"The movie, by sheer speeding up the
mechanical, carried us from the world of sequence and connections
into the world of creative configuration and structure. The message
of the movie medium is that of transition from linear connections
The aesthetic history of media can be described on the basis of a drift towards greater realism for improved immersion of the viewer: with images becoming more detailed and spatial due to the introduction of perspective, with photography as means of mechanically reproducing specific views, with pictures that began to move, later even to talk and take on natural color. On the other hand, there have been setbacks too: for some reason, neither 3-D movies nor Smell-O-Vision really worked. But what about the one-way relationship between media and audience as the crucial obstacle for realistic media? Even a high-resolution Imax movie will not give the viewer feedback the way "real reality" does so easily. So wasn't it time that media began to show some interactivity in order to become a serious surrogate for reality?
The idea of interactive media has not only to do with the demand for realistic representation but also with the fear of being controlled by media (which people would then call propaganda). In certain early theories of cinema, the main concern was to tell a story effectively and according to the inner rules of the medium. Efficiency in this case meant first of all to restrain the reactions of the viewer. After World War II and the experience of fascism, this idea became unpopular. One of the first theorists to develop a humanist approach to cinema was André Bazin. As a means of counteracting manipulation through editing, Bazin favored the deep-focus photography he found in the work of Orson Welles and others. Instead of combining a series of different views, Welles showed a whole room in which the viewer could decide what he wanted to concentrate on, with the movie thereby being completed in the head of the individual viewer. Though this structure wasn't unusual in painting, theater or literature, for the film medium this kind of "open artwork" can be regarded as a first step towards audience engagement.
The structure of a traditional movie looks like a plain line with a certain amount of little dots representing plot points. Plot points are important events that change the narrative situation. The popular notion of what an interactive movie should be is about making these points nodes, so that the straight line turns into a tree with forking paths. The important decisions in a story are no longer made by its hero but by the viewer himself. That means a movie would be interrupted from time to time for the viewers to choose among two or more possibilities of how the story goes on. The traditional model of viewer identification with a central character turns into a unity of viewer and hero: we can call it the kiss-it-or-shoot-it-model. A first version of this idea was the "Kino-Automat" shown at the Expo 1967 in Montreal: Radusz Cincera's One Man and his World was stopped several times to give the audience the possibility to decide how the film should continue. 
Already for economic reasons, this model seems rather unsuitable, since it not only increases the amount of final footage required, but also limits distribution to specially equipped cinemas. Considering the growing number of Imax cinemas, it is surprising that the Kino-Automat remained a curiosity in cinema history. One reason might be that the apparently motionless cinema audience is actually quite active. The resolution of a 35mm film, compared with that of television or video, can occupy a much bigger part of the viewer's visual field without revealing its material. In the cinema, the viewer's eyes are constantly moving in order to grasp the entire screen, whereas the monitor fixes the view and finally sends the brain to sleep. In order to avoid that problem, TV and computer programs require us to constantly act - be it to zap, phone, write, fire or go shopping.
The monitor appears to be the perfect media for direct viewer-user feedback, and the computer game may be regarded as one of its most demanding programs. Economically, the game outdoes interactive prerecorded film footage because it uses audiovisual elements with an extremely flexible structure that can even, in the case of 3-D data, be viewed from all sides. But games can hardly be seen as narrative, although some are placed inside a narrative frame in order to emotionalize the action. A number of oppositions exist between the forms of narration and game. Stories first of all evolve out of protagonists who act according to their specific characters. But a protagonist who is solely dependent on the player's will cannot have a soul, even if the game design tries to display one. A figure like Lara Croft in the Tomb Raider series first obtains identity outside the game - via marketing, fan sites and performances in other media. Due to its high level of abstraction, the game no longer needs the human or humanoid characters essential to any kind of story. Besides, narration works due to the different speed of what the structuralists called fabula (the narrated content) and syuzhet (the narration itself). The entire life of a person can be told on four hundred pages, two hours or ten minutes and even a movie like Hitchcock's Rope (1948), which avoids any visible cuts, secretly speeds up the syuzhet. Games, by contrast, depend on simultaneity of the dynamic content and its representation. The most successful examples, for example Doom or Quake, meanwhile lack any kind of plot. To have no story but plenty of high-end graphics with constantly new effects has become a mark of quality.
