What do you think?
Rate this book
304 pages, Paperback
First published January 1, 1993
Norman is such a pleasure to read. His prose style is light and easy, he has brilliant examples, and he mints some very useful concepts. The Things That Make Us Smart picks up many of the themes of The Design of Everyday Things, but gives them a different emphasis. DOET focussed on physical artefacts like plugs and door handles, whereas TTMUS focusses on information technology. DOET focussed, rather humorously, on design failures. TTMUS is more abstract and general, as Norman tries to explain how 'human-centred design' can apply to incorporeal objects like software programs.
The crucial chapter in this book is Chapter 3, 'The Power of Representation'. In this chapter, Norman introduces the book's three key concepts: problem isomorphism, representation and cognitive artefacts. 'Problem isomorphism' is the ability of a single problem to appear in different forms. These different forms might be logically equivalent, but can have very different meanings for a human user. A 'representation' is a particular way of displaying a problem to a human or a machine. 'Cognitive artefacts' are machines that humans use to extend their cognition—for instance, a pen and paper to extend our memory, or a calculator to extend our arithmetic. Norman combines these three concepts to explain why electronic systems are often so poorly designed. The system chooses to represent the problem in a way that is convenient for the machine, but inconvenient for the human user. The system therefore fails to function as a cognitive artefact, making it harder for the user to think through their problem when it should be making things easier. But since problems can be represented in multiple logically equivalent ('isomorphic') ways, it should be possible to design systems that fulfill the inner requirements of the machine while also fulfilling the requirements of human cognition.
It is an elegant theory, and Norman supports it with many examples and arguments. He draws together many ideas from cognitive science, data visualisation and other fields to buttress his conception of human cognition. I must say that the book does tend to rehash material from DOET. The familiar concepts of 'natural mapping', 'slips vs. mistakes' and 'affordances' return. And there are not quite as many examples in TTMUS as in the earlier book.
There is no doubt, however, that this is a brilliant work on the subject. And it is edifying to read a book on machine intelligence that focusses on how human and machine can work together intelligently. There are no silly fantasies of 'superintelligence' or 'singularity' here!