Top critical review
24 people found this helpful
A missed opportunity - laden with untested opinion
on 7 November 2011
This book is a tragically missed opportunity to sell Specification by Example into organisations.
Whilst brimming with "good ideas", Specification by Example ironically seems to lack the very thing that would make it more useful i.e. examples. Instead, the author, Gojko Adzic, relies on a series of quotes and opinions from interviews with research subjects in organisations that he identifies have adopted these practices. This is fine if you already "get" what the practices are about, but it doesn't really teach you how to go about doing it. Consequently, it's not a particularly engaging read.
One of the things that niggled at me throughout the book was the assertion that the primary goal is to automate your acceptance tests. He then goes on to suggest (from his research anecdotes) that this is actually very hard to do in a maintainable way. From personal experience, I would agree with this. So why set your readers up to feel they've failed?
My own experience has lead me believe that you don't have to automate your acceptance tests to get value from Specification by Example. Simply having requirements that stand up to the scrutiny of testing adds significant value and is something I have explored successfully with a number of teams. This has lead these teams to automate certain tests where appropriate and leave others to the domain of manual or exploratory testing.
The supporting practices explored in this book (including TDD and automated testing) are a hard sell in most organisations as there is a perception that they slow down delivery. Whilst this is an obvious over-simplification, the counter argument needs to be presented in more management-friendly language than is used here. A series of simple case studies would be more compelling than the quotes and vox-pops favoured by the author.