digital arts & culture 1998

Gonzalo Frasca:

Don't Play It Again, Sam:
One-session and serial games of narration

Extended abstract.

About the author


Adventure videogames, multi-user dungeons (MUDs) and hypertext fiction are usually described with terms like "interactive narrative", "non-linear fiction" or "interactive fiction". In this paper, we will refer to those works under the term "games of narration" (GON). This term, proposed by Espen Aarseth, introduces the game component, which is essential to fully understand those particular kinds of computer software. In addition to this, the term does not assume that GON are necessary part of narratives (actually, we will refer to GON as just being narrative-like instead of narrative). For these reasons, we think that the term GON is clearer than the ones that are usually used. We will also use Aarseth's definition of "ergodic" to refer to GON, (applied when "non trivial effort is required by the reader to traverse the text").

In this paper we propose to change a main characteristic of computer software: the ability to use a program more than once. We will focus on GON, particularly adventure videogames. However, the described techniques can also be applied to other kind of software, like hypermedia-based GON.


One characteristic of games is that players can always restart from scratch. Videogames even let the player to do something that is not available in usual games: to restart not from the beginning, but from a past moment that was "saved". This is also true for traditional narrative media. Books and movies also allow the reader to freely restart his/her "reading" or "watching".

We propose to alter this scheme, creating GON that can be used just once (One-session games of narration, OSGON).

William Gibson's Agrippa: a book of the dead, is a clear precedent of one-session computer-based literature. The program, distributed on floppy disk, allowed the user to read a poem just once. However, the poem was not ergodic.


Many techniques can be used in order to create one-session games of narration (OSGON). One can be a floppy disk that is erased or encrypted after use, like in Gibson's Agrippa. This is problematic because floppy media can not hold much information (if re-writable CD-ROMs became popular, this would be solved).

Another possibility is to keep the program out of the reach of the user, by installing it on an Internet server. The user can only access the information by getting (or buying) a password. The server will remember the "used" passwords, in order to prevent the access to the OSGON after it has been used.

This option would be perfect for OSGON like hypertext fiction. However, because of bandwidth limitations, it may be problematic for bigger programs, like hypermedia fiction and adventure videogames. This may be solved using a web-hybrid CD-ROM that would not execute unless it gets authorization from a web server.

A similar technique is being used by DIVx, a particular kind of DVD (Digital Versatile Disc) that allows the user to watch videos just once. Every time the user wants to watch the movie again, his/her credit card is automatically charged using a modem. However, DIVx focus on non-ergodic content.

Agency and life in GON

Janet Murray, in Hamlet in the Holodeck, defines agency as one of the main sources of pleasure in digital environments. "Agency is the satisfying power to take meaningful action and see the results of our decisions and choices".

Videogames are an environment that usually gives a high level of agency to the user. Adventure videogames usually let the users to perform many real life actions: walk, eat, sleep, manipulate objects, and die.

Computer agency has two major conflicts with the concept of agency present in real life and narrative. The first main conflict is about life and death. Videogame death is very different from traditional death. Generally, videogame's characters have many "lives". That means that if, during the game, the character dies, it will resurrect later. Players know that they can always restart the game from scratch or resume a previous saved game.

The "resurrection" ability of videogame characters is now an accepted convention among players. Nevertheless, "videogame death" is absolutely different than "narrative death". Generally, when a narrative character dies, there is no way back.

The second major conflict has to do with agency. GON can be restarted and each game session can be completely different from the previous, allowing the player to experiment many possible actions and behaviors. This can produce a problem for the player's feeling of agency: he/she knows that his/her actions are not definitive, because they can be undone later. So, actions become trivial, and the feeling of agency can lower. The player can easily become lazy and take decisions without considering consequences, because it is always possible to start over.

As OSGON can not be restarted, all the situations previously described can be avoided. An OSGON character's behavior would be much more similar to reality: he/she will have just one life and his/her actions would be definitive. These characteristics may increase the user's illusion of "being" the character.

Forbidden games, forbidden narratives

The thematic present in today's videogames is usually related to fantastic themes (science fiction, epic fantasy), mystery or detective stories. Unlike what happens in hypertext fiction, there is a lack of "serious" real-life topics, like philosophical issues, social problems, relationship conflicts, etc.

The reasons to this may be varied. However, there is also a cultural problem. Our culture reminds us that we are not allowed to "play with fire", and that some games "are forbidden". For example, we will probably offend many people if we design an adventure game, where the main character tries to survive during the Holocaust. Themes like the Holocaust remind us about the real value of human life, and in videogames, life is valueless: you usually have three or five and it is not big deal to die.

This same characteristic is the one that haunts the protagonist of Fearless (a film directed by Peter Weir). The man almost died in a plane crash. In an attack of anger, he throws his son's videogame console into the trash bin, telling him that in real life you just have one life.

OSGON can give an alternative way to face this problem, because it kills the start-over characteristic of games. As you can not restart, it is also trivial to talk about winners and losers. The player will focus on action, getting closer to the character, and feeling more the consequences of the actions that have been taken. If the character gets hurt, he/she will stay hurt during all the session. Anything that happens is irreversible. If the character dies, the action will literally end, creating a more realistic representation of death.


