Where Text Adventure Games Live.
What Makes a Good Text Adventure Game?
In judging the quality of any computer software, you should keep one point in mind: publishers of this kind of software suffer a theft, or piracy, rate that would frighten any store owner. It's hard for someone to justify large development costs when he knows that a large percentage of the people who take advantages of his efforts won't be paying him anything. Under the circumstances, we should be grateful that people keep producing new computer software at all, instead of moving into a more financially secure area. [ All of this is perfectly true and valid today. - Editor ]
Evaluating text adventure games is trickier than evaluating most other kinds of software. A strictly practical program, such as an accounting package or word processor, is judged by how easily it gets the job done. Even with an ordinary computer game. the critic expects the commands or controls to be as easy to use as possible. And in either case, the documentation is supposed to be complete; if you aren't told how to print a letter or shoot the invaders, you have every reason to complain.
With a text adventure game, on the other hand, figuring how the program works is part of playing the game. You don't know what all the commands are or whether they'll always work. The text adventure game might even go into a special mode without warning, requiring you to use an entirely different set of commands until you get back into the normal situation.
Even so, there are definite standards for a good text adventure game, just as there are standards for a good novel. First of all, the difficulties in a text adventure game should lie in the situation it presents, not in getting the interactive fiction title to understand what you mean. There is nothing more frustrating than knowing exactly what you want to do but being unable to convey your intent to the program. To avoid this frustration, the program should have a powerful command parser, a large vocabulary, and understandable error messages. As the authors of Zork put it, "the game should simulate the real world sufficiently well so that the player is able to spend most of his time solving the problems rather than solving the program."
In a text adventure game with a good parser, you can express your actions directly. In a text adventure game that accepts only two-word commands, it is difficult to express an action like putting a coal into a fire. The text adventure game must either arbitrarily assume that DROP COAL puts it into the fire or ask for a second command to specify the destination of the coal. Being able to type DROP COAL IN FIRE is much more convenient.
A large vocabulary can increase both the number of available actions and the number of ways the player can express those actions. After all, kicking an ogre might do something useful; why not let the player try it? Having a small vocabulary encourages playing the program rather than playing the text adventure game, since it lets the player zero in on the correct actions just by noticing that certain words are in the game's vocabulary. If the vocabulary is large, the mere fact that a word is acceptable doesn't mean that it's necessary to solving the puzzle.
Good error handling makes it easier to figure out just where the program's limits lie. If the response is simply "I don't understand," you can't tell whether one of the words you used isn't in the program's vocabulary (or which word) or whether the way you combined the words made no sense to it. This means you have to rephrase a command several times before you can figure out why you weren't getting through to the text adventure game.
The text adventure game's responses should be clear and grammatical. Exits and objects should be clearly noted, unless they are supposed to be difficult to find. Descriptions should convey a sense of your actually being there, just like descriptive passages in any other work of fiction.
They should also follow the rules of the language on grammar and spelling. Sadly, a number of text adventure game publishers have allowed errors to slip into their games -- errors that even the sloppiest magazine editor would catch at a glance. The trouble is that nearly all text adventure game authors are programmers first and writers second.
Responses shouldn’t be misleading. It’s acceptable for the text adventure game to misdirect the player, to disguise the nature of things, but it shouldn’t produce a response that flatly contradicts the way the adventure world is intended to be. That may sound obvious, but it’s tricky to accomplish because of all the possible actions that a player can attempt. The first version of Zork, for instance, responded to attempts to LIFT the rug by saying that fiddling with it wasn’t useful, when in fact there was something hidden under it. The program handled the intended verb, MOVE, correctly, but the writers just hadn’t considered the possibility that someone would try LIFT.
If the adventure game includes graphics, they should do more than just illustrate what the text says. They should supply additional facts or at least expand on the ones the text mentions. Many graphic adventures do this by including items in the graphics but not mentioning them in the list of items present. In many cases, the graphics can also help to clarify the geographic layout. Some adventures, however, add graphics only as an afterthought. In those cases, they only slow things down and leave less room on the screen for text.
The speed at which the text adventure game runs is important. If you have to wait ten or fifteen seconds after each command, it’s easy to get bored with the adventure. (editor's note: twenty years ago, game response time was a major concern. Today such problems are non-existent.) This can be a problem if the adventure game is constantly bringing in graphics or text overlays from the disk, especially on systems with slow disk drives. The language in which the adventure is written can affect its speed; any program that runs under a BASIC interpreter is going to be slower than an equivalent machine language program.
Finally, the quality of the theme, story (if any), and associated puzzles is all –important. Not all adventure programs spin out a story; some just present an assortment of puzzles in a common setting. But in either case the situation should hold the players interest, or the program degenerates into a dry exercise in problem solving. In adventure, much of the pleasure comes from the sense of going deeper and deeper into the cave and discovering unexpected passages. Monsters and treasures aside, it conveys the feeling of exploring a spectacular area. In Witness and other mystery adventures, it comes from interacting with the characters and discovering their motives.
To have this quality, the text adventure game should fit together as a self-consistent world. This means that the puzzles should play fair, but it includes much more. Elements that are foreign to the game’s milieu shouldn’t intrude. If the setting is based on science fiction, a puzzle shouldn’t arbitrarily throw in magical elements. If it follows the world of Greek mythology, Wotan and Brunhilde shouldn’t appear without good reason.
Deliberately mixing milieus can be the basis of an amusing story, as in Paul Anderson’s novel, Operation Chaos, which gleefully combines magic and science fiction. But the mixing should be international and well should be intentional and well prepared, or the player will get the feeling that the author is just trying to confuse him.
The difficulties that the player’s character encounters should make sense in terms of the situation. It makes sense, for example, that an elevator button might be gimmicked so that trespassers will have trouble getting to restricted areas. However, it doesn’t make sense to have the elevator button at the other end of the building. The player should be able to recognize patterns and follow them through logically. If he’s reduced to trying every possible action at random, then the adventure (or at least that particular puzzle) is a failure.
The adventure game should also create the illusion that the player’s character is a clever person who anticipates difficulties, senses dangers, and deals with them successfully on his way to victory. When the player has finally solved the adventure and can play it from beginning to triumphant end, the actions he takes along the way should seem reasonable to someone looking over his shoulder.
This means that every danger should be preceded by signs that, at least in hindsight, can be recognized as a warning. It means that the character shouldn’t have to take actions that are motivated only by information he gained in a “previous life.”
Some text adventure games give the character a limited number of resurrections; this makes it more likely that he will learn from fatal mistakes. However, such an approach makes sense only in a capricious world like that of Adventure or Zork, or in one in which science, magic, or mythology permits a return to life, such as Lords of Karma. Adventures with more natural environments don’t have this recourse.
Consider, for example, the case of a room full of poison gas. The way to get through the room is to give command HOLD BREATH before entering. If the character has no reason for holding his breath except that he choked to death in that room the last time he played, his actions become illogical.
However, things can be kept reasonable if the description of the previous room states that wisps of green mist are coming from under the door. Giving the command SMELL MIST might elicit a stronger warning, and then it would make sense that the character should take precautions. The point isn’t that a really good player should be able to get through the adventure on the first try, but that the character should stay within the bounds of the game’s reality.
So what is a good text adventure game? A good adventure is one in which the player can vicariously experience the sense of being the character in a story. To the greatest extent possible, the player shouldn’t have to worry about the right phrasing. He should be able to feel that his opposition is the situation itself, rather than programmer who’s trying to confuse him. The adventures currently on the market vary tremendously in how closely they approach this ideal, and even the best have a long way to go.