We thought you and your visitors might be interested to hear about a new Alternate Reality Game that we’re putting together.

“Traces of Hope” is being launched as the first ever charity online ARG and is being built by the British Red Cross to coincide with its Civilians and Conflict month. The game features Joseph a sixteen-years-old caught up in the Ugandan civil war, separated from his family, hungry and alone in a camp overflowing with thousands forced to flee, Joseph is desperately seeking his mother. But he needs your help…

Registration will open on Sunday 28th but until then there’s a teaser page at http://www.tracesofhope.com and a teaser video at http://www.vimeo.com/1811645

We’d really appreciate a link to the game or any coverage you could provide. Any questions please let me know,

Thanks

Alex

Actually I agree with some of Matthew Murray’s criticisms, but of course not to the degree of severity he paints them.

The nuanced tone of the characters in verbal intonation, facial expression, and positioning in space all add to the emotional resonance of the work. There is a lot of subtle work in the characterization of this “materially successful” couple. The visual design of the system supports this as well– two-dimensional world and avoidance of primary colors in the color scheme serve the tone of the narrative well. Visually it is conspicuously different than other game-like systems, this becomes a marker of the authors’ intentions to create a serious interactive dramatic experience for adults in contrast to the numerous juvenile targeted narrative computer games that exist.

Façade is quite different than narrative adventure games for example, because in those games the moment-to-moment interaction is largely not dialogue-based but action based. The use of dialogue in Façade seems to be about influencing characters and uncovering hidden levels of meaning. In a typical narrative computer game the dialogue is used to move the plot forward and to make decisions that affect player character resources — with development of the player character (enhancement of statistics and acquisition of possessions) often given equal status as a goal to the plot driven goals.

As in the Mateas, Vanouse, Domike collaboration “Terminal Time,” there are fixed narrative attributes and variable narrative attributes between each execution of the system. In my own work I have found that a potential power of a computational narrative artwork is meaningful difference between repeated user experiences. For example, meaning comes out of the contrast between the output of different executions of the software. Terminal Time uses this characteristic as its central feature — it critiques the dominant and monolithic voice of traditional documentary film with clever rhetorical differences and political positioning between each system execution. Façade also takes advantage of meaningful difference between user sessions , but this occurs more in the user’s accumulation of experience with the story world and interaction mode. We learn how to avoid certain paths or dead-ends. We learn how to move Grace and Trip to discuss one subject matter or another, and potentially how to delve deeper into that issue next time. We learn of the paths that there are to explore and we can keep mental notes of them for the next iteration. Thus, the meaningful difference between runs here has more to do with uncovering more of the narrative space (learning the full story between the characters) and game-like accrual of user experience.

An additional effect of this characteristic, is that though the experience is more “narrative” than “game,” many users will still adopt a goal-based strategy when interacting with the system. For example, users may feel a sense of failure when failing to reconcile the arguing couple. My suspicion is that even while exploring the narrative space, users will have a sense that there is the “correct” set of paths that end in the couple’s reconciliation, and a set of failure paths with the user either being kicked out or the failure of Grace and Trip’s marriage in one or more ways. I believe that this sense comes out of (1) and expectation for game-like conventions yielding from popular computer gaming and interactive experience, and (2) the characteristics of the meaningful differences between user sessions.

The experience also surprised me in the degree to which its interactive fiction (IF) heritage manifested through my subjective experience of interacting with it. For me, the character of moment-to-moment interaction with an interactive fiction system, involves learning the types of interaction the system expects from me. Since it is never possible to anticipate everything a user might attempt, there is always some effort to deal with unexpected user input. Interacting with such systems has a qualitative feel to it, whether it is a generic response such as “I don’t understand what you typed,” a clever response in the tone of the game (that still indicates that the user should make another choice) such as in the LucasArts “Monkey Island” series of games, or an attempt at a more graceful response that tries to deal with the user’s actual intentions.

In Façade, I found myself trying to understand the sorts of input that the system would respond reasonably too. Hence, as in IF, there is this feeling of testing the boundaries of the input system. On several occasions the system did not understand what I meant as earnest input, e.g. (if I recall correctly) when I typed “Grace, I really like your painting” for some reason Trip responded with a statement like “You think I should start to paint?” I remember in my youth a similar process of having to reign in my input patterns to something that the system would recognize. All this is not to say that Façade is qualitatively equivalent to these older interactive fiction systems, rather that Façade is a quantum leap forward in an underdeveloped legacy of interactive narrative systems.

This interactive fiction character did not only have to do with the limitations of the Natural Language Processing (NLP) system. It also came from the narrative constraints. The tension between the ability to interact with the characters and environment and the dramatic arc of the story is more apparent than I would have guessed. The user’s agency exists only within the dramatic world of the system. I do not see this as a fault. I agree with Eric Zimmerman’s assessment of successful games as being systems of constraints, as noted in the book “Rules of Play” co-written with Katie Salen. Mateas and Stern’s work provides an example of a similar value for interactive narrative, the user should *not* be able to do whatever s/he wants within the narrative world. The user should be able to make meaningful choices that result in meaningful differences within the story world.

