Drucker -- Chapters Three and Four

In the first of these chapters, Drucker extends the ideas she developed in discussing static information visualizations to the theory of web interfaces and user interfaces in general.

She begins by detailing the historical development of user interfaces and provides a critical overview of interface theory in general. Her main criticism -- similar to her earlier criticism of information visualizations, as they are -- is on the emergence of Human-Computer Interaction (HCI) from highly technical fields like engineering. She describes how traditional HCI methods tend to abstract away any sort of nuance, subtlety, or ambiguity about the behavior of its subjects, instead viewing them as effectively mechanistic "users" who can be sufficiently modeled by feedback loops and related systems. Drucker advocates instead for viewing these "users" as human "subjects," the change in terminology emphasizing her argument that users are ultimately human, that they cannot be fully understood, and that acknowledgment of this ambiguity of human-ness should be explicitly incorporated into user interfaces.

To further elaborate on these ideas, Drucker briefly discusses Media Theory, saying that its ideas can be effectively applied to Interface Theory. She explains how reading a website is "an act of self-production," saying that a subject-oriented interface (rather than the existing user-centered interface) should take into account that, in using an interface, the human subject is constantly trying to situate herself in the interface she perceives. In this way, the information that an interface successfully communicates is heavily dependent on the changing subject that interacts with it -- of course this subject cannot be statically modeled. The interface, Drucker says, "becomes a codependent in-betweenness in which speaker and spoken are created" (151).

She then goes on to describe the development of reading interfaces, explaining how web-based reading interfaces are unnecessarily similar to the structure of physical books. She explains how, in the development of books, the form was a result of careful consideration of how a book was used and argues that effective web-based reading interfaces should be designed in the same way -- that we should consider the nature of how we read on the web and capitalize on that, not simply rely on simplicity and familiarity.

In the following chapter, Drucker describes further how interfaces should be designed to support "knowledge production." As in previous chapters, by this she means that the design of an interface should acknowledge that information conveyed will be a result of interpretation on the part of the subject. To maximize communicative ability, then, Drucker advocates for interfaces that can be read and navigated combinatorially, which helps explicate the interpretation of the subject.