Notes

Chapter 1: The Nature of the Beast

Laurel opens with Spacewar, the first computer game, which was developed at MIT in 1962. The game was the “natural” thing to do with the computer when it arose. Laurel claims that this is because games have the capacity to represent action, and the key ingredient is the human element. That fact makes the interface a matter of significant importance. Agency is shared in computers, both the user and computer act and have a common context for action. This idea is required for the standard dialogue and communication model of UIs. The idea of agency and action is derived and anchored in Aristotle.

Laurel looks at psychology, which has been a longstanding influence in HCI (especially through Don Norman), and finds that psychology and theatre are closely related. Psychology “describes what goes on”, and theatre “represents what might go on”. The approach to interfaces defined by Norman and in psychology in general are found to be the same as the definition of theatre: “representing whole actions with multiple agents.” (p. 7)

Laurel finds some criticism with traditional definitions of interfaces. Interfaces may be a simple layer between the user and the machine, but cognitive science introduces the notion of mental models, so the user and computer have models of each other. However, this can easily turn into a “horrible recursion” with self reference and abstraction. Another classical approach is the view of the interface as a prosthetic, mapping directly between the user and the machine, where actions are mapped nearly one to one. To Laurel, none of these approaches is satisfactory, so she turns to look at a theatrical model of interfaces. The idea with this is that the program performs for the user, and that the UI should be modeled after the technical support for a theatrical production, which, when it works, is totally invisible and does not matter or affect the user’s experience. This idea seems to reverberate in contemporary “rich client” design, which uses display effects and lots of theatrics and feedback to demonstrate interaction. However, like in many theatre productions, technical support does not always conceal the nature of the play (for instance the moving mechanical set in Sweeney Todd) Not all models of theatre apply to transparency, some are highly reflective, and this should be seen as making quite a bit of sense, especially when connected to rich UIs.

The challenge with the theatrical model of the interface is the user’s role within the production. Clearly, a play can not easily welcome audience participation. Instead, Laurel proposes that the interface treat the user as a performer (or character), and this again echoes with others, notably Goffman. This tradition is generally known as scripting the interactor, and is commonly used in tutorials and games. Defining interactivity leads to another conceptual problem. This challenge also percolates into Chris Crawford on interactive storytelling. Laurel’s definition is this: “It enables you to act within a representation that is important.” (p. 21) This definition complicates standard assumptions with interfaces, but is nonetheless quite evocative. It however depends on ideas of action and representation that need to be fleshed out further.

Imagination plays a significant role in models: It is used by humans for planning (in traditional AI sense), but it is also generally used for much more. Art is an external representation of imagination. The things represented in art often tend to some with whole worlds of meaning, even when the things are wholly imaginary. The idea of an interface metaphor convolutes the relation between representation and reality, whereas the language of theatre better establishes that relation as a knowable boundary.

Chapter 2: Dramatic Foundations

The focus of drama here is of the Aristotelian variety, focusing squarely on the Poetics. The question of “Why Aristotle?” is addressed, and this seems to be because Aristotle is complete, and one might also say that Aristotle is easy to model. The Greek cultural frame of drama, specifically with the idea of divine possession inspires a more modern sentiment of immersion, which links to immediacy and transparency where interfaces are concerned. Divine inspiration precludes the boundary of a medium.

Aristotle outlines four causes of representation, and Laurel expands on these and applies them to HCI. Programs themselves are difficult to reduce to this model because of their complex layers and patterns. Laurel’s investigation of programs is persistently grounded in the user experience of the software. So, while a program may be very good at a task, if it is nearly impossible for users to perform it, the task is not effectively performed by the software. Laurel finds two things specifically: “functionality consists of the actions that are performed by people and computers working in concert, and programs are the means for creating the potential for those actions.” And: “The most important way in which applications, like plays, are individuated from one another is by the particular actions they represent.” (p. 45)

• Formal cause: This is the idea of applying to the essence of something, the abstract idea that the representation aims to represent. The formal cause of a play is the action and plot that is performed. The formal cause for a human/computer action is a representation of action (the functionality?)
• Material cause: This is the physical construction of an object, its substance and composition. The material cause of a play is its enactment. The material cause of an interface (not a program!) is its enactment as well, the presentation and feedback given to the user.
• Efficient cause: This is how the object is made, involving the maker and the tools. The efficient cause of plays is its technical construction, and with interfaces it is the same. However, interfaces also are very notable in their use of models and concepts, especially in reference to the underlying application itself.
• End cause: This is the purpose of the thing in question in its existence and application. The end cause for a play would be its catharsis, for the Aristotelian model. The end cause for an interface is the successful use of the functionality by the user.

Laurel also presents Aristotle’s six elements of qualitative structure in drama. These are: Action, Character, Thought, Language, Melody/Pattern, Spectacle/Enactment. Each of these are connected between UI aspects as well as classical dramatic ones. This is notably the same scale used by Mateas and Stern in describing Facade. An effect of these is that material causes propagate backwards, originating in the enactment and affecting the other levels towards the top. Formal causes originate at the level of action, and then move downward affecting each other aspect of the production. The conflict framed by this confluence of factors resembles the problems found in design of all sorts. Even still, this stratification of elements is found in many other sources (pertaining to architecture and design among other things), not all of which are Aristotelian.

In terms of the behavior of programs, Laurel emphasizes that it is important to connect the idea of character. For Aristotle, a good (virtuous, even) character is one who successfully transformed thought into action. A virtuous object is one which fulfills its purpose. This idea also implies that a good character is one who fulfills expectations. This in mind, that aspires to the idea of mythologically emphatic plots, closely perscribed experience, as well as to transparent user interfaces. What complicates the situation is that characters and interfaces must be appropriate for the function or action.