Jesper Juul describes computer games as a two-layer composition of material and program in which the latter gives a meaningful structure to the game's objects (texts, graphics, sounds, and so forth). This model corresponds to Lev Manovich's description of the database as "a new way to structure our experience of ourselves and our world," and therefore as the contemporary counterpart to the traditional form of narration. The database became an expression of the variability of new media: entries can be modified, added and deleted without consequences for the entity of the database; the user is granted random access to different kinds of multimedia objects. In the simplest case this may be by engaging in the linearity of a movie by jumping to different positions. The rise of the database changes the relation between paradigm and syntagm as described by de Saussure and Barthes. Traditionally, the paradigm becomes visible as an ordered collection of signs only through the syntagm as a meaningful arrangement of selected signs. When narration is superseded by the database, the paradigm becomes real and the syntagm virtual. In comparison to narration with its one dimensional-structure of time, database information is structured in multiple dimensions. Time is usually too abstract and dysfunctional as an interface to databases, while a spatial construction (be it a house, a landscape or something abstract) seems more appropriate. We can say that the form of narration is temporal and authoritarian because the author has organized the information in advance. The form or interface of the database is spatial and interactive or semi-authoritarian.
One of the first applications permitting navigation through multimedia data was the Aspen Moviemap made in 1978 by Andrew Lippman and his team at MIT. The streets of Aspen, Colorado were filmed from a moving car and the footage was stored on laser-discs. This interactive movie map could be watched on a touchscreen display allowing users to choose their own routes at each junction. The relatively new MPEG-4 specification now offers authors a standardized language for organizing interactive data. The user can interact on the basis not only of shots or scenes as the smallest dynamic unit, but also of objects within the picture. Additionally all data can be accessed in a three-dimensional setup. But the crucial opposition between database and narration endures. New technological standards like MPEG-4 will probably bring new hybrid formats. Soap operas and TV shopping can be merged so that users can obtain additional information about characters and order products featured in a show. It is more interesting, however, to see how cinematic narration responds to the database challenge. Manovich refers to Vertov's Man with a Movie Camera as the first database film, and also to Peter Greenaway, whose "favorite system is numbers. The sequence of numbers acts as a narrative shell that 'convinces' the viewer that she is watching a narrative. In reality, the scenes that follow one another are not connected in any logical way." Consequently Greenaway has moved away from temporal cinema in recent years and now concentrates on spatial installations and new-media projects. Alongside Manovich's database filmmakers, we can find a lot of influences in contemporary films. The end of narration as dominant cultural form posited by Manovich seems to be demonstrated in very diverse forms such as the para-narrations of David Lynch, blockbusters like Quentin Tarantino's Pulp Fiction (1994) or Christopher Nolan's amnesia thriller Memento (1999). The fact that in Pulp Fiction the linear form of narration is already partially suspended has nothing to do with random order: even this post-linear narration is precisely constructed under full control of the author. Whether the viewer would prefer to change the story remains an open question.
An attempt to spatialize narration can be found in synchronized multi-stream dramas that offer the viewer two or more corresponding linear programs to jump between. On 1 January 2000, seven Danish TV stations co-broadcasted the D-Dag project set up by the "Dogme 95" initiators Thomas Vinterberg, Lars von Trier, Soren Kragh-Jacobsen and Kristian Levring. On the preceding eve of the new millennium, the four directors shot four connected real-time films that were then shown synchronously on four different channels. The fifth channel showed the first four films in split-screen format, while the sixth and seventh broadcasted the behind-the-scenes communication among directors and actors. The passive viewer became user and with the familiar remote control could edit his own film.
