand to a particular location on the item, perhaps using your "rotate wrist" and/or "rotate arm" keys in the process. In case you’ve lost count, that’s up to five keys being held down at the same time while aiming with the mouse: an almost ultimate level of control, but already past the point of enjoyability and well into the realm of cumbersome.

I think this pretty clearly indicates that in even something as simple as picking up an item, you can introduce as many or as few decision making opportunities as possible. It’s up to the designer to decide which decisions are fun to make, and which are appropriate for the game to assume for you. In the above examples, the first three games have appropriate levels of interaction with regard to picking up an item: in Quake, I don’t want to think about whether or not I want to pick up the rocker launcher. Of course I want to pick up the rocket launcher, don’t be absurd. In Grim Fandango, I don’t want to pick things up automatically, because finding things and determining their importance is half the game: take that decision away from me and you may as well go the rest of the way and make it a movie. In Ultima Online, picking up an item has direct bearing on my character’s abilities, and whether or not an item is worn, wielded, or stowed in my pack is also of great significance. The interface therefore introduces enough complexity into the decision sequence that it is harder to make unintentional or disadvantageous choices.

However, Trespasser goes too far in presenting me with options: I’m already at the point of expressing interest in acquiring the resource, must I jump through hoops that should be obvious decisions? Do I want to say, "Pick up this item," yet still have ample opportunity to miss the item? If that’s the way to get interactivity, my next game will have "inhale" and "exhale" buttons. You’ll all pay then, muahaha.

So clearly, we have one obvious conclusion to draw:

Don’t mistake "navigating the interface" for "playing the game."

"Navigating the interface" is the process of implementing the decisions you make when you are "playing the game." Playing is done in your head, navigating is done with a keyboard, mouse, or joystick. Games must balance both aspects, and should choose an interface that fits the style of gameplay.

Irrespective of gameplay style, however, there are still certain decisions that are never fun to make, and among those are "having to make five choices when one would do." No matter what the game is, if I often make a decision sequence that never varies, I don’t want to have to explicitly make those same choices every time. Suddenly I want a scripting language, and on some level, I actually want to not play the game. I’m wishing the game would play itself for me in this respect, and that’s a bad thing to start thinking.

There’s a second conclusion that can be drawn:

Obvious decisions are not typically fun decisions to make. Don’t repeatedly force the player to explicitly make obvious decisions.

Exciting, world-altering decisions are the ones the player wants to be making, not mundane, boring decisions they grudgingly make in order to get to the fun ones again. Common activities should require minimal explicit input from the player, while special-case decisions or difficult puzzles are areas where you can get them to work through the interface extensively.

A simple example of a brilliant minimal interface is mouselook in Quake, to be compared to earlier interfaces for doing the same thing, like in Heretic. Heretic never became the deathmatch game of choice for me even though it seemed in many aspects to improve on Doom. The interface of looking up and down became a sticking point for me, because I found I couldn’t have fun in situations where I had to fight well but also utilize look up, look down, and center view buttons: looking around should be easy, not something to think about at the same level as thinking about strategy, or choosing weapons. When I first saw Quake in its early stages, the mouselook was a revelation. It seemed such an obvious solution, and the years of experience with it make it hard to think about a different interface for looking up and down. It expresses best what good interface does: makes itself as invisible as possible to the player, so the player can get on with playing the game.

Once you state the basic interface guidelines clearly, they sound like truisms, but finding games that have cumbersome interfaces is as easy as a trip to the software store. I think that these rules, by sounding so incredibly "obvious," disguise themselves as "easy" as well, which is why they become deceptively simple to get wrong.

To finish, let me share with you my favorite game interface decision ever. Years ago, my friends and I loved to play the paper RPG Traveler. A fine game in many respects, and like all RPGs, it had a fairly detailed character generation process. You created your basic statistics, and then rolled dice to see what tours of duty your character served, to determine what skills you would start the game with. Some tours, which generated the best rewards of skills or items, also involved the most danger. It was the only game I’ve ever encountered where your character could be killed before you ever started to play. Now that’s an efficient interface!

  - Harry Teasley is a game designer for Valve Software, the developers of Half-Life.