Which brings me to an issue debated between Chris Crawford and Michael Mateas on Grand Text Auto (or was the debate elsewhere and only recounted on the blog?). In paraphrased form, Crawford argues for iconic user input languages whereas Mateas argues for allowing users to be able to express himself or herself in natural language. Crawford is interested particularly in guaranteeing the salience of user decisions with regard to the story, while Mateas is interested in allowing users a means to express themselves in their own manner as well as avoiding exposure of the deep structure of the system’s narrative content.

My feeling is that both approaches probably have their place, but what is important is the qualitative experience of the user. Interpreting natural language alone will not allow for individualistic sense of expression for users. This is not to say that it should be uniformly avoided, however.

Since there are limitations to the state-of-the-art for natural language processing, and since it is a hard problem in general, the system ends up ignoring a great deal of the user input — focusing on noun and verb combinations. Because of this, individualistic and idiosyncratic input is incorporated at meta (social) level in the output scripts generated (as seen in some humorous examples of Façade generated scripts), but not at the level of user interaction and meaningful difference in system execution. Since the goal was to allow users a mechanism for personalized interaction this is probably not a desirable situation.

It is useful to note that an iconic input system could be built atop natural language system that takes iconic input and maps it to an input form recognizable by the NLP system. Likewise, atop an iconic system an NLP system could be built that maps parsed input to iconic input recognized by the system (perhaps even presenting graphical output of the iconic system as a form of acknowledgment of exactly how the system has parsed the user input). The point of this duality is that while both systems carry debatable features and drawbacks, the important debate is about how well the system supports the stated goal: meaningful interpretation of personalized input patterns.

For design of such systems I usually look at mappings between desired functionality and how these are represented in the system’s interfaces. Thinking about the problem this way, the idea of “input that expresses a user’s personal style” is too broad, it is important to ask “personalized how?” It is also important to ask what the core function of user input is.

For example, it may be important to allow users to input action/object combinations. This means that a verb/noun combination or possibly verb/noun/direct-object combination is the core user input type. It may be important to allow the users a range of emotional modifiers for these action/object combinations. Then perhaps the system should also allow for a set of emotion oriented adverbs. The main idea here is that a tight system could be produced from a minimal set of input options that constrains system to create a narrative world where user-action and intentions are always meaningful. The user can still have a sense of personal approach to interaction, but only in a way that is incorporated into interaction with the system. I offer this “emotion adverb” example only as a loose sketch to illustrate a general idea. A tight loop between user input and narrative world results probably will create a sense of user style meaningful within the story world (as opposed to a sense of user input style visible at a level outside of the system). Such a tight loop could be implemented successfully via iconic languages, a restricted natural language, or even a menu-based system. The important issue is for the author to carefully describe a range of meaningful user actions and a means to enable them unambiguously in the story world.

In summary, I believe that Michael Mateas and Andrew Stern have offered a valuable contribution to the world of procedural story-telling. Their production model (independent, experimental, academy-industry collaboration), their vision of artificial intelligence, their vision of interactive narrative, and even their distribution model (free!), represent vital work within the community of people interested in gaming, interactive narrative, computational fiction, and all related areas. The fact that Façade is a fresh and engaging experience is a bonus.

Traffic update: downloads have reached 40,000 today.

For those with American cable TV access, the videogame channel G4 has a show called Cinematech that apparently will be doing a piece about Façade called “Who’s Afraid of Interactive Drama“, at 10pm EST on July 27 and repeating the show a few times until August 2. (I don’t have cable, but I’ll try to find someone who does so I can watch.)

[Ed's note: for those who do not know, Richard Evans was the AI developer behind Lionhead's Black and White, and is rumored to be currently working on next-generation Maxis projects.]

It is hugely ambitious and very entertaining.

The biggest “Ooo, wow! I am witnessing the future of interactive entertainment and I want to kneel before it!” moment was when Trip said told me I had really been pressing his buttons tonight. He redescribed my behavior back to me quite accurately (praising him, praising his flat, flirting with his wife, asking if he had been having an affair), and asked if this was me trying to get him to realize something about himself. He then revealed more about himself, which was something of a shock, but also seemingly inevitable. This really felt like character self-revelation caused by an understanding of player behavior.

P.S. Trip’s pants are too high. You might want to think about pulling them down a little in a future version. :P

Considering the installer is 800MB, that’s over 20TB of data moved over the Internet, all virtually for free for us as developer/publishers.

They showed a brief demo of the host’s gameplay and gave a favorable
review, alluding to how great it would be if you could have first
person dialogue interaction in a first person shooter game. In
total, it was maybe 5 – 8 minutes of coverage I think?

We’ll add more links and info to the Wikipedia as needed.

http://en.wikipedia.org/wiki/Fa%C3%A7ade_%28Interactive_Story%29

A basic page! Could use some of your edits -

Nice work on this project! Thanks for freely sharing your results.

I’m going to pick up Brad O’Donnell’s entry in our second take over at WRT.