Interfacing with Computer Narratives: Literary Possibilities for Interactive Fiction
Other attempts at interface design range from combinations of the graphical and textual interface, which promise better usability but no essential improvement in literary potential, to controller-specific interfaces for highly engaging environments. Unfortunately, development of fictional 'virtual reality' works has centered on spectacular action games, not the creation of narrative environments, and the interfaces to such 'virtual realities' usually are focused on the one means the interactor has of influencing the environment - almost always a weapon of some sort. Despite the narrative vacuity of the environments of such games as Doom and 'Dactyl Madness, their interfaces are worth examining, because their success in conveying limited information in the context of massive sensory engagement may hold lessons for more communicative interfaces. Finally, some products from Activision, including the interactive fiction Portal, abandon the attempts of other interfaces to make the computer transparent and instead make the software interface an explicit component in the narrative.
One way to bring the advantages of different interfaces to an interactive fiction is to use both a text parser and a graphical interface. Most interactive fictions using this style of combinational interface only require textual input - all of the characters' possible actions can be performed using typed commands. But there are added "shortcuts" for common actions, so that a series of mouse commands (e.g. selecting an item from a menu, dragging an object from one window to another) can take the place of a typed command. The Infocom product Arthur: The Quest for Excalibur, which does not have a mouse interface, has one of the slightest expansions on the traditional text interface; it simply uses the standard Interlogic parser with six added options (time-saving commands to give a description of the current area, the possessions of the character, etc.) that can be activated by function keys on the keyboard.
In fictions by Magnetic Scrolls and in The Multidimensional Thief, there is an extensive interface combining a text parser (though which all interaction can take place, if so desired) with a graphical interface. In The Multidimensional Thief this interface consists of icons that, when clicked on, produce the appropriate text for a particular command (e.g. "examine," "drop") at the command prompt. In Fish! and other Magnetic Scrolls products there is an entire windowing system, complete with the ability to move and resize different windows (which contain a map of the interactor's progress, the interactor's possessions represented as icons, the items in the current location represented as icons, etc.) and even features like a modifiable background. In these systems the consistency of a text interface is maintained, since each command issued with the mouse has a uniform textual translation. There may be numerous alternate ways, though, of performing many commonly typed commands. For instance, instead of typing "Take the gold disc," the interactor can simply click on the gold disc icon (in the "items in room" window) and drag it to the "inventory" window (in which icons representing the character's possessions are shown). Or the interactor can select "Take" from a pull-down menu and choose "gold disc" from a submenu.
While a complete environment of the sort Fish! provides offers the interactor a great deal of control over the display of various pieces of information on the screen and simplifies the entering of common commands, it has noticeable difficulties. There is more overhead involved in beginning the interactive fiction, for instance, since the interactor must learn the particular conventions of the environment before using the graphical features of the system. Also, the semantic limitations of a graphical interface has can manifest themselves in a combined interface. For instance, if the interactor clicks on an item's icon from the inventory window and drags this item, not to a blank spot in the items in room window, but to an icon representing another character, does this have the semantics of "drop <item>" or of "give <item> to <person>"?
Some immersive computer experiences, like the three-dimensional, first-person game Doom, have elaborate graphics and respond profoundly to input from a player, but are focused on a simple interaction and so use a simple interface. In the case of Doom, the simple interaction is that the player controls a character who moves through an environment in a standing position. The character can run, walk, turn and sidestep. The character automatically picks up useful items, as many as he can carry. There are only two ways of interacting with the environment: press one key (or mouse or joystick button) to use a weapon (the player can select any one in the character's possession to hold at the ready) or press another to perform the nonviolent actions of opening doors and throwing switches. Although deeply different from interactive fictions like LucasArt's Loom and Myst (in these, neither the interactor's character nor any other character can die; in Doom, the object is to kill every creature encountered with weapons ranging from a chainsaw to a rocket launcher) Doom shares with these fictions a simplicity in interface aimed at making the experience more immersive by making the interaction require less attention. Doom is not much of a narrative
and certainly provides the player more the feeling of being at a shooting gallery than of participating in or helping to craft a literary experience. Yet it demonstrates the engaging power of an interface that allows for continuous interaction, where the character's world constantly responds to the actions of the interactor.
Another approach to making the interface seem natural is to make the computer an explicit part of the interaction. Activision's games Hacker and Hacker II employ this technique, as does their interactive fiction Portal and Infocom's A Mind Forever Voyaging, in which the main character is actually a sentient computer who enters simulated futures in order to test the effects of a proposed policy. In these works, the premise is that a computer, in the world of the fiction, is itself the interface between the main character and the fictional world. Therefore, it is easy and natural for the interactor to assume the role of the main character without having to suspend any disbelief with regard to the interface.
The way this technique fits the interactor into the story gives it an outstanding advantage. All constraints on the interactor's actions, even the extrinsic constraints (those that "have to do with the context of the person as operator of the system" - Laurel 103) are made to be part of the story; for any computer-related action (e.g. turning the computer off, hitting the reset switch) "[w]e need not shift gears to consider the effect ... upon the computer or the game program." (Laurel 103-104) If "what makes a simulation seem real is the degree to which a user's interaction with it is like his or her interaction with the real world," (Kelley 52) the ruse of using the computer as a bridge between the real and narrative worlds makes for the perfect simulation.
This technique does have disadvantages, however, which include the difficulty of learning a traditional style of computer interface in order to communicate with the fiction. In both Portal and A Mind Forever Voyaging, for instance, the interactor has to access files and learn to navigate a directory structure using techniques specified in the manual. The Portal documentation describes this task as "one of the more difficult but essential things" the software user has to learn to do, and follows this warning with a list of five steps. Also, this method of interface restricts the characteristics of the narrative that can be told - they must take place through a computer in the world of the story, so all narratives told are constrained to have this premise.
While conjoining different types of interfaces may make software easier to use, the ways in which this has been done as of yet have usually afforded little or no additional literary possibility. The one way that combination text/graphical interfaces may enhance rather than diminish the engagement of the interactor by fitting him or her consistently into the narrative is when they serve to make the computer mimic a computer in the narrative, and thus become a total bridge to the story. Even then, if the computer-within-a-computer's interface is complex or unwieldy, the interaction may not be pleasant.