Expressive Systems

This is an excellent and very important book. In the mid-1990s we were really starting to understand what made computer systems usable, how human interactions with their computer systems could be tuned to make a more rewarding experience. The user would reach the goal of their computer use and be helped to solve any problems on the way, without any of the common frustrations and mistakes.

Then the Internet happened, and suddenly we were put back 10 years. Everyone wanted "web based" systems, which had a some advantages, but one big disadvantage, Although such systems were superficially modern, with pretty colours and pictures, their models for the human-computer interface were mainly out of the dark ages of mainframe computing.

Richard’s book puts us back on track. He reminds us that we should be building software for people to use, and that if we get that wrong, most other considerations are secondary. He also has a clear vision of how systems should work, exposing the underlying object model to the users, so that they can directly manipulate a representation of their problem or task.

He starts by exploring the shortcomings of existing systems, and asking how a better interface would improve matters. He then develops a clear understanding of the characteristics of his "expressive" systems, as well as explaining why these are very different from capabilities like end user programming. Unusually, this is explained as a "business case" which business managers as well as IT specialists can understand. By doing so, we understand why this should be important to senior stakeholders, and not just something to keep the users and "techies" happy.

The next section of the book explains basic concepts of object-based and component-based systems, and explores why these are the basis for systems which put the user in control, in a problem-solving role.

Finally, the book discusses a range of design principles for expressive systems. Some of these are quite radical, such as "Don’t start by asking your customers what they want". A recurring theme is that IT may have to move beyond the limits of traditional requirements, which limit systems to very narrow definitions of both the problems and the system’s relationships to the business and the users.

This is a very readable book, designed to be read through in just a few sessions. Even this may leave you inspired, but the book is rich in both ideas and examples, and hopefully you’ll go back again to get more and more value from it.

Don’t expect a tutorial on the details of interface design. For this, read Alan Cooper’s "About Face" or one of Bruce Tognazzini’s books. Instead, use this book as a source of both inspiration and justification when you want to do more than traditional analysis and current fashions allow.

The book is beautifully presented, with most topics handled in a single two-page spread which makes it very easy to dip into. It’s richly illustrated, although some of the material originally from CSC’s "Foundation" research programme is not so strong. I’m proud to be a contributor in my own right, my ConQuest system making an appearance on page 63.

If you can get hold of it, the original paperback version is spectacular value for money. I haven’t seen the new hardback edition yet, but I expect it to be equally good.

I heartily recommend this book.