Rating: 
Summary: Shotgun approach to requirements gathering.
Review: Well, I bought this book and I have to say that, even though Ed Yourdon is a co-writer, I am pretty disappointed with it. The problem with it is that it has just too much information, presumably aimed at creating requirements specialists. Technique after technique is discussed, each in great detail, but never in such a way that I feel convinced that any one technique would actually be useful to the development process. In fact, while reading the book for the first time I became convinced that the book would be perfect for requirements consultant types, i.e. people who end up doing nothing but requirements gathering and giving advice in that regard. Now, maybe it is just me, but I have found this type of partitioning of roles really counter productive. In most projects today we have the situation where one group does the requirements, another group does the analysis, another the design, and ultimately some poor individual sits down and writes the code, tossing out most of the prior work from the other teams just because it is so completely wrong. I have come to believe that if the 'requirements phase' (in my opinion a phoney development layer, like the so called 'analysis phase'), is going to be any use at all, then it must be performed by the same guys who end up writing the code. The same goes for all the other big project development layers. The same team of people should be involved. If you don't believe me, and you're not in IT, just asks any developer what really counts in a project. Sure requirements are essential, and sure design is also essential. But what really counts (aside from the actual skills required) is the continuity of thought and knowledge throughout a projects life cycle. Having specialists who concentrate on one subset of skills breaks that continuity. I like Ed Yourdan's stuff so I do feel a little apprehensive giving the book such a low rating. But in my opinion this work needs to be firstly, mercilessly edited down to around half of its size. So, if there is to be another edition, kill your babies' guys.Secondly it needs to be rewritten to give some real direction to developers, not contain a sequence of 'you may want to do's'. Consultants from the PWC's and Andersons of this world, and Dilbertesque managers will love this book, but most developers will most likely find it a waste of time.
| 
