A Pattern Approach to Interaction Design

Rating:
Summary: Great concepts and ideas, despite the examples
Review: This book addresses a basic issue of obtaining repeatable success in user interface design with the use of interaction design patterns. I seldom come across a book that poignantly strikes a chord where I have personally witnessed so much opportunity for improvement in our field. My personal work on hundreds of user interface applications over the past decade have led me to pursue the approach of Visual Design Patterns as a means to achieve more successful and consistent user interfaces. Jan Borcher adds rigor to this approach by providing an opening round of design techniques and pattern languages for documenting and communicating Interaction Design Patterns using an XML-based notation. His pattern approach draws upon object oriented pattern work by Kent Beck whom I personally was introduced to back in the late 1980's when Kent assisted me designing large-scale applications using Smalltalk for the financial community. The author does an excellent job of explaining how patterns can be used to capture, share and structure user interface design knowledge from their projects and how to use the patterns to allow for better communication among multi-disciplinary teams. My only issue with the book is the detailed working example used to drive home the concepts. For this, Jan uses an Interactive Music Exhibit to explain the detailed implementation of his ideas. While this approach can be very effective for some designers, my fear is that a large part of the book's audience will be left scratching their heads as they try to relate the interactive music case study examples to more business oriented examples facing them in their daily life.

Rating:
Summary: A disappointing book, but with some good ideas
Review: This is a deeply disappointing book, typical of a field in which approbation by academic peers seems to be much more important than delivering real help to those designing real user interfaces.

I had hoped for a book dominated by a collection of patterns describing how users interact with technology, particularly computers. Instead the first 3/8 of the book is given over to an almost blow by blow account of each contribution to the development of patterns, their application to IT and finally to HCI. There is actually one important message, that patterns can aid communication not only between IT professionals, but also between professionals and users, as a way of describing both the problem domain and the proposed solution. However, this was almost buried in a detailed discussion of pattern format, which is much less important.

I very nearly gave up reading at this point, which would have been a pity, since the central chapter of the book contains some interesting patterns, albeit of limited practical use. The "How to Play the Blues" pattern language is an entertaining demonstration of how to use patterns to describe a problem domain, while the patterns for interactive exhibits are a good set of HCI patterns, although focusing on the interaction of regular users with business computer software would be of more general practical use. I also really liked the pattern layout, relying on typography and styles rather than headings to standardise the structure, which definitely enhanced readability.

Unfortunately the book does not sustain the interest, and after a single chapter on patterns returns to a strange and lengthy self-review, to the extent of reprinting the review comments the author received on a draft version. This is ridiculous.

The book's malaise seems to be symptomatic of the whole discipline. The referenced web sites and books are not much better, with all but a few noble exceptions concentrating on academic discussions about patterns, rather than building a real pattern collection.

I am convinced that patterns can be really useful in the area of human-computer interaction and user interface design, just as they have been in many other areas of IT. I also think the arguments about style and format will be won by the author of the first good book that concentrates on building a comprehensive and readable set of useful real-world HCI patterns. This isn't it.