D-Dag is only one of several similar projects. In 1991 two German TV stations produced Mörderische Entscheidung (Murderous Decision) by Oliver Hirschbiegel, a cross-genre story somewhere between film noir and detective movie. The film was shot in two versions: one was from the perspective of a woman, the other followed a male figure. Both films began identically, then separated, sometimes met in double version of scenes with both characters, and at the end became identical again. What is interesting about Mörderische Entscheidung is that it demonstrates in almost didactical fashion all possible relations between the two narrations. Hirschbiegel uses the narrative voids we know from film noir as a general style to give the viewer the feeling that a lack of certain information is not caused by zapping incorrectly. To make sure that main story remains understandable important information was given on both channels at the same time. Hirschbiegel also tried to direct audience attention towards one channel - if not to say make people zap due to boredom - by reducing the amount of information given on the other channel. In an empirical study of Mörderische Entscheidung, Kay Kirchman revealed that the more similar the two versions were, the greater was the viewer satisfaction. In particular, most viewers misunderstood a more experimental party scene that was shown realistically from the woman's perspective in one version and in the other through the eyes of the narcotized male protagonist. The experiment worked best when both versions showed the same information from different points of view - be it a classical shot/reverse-shot-relationhip or a scene that was broadcasted and filmed from the monitor. The Aristotelian unity of space and time had to remain untouched, and along with it the narration. By contrast, the similarly structured DVD The Last Cowboy (Petra Epperlein and Michael Tucker, 1998) from the outset avoids straight narration in favor of three parallel video-essays about the myths of America. The film contrasts memory, imagination and reality among which the viewer can zap without losing the trail.
Valdis Oscarsdottir, editor of the theatrical release version of D-Dag, describes the Danish project as follows: "You zap around as you normally do. You start by checking out the first channel, and then you decide to check the next channel. If that isn't interesting, you zap to the third channel. If that doesn't seem interesting you just try the fourth channel. If the fourth channel turns out uninteresting as well, you can go back to the first channel. If you are out of luck and that turns out somewhat boring as well, you can just, zap, zap, zap through all the channels." This statement as well as Kirchmann's analysis of Mörderische Entscheidung show that the synchronized multi-stream drama format does not function over active decisions in favor of one channel or the other but rather over decisions against a specific channel at a specific time - much like the way people normally zap through the channels. The effect might be called a "negative aesthetic of boredom."
The question therefore remains: What can be done to overcome linear narration and deconstruct the author's authority without forcing the user to assume the responsibility and not always pleasant duty of co-authorship? Mike Figgis brought parallel narration in split-screen format to the cinemas with Timecode (1999). Viewers were not required to choose among the four pictures, but instead were guided by the director's sound design. In an interview, Figgis took what can only be regarded as a step backward: "We also hope the DVD release people will get the four unedited soundtracks, and the music as a separate element, so they will be able to do their own mix. If we get enough space on the DVD, they will be able to (view) each film separately, to build the narrative their own way." Somewhat more interesting is the fact that Figgis held screenings in which he performed live sound mixes, and in doing so reactivated the projectionist - a figure that nearly fell victim to automation and digital satellite distribution. Before the rise of VJ culture, projectionists tended to be regarded as a source of mistakes and irritation, but might now play a key role in turning linear cinema narration into post-linear performance. 
The Taiwanese director Ko Yi-cheng made his film Lan yue/Blue Moon (1997) specially for remixing: "I divided the story into five episodes and devoted each twenty-minute episode to one set of things happening to this trio. In two thousand feet, I had to finish every episode. The reels can be screened in any sequence, so you'll have five different endings: either the trio walk down the road holding hands, or they separate, or the woman rejects both men, or she falls for one of the two men. 120 versions, 120 possibilities."  This approach recalls the early days of cinema when theater owners bought single scenes from the producers and combined these scenes to create a program. In line with this standard practice, Edwin S. Porter left it up to the projectionists to decide whether his famous take with the cowboy shooting at the camera be shown at the beginning or the end of The Great Train Robbery (1903).
Two points about the Taiwanese director's approach are interesting: First, Ko Yi-cheng pursues the humanist concept of a flexible work for viewers who nevertheless remain passive. Second, he transfers interactivity to the level of representation. While the traditional notion of interactivity aims at altering the fabula itself, Ko Yi-cheng works with already established syuzhet schemata like flashback or ellipsis. Obviously, the story ultimately changes according to the sequence of the single episodes, but this change comes about only through the viewers' individual interpretation of what they see. As a pre-cinematic occurrence, the fabula remains undefined.