User Interfaces and Design

The Design of Everyday Things, by Donald A. Norman

Very thought-provoking. 10 out of 10.

Designing Large-Scale Web Sites : A Visual Design Methodology, Darrell Sano.

Sano's book has perhaps one and a half chapters (3 and 4) that live up to its title, and the rest is either filling or out of date. Having said that, the one-and-a-half chapters are essential reading. Sano emphasizes the need to provide a clear organizational framework -- an information space -- through which a visitor can navigate. He proposes hierarchical division of the space, with well-defined, coherent paths through and between components of that space. When you read this chapter, you suddenly realise what is lacking on so many of the sites you visit.

The failing of this book is that it loses focus on the key, eternal design issues, and starts rambling on about particular features of Netscape (Sano works for Netscape). In the world of the Web, this is guaranteed to doom any book to early obsolescence, and Sano has done his book a disservice by focussing on the features of Navigator 2.0. I give this book a recommended rating, but with reservations about spending your own money on it because of the amount of obsolete material.

John Reekie, January 31, 1998.

Understanding Computers and Cognition : A New Foundation for Design, by Terry Winograd, Fernando Flores

This book is a somewhat dated (published in 1986), but it shows it only in its efforts to challenge the directions of AI research about 15 years ago. Other than that, it offers much interesting food for thought on the nature of cognition and interaction. The writing is on the wordy side -- "Before we could become open to their relevance and importance we needed to take a preliminary step towards unconcealing the tradition in which we lived, thereby recognizing that is was in fact open to serious question."

Some of the insights I gained from this book (in my own interpretation):

• The importance and nature of "thrownness." Similar to Normans' experiental cognition, thrownness is the condition of being in and responding to the external world.
• We "represent" information in our responses, not in passive data stores. I think of this as being "higher-order": when we learn a skill, we modify our organism such that it changes its behaviour in response to certain needs. We do not merely collect information about the behaviour in order that we might reproduce it later -- the author's call this the "metaphor of information processing," and argue against it.
• The breakdown of thrownness: when this mode of operation fails, only then are we required to realize what we are thinking about, to "process information." in order to overcome the breakdown.
• Winograd and Flores offer a particularly concise and revealing summary of what makes computers different from other devices we construct: apparent autonomy, complexity of purpose, structural plasticity, and unpredictability (pp 93--95).

This little book certainly deserves a second reading.

JohnR, 12 January 1997. 7 out of 10.

Things That Make Us Smart : Defending Human Attributes in the Age of the Machine, by Donald A. Norman

After The Design of Everyday Things, I was very disappointed with this book. In the first half, Norman presents the two key kinds of cognition: experiential and reflective. Experiential cognition is reactive, automatic thought: the kind that an expert at something calls into play, and which is most apparent in activities that include motor action. Reflective thought, in contrast, comes into play when deeper decision-making is required. (I believe this is what Winograd and Flores call a breakdown of "thrownness" -- the state of being in and responding immediately to the world.) He presents some interesting experiments in which the presence of extra rules in otherwise isomorphic problems affect the speed at which they can be done.

After that, it's very uneven indeed. Instead of suggesting how this knowlege might be incorporated into the machines and user interfaces we build, Norman spends a lot of time complaining about how poorly specific examples of modern technology meets their users' needs, but very little suggesting solutions. In a particularly bad chapter, Norman predicts the future, and ends up confusing himself (as far as I can tell) about what the experiental mode of cognition means: he laments that technology is being used for pure entertainment -- too much experience and not enough reflection -- when earlier I understood that experiential mode is what makes us able to perform certain tasks effectively.

Norman somewhat petulantly demands that retrieval of information from electronic databases should be by description: "What matters most is that the users be permitted whatever descriptions are most relevant to themselves." Natural language, perhaps? But Norman has just been arguing that imprecise pattern-matching is what we are good at, and is the standard by which we should judge intelligent behaviour: why, then, not build affordances into the machine that do not require them to be as smart as we are? I don't ask my power drill to place a hole just over there, or point to a location on a map and expect my car to drive me there -- yet no-one ever complains that these machines are not responsive to our needs. They have the affordances that enable us to accomplish certain tasks effectively and -- with practice -- extremely efficiently by using experiental mode. By the same token, what is needed for electronic artifacts are constructed affordances that enable us to achieve our goals effectively (and precisely -- a failing of natural language). It seems clear from Norman's earlier groundwork that we should use experiental mode for instructing the device, and reflective mode for high-level navigation, but Norman never quite says it or develops the idea.

The only real insight in the second half of the book is the way in which physical artifacts (airline controls is an example) that are no longer needed technologically are nonetheless essential cognitive artifacts for group work.

In summary, if you haven't read Design of Everyday Things, read it, and forget about this book. Much as I hate to say it, this one is a real let-down.

John Reekie, January 12, 1998. 5 out of 10.

Interface Culture : How New Technology Transforms the Way We Create and Communicate, Steven Johnson.

I liked this book very much, because Johnson takes a look at user-interfaces and the Web from sufficient distance to be critical and perceptive. He provides interesting historical background to user interfaces and hyper-text from the perspective of what the developments have meant to "users." I read the book somewhat hurriedly on a long plane trip, and the book deserves a more careful reading. However, two perspectives in particular that I found illuminating:

1. The metaphor of computers, the Web, data in general, as a space. Sure, nothing new here, since the metaphor pervades computer culture from the "desktop" through to William Gibson's Neuromancer landscapes, but I found it enlightening to realize how ubiquitous the spatial metaphor is (in an abstract sense), and how important navigation of that space is to effective user interfaces. The more I think about, the more I realize how pervasive the metaphor is -- or should be -- simply because of the way we think.
2. The concept of user interface as a filter. Johnson's analogies were a little strained here (as I recall -- I haven't got the book here right now), but the idea that the interface is not really presenting you with information, but hiding irrelevant information was quite compelling.

I thought he got unstuck on the chapter on agents, and I couldn't follow why he considered agents a new form of user interface. Johnson takes a few pokes at Sven Birkerts and his Gutenberg Elegies, which suits me fine.

JohnR, Dec 12, 1997. 8 out of 10.