Software Requirements and Specifications

I'm a technical neophyte but I have to communicate ideas concerning software development to developers. Jackson's book takes a complex subject and puts it in simple terms. He creates common ground for the technical and non-technical to communicate. A must for implementation, business analysts, and project managers if you to translate the customers needs to the developers process.

A great deal of software uses the "where the arrow lands, draw the target" approach. Jackson presents formal terms and techniques to pursue an alternative. Anyone collecting requirements and writing specifications will find benefit from this book. His discussion of various problem frames is wonderful.The book is written in tool box style, and Jackson makes clear that he believes in fitting the right method to the task at hand.Thought provoking and well written, the book borders on philosophy (epistemology) and predicate logic