Roger C. Schank
Northwestern University Press, 1990
Roger C. Schank
Northwestern University Press, 1990
ISBN: 0810113139 paper.
Buy it from Amazon.com
Buy it from Amazon.co.uk
In this pioneering, enjoyable and insightful book, Schank leads the reader to start from scratch when thinking about what we mean by knowledge, memory, and intelligence.
His thesis is essentially that being human means telling stories:
"...all we know is embodied in stories"
We remember stories, and perhaps they are the key to how we organize our memories, which is to say our minds. We know what to do in a situation because we have a wired-in script that we have (maybe painfully) learnt -- and we behave with skill in that situation because we can follow that story efficiently. We can intelligently plan and reason about alternative plans, because we can tell stories with happy or unhappy endings for each variation.
I hope this is enough context to make it quite evident that Schank is totally relevant to thinking about how to capture and organize requirements. His work on Scripts set the tone of much of the Artificial Intelligence research in the 1980's: what was then called "knowledge elicitation" has of course transmuted into "requirements elicitation" today. Schank consistently argued that people ran their lives by stories, such as the famous restaurant script -- you go in, hang up your coat, get shown to a table, sit down, read the menu and so on. This was taken terribly literally as a precise set of rules at the time, which (of course) didn't work; but Schank's work has also always emphasized the vague and associative nature of human thought, such as the way we connect one story with another, or events and details with a story. This should have warned people not to be too rule-bound and literal.
Schank writes fluently, cogently, and with a sense of fun as well as of exploration -- he dares to break the rules (and across the cover of the book is a strapline that says "rethinking theory"). The rethinking has to reach back far into the past, indeed to Aristotle, whose view of reasoning was that it had to be theoretical and hence rule-based, episteme -- and for the last two millennia, the theory of knowledge has been Aristotelian "epistemology". A competing view even in ancient times was that you could reason from instances or cases, casuistry, which like rhetoric and sophistry was not originally a pejorative term. Only since the 1980's has case-based reasoning come to be honoured again, and rulebases have come to be seen as hopelessly limited except in very tightly-delimited domains. It's almost funny that it has taken 2000 years to get back and fix mistakes made by the Ancient Greeks, but that's the way it is. For me, this is extremely important -- it is a revolution in thought, and it goes a long way to explaining why traditional arguments always break down. Another (closely-related) part of the repair and replacement of outworn arguments is Polanyi's demolition of the idea that knowledge is conscious -- much of the way that skill works is plainly tacit.
Given the contributions of Polanyi and Schank, requirements engineers -- indeed, system thinkers and developers in general -- need to come to grips with the fact that people, even if highly skilled and even 'expert', cannot give us precise rule-like requirements. If we try, we will get bad and wrong requirements. Instead, we urgently need to start thinking and working in stories -- scenarios, use cases, scripts, user stories, whatever. We can use these as hooks to retrieve connected ideas and build a shared understanding of what people need, by talking about what they actually do. Then we'll have requirements that are somewhat closer to reality.
The book looks at where stories come from and why we tell them; how we understand and index stories; how stories shape memory; "story skeletons", i.e. the structure of stories; the stories of different cultures; and the role of stories in intelligence.
There is a very short list of references, but Schank sprinkles the text freely with quotations and jokes, examples and stories from all over -- if you don't know what Chutzpah is, you soon will from Tell Me A Story.
Everyone interested in writing better requirements should at once read this book. Academics will find many stimulating ideas for things that ought to be researched; practitioners will be able to reflect in an easy-to-read book on their daily work; students will get an insight into one of the leading minds in the field of human knowledge.
© Ian Alexander 2002
You may also like:
