Complete Systems Analysis

Unlike many books on systems analysis, the Robertsons' does not confuse analysis with "design". This book focuses on developing an unambiguous system specification which both the sponsoring end users and the developers can understand. The examples are clear and well chosen, small enough to be understood, big enough not to be trivial. This is an excellent main textbook for a course in systems analysis.The documentation methodology is derived from the classic DeMarco "structured analysis". The book's main weakness is insufficient treatment of object-oriented concepts, especially in data definition.

For persons needing a well written platform independent non-dogmatic textbook and workbook to bring them up to date on structured systems analysis, this book is excellent. Includes a brief explaination of how to transition to the design process (including Object Oriented Design (OOD) as well as old fashioned 3rd and 4th generation languages), but refers you to other books for design (and project estimation) details. Alas, has no recommendations for best CASE tools (even though they would be circa 1994). The book will take you at least 20-40 hours to go through the "Easiest" trail even without doing the work. I suggest that you read it 3 times: 1) Easiest (but don't do the work, 2) : Promenade, 3) Easiest again but doing the work and simultaneously creating mini-specifications and data dictionary of the analysis strategy (see chapter 1.18 before you start). Quite frankly, while the book implies you can do this with index cards and paper (you can for the case study), in real life you will need a good CASE tool (which is why no recommendations are disappointing). I have tried to cobble together my own "CASE toolkit" using commonly available workprocessing, project management, and flowcharting packages, and it is extremely cumbersome and impractical to use. Also useful would be more practical interview tips (which are short-changed in exchange for more empasis on how to properly analysize DFD and ERDs) such as 1) avoid tape recording users (it makes them nervous and is impractical), 2) listen to the user discribe his system, then when he has finished, then draw the DFD and ERD while repeating back what he said (writing/drawing while users are talking distracts you and the user), etc, etc, etc. Thus this book is a good book to bring an "old hand" up to date as well as a good first semester course book for Computer Science majors, which would need to be followed by a second semestor "live" internship. Together with its bibliography, this book will get you started in system analysis (again). I suggest you read this book first before attempting to read OOA (Object Oriented Analysis) books and implimenting OOA stratagy, especially if you are an "old hand" and are unfamiliar with C++ or JAVA