Many narrative serials are available in the World Wide Web. Some of them are non-ergodic, but most of them (The Spot, Homicide) usually have some ergodic techniques. The interaction usually is though hyperlinks and bulletin boards where readers can post information.

Janet Murray is one of the few authors interested in the development of ergodic serials. She imagines a new kind of ergodic serial, the hyperserial, mostly as a database-enhanced traditional serial. In such a serial, the user would also have access (through a web site, for example) to a big amount of information about the characters and the story. For example, users may browse the character's personal objects, and freely explore the places where they live or work. Users would also be able to explore information about previously emitted chapters.

However, none of the existent or proposed ergodic serials include one important technique that is usual in other kinds of GON: multiple-path plots. An ergodic serial with this feature, based on the normal adventure videogames paradigm would be impossible. Serials usually have a coherent plot that continues from chapter to chapter. Using the "restart as many times as you want" structure, the user would be able to have different performances on the same chapter, making impossible to have plot coherence in the next chapter.

This problem can be solved serializing many OSGON. Each chapter would have access to each user past performances, giving coherence to the plot. This will help authors to create web or CD-ROM based serials, that are actually much similar in structure to other media serials, like television, but that also give a bigger set of ergodic resources, like the ones available on adventure videogames.


Goals - Games are defined by philosopher André Lalande as a "kind of play that results in a triumph or defeat". One consequence of this is that games have goals. Main goals can be, for example, to rescue a princess or to solve a crime. Secondary goals are usually things that have to be done in order to accomplish main goals. For example, finding a key to open a locked door.

As OSGON can not be restarted, it will be unfair to design a game with a main goal, because the user will feel frustrated if he/she can not win in the first try. A much better possibility seems to be giving the user many secondary goals, or let him/her to define its owns goals. Those options are usually available in videogames known as "simulators". Videogames like Sim City let the users to freely experiment with the actions and the environment, setting his/her own goals. You never "win" in this kind of games, but you can lose. The game may also impose some secondary goals, like keeping the major popularity high, or administrating with care the city funds.

Time - As OSGON can be used just once, they necessarily have to have a limited duration. The author may design the program so its closure would be a direct consequence of the user's actions. The story would end if the character dies or wins.

Another option would be to have a limited time, like in theater plays or movies. The user would be able to experience the program for, let's say, a couple of hours.

The critic's role - Critics may have a hard time dealing with OSGON and serial OSGON. As in videogames, the narrative characteristics depend exclusively on the user performance. However, in this particular case, there is no way to test how the program deals with different kinds of input, because there is just one possibility per person to experience it. Critics will not be able to experience the whole work, so probably their reviewing task would be different, and probably more difficult.

User perception of depth - As the user can not experience all the possible actions and content of the program, he/she will have a very particular perception of its characteristics. The only possibility to have an idea of alternative experiences with the software is by exchanging opinions with other users.

Unlike what happens in videogames, its is almost impossible that the user can have a real perception of how big or good the whole program is.

User frustration - As the user has just one chance to experience the OSGON narrative, it is very likely that he/she would not be able to see all the elements (like characters, environments, games, texts, etc.) included by the programmers. It will also be impossible for the user to know what would have happened if he/she did not behave in a certain way.

As exploring new places and alternative plot paths may be a source of pleasure for the user, the designer should try to focus on showing that it may be also interesting to explore a virtual world where actions behave like in the real world.

Consumer perception - It is still early to know what would be the final consumer reception of DIVx, the only massive one-session medium available right know. Many reviewers have considered that DIVx only advantage was not for the user, but for the movie industry, which now can get money every time someone wants to play a DIVx.

We believe that OSGON give real advantages over traditional GON. However, consumers, based on previous experience like DIVx, may judge it just as a malicious system designed just to make money with non re-useable software.

Economics and Production - The OSGON may take the same amount of production time as a traditional adventure. However, its duration would be much shorter. As OSGON are probably going to perceived by consumers as something "smaller" than adventure videogames, they probably must be charged a lower price.

So, OSGON may not be a very good economic deal, at least for a while. The production work would simply be too hard with nowadays techniques for the possible revenues. However, this may be just true for adventure-like OSGON. OSGON based on hypertext or hypermedia has lower production fees, and may be a good medium to start with.
Right now, videogame production system is very similar to movie production. OSGON production would require a more television-like production scheme. This means to have pre-built and re-usable components (code, graphics and sounds) instead of starting from scratch on every project. Such GON would not look as good as the latest videogames available, but they will be much easier and cheaper to produce.


Our goal is not to say that OSGON are better than traditional ergodic literature, like adventure videogames or hypertext fiction. Neither do we believe that OSGON will replace those genres. Our main objective is to describe OSGON and OSGON serial's main characteristics, and why we believe they have an interesting ergodic narrative potential.

It is yet impossible to predict whether OSGON would become popular or not. However, we believe that it can be a good example of how new paths in ergodic literature can be opened, by simply experimenting with elements in GON's structures.

Gonzalo Frasca
Jacaranda Multimedia (Uruguay)

(1) 541 -754 8031






about Bergen

Please ask Jill Walker (local coordinator) or Espen Aarseth (conference chair) if you have any questions.
Last update: 22
nd of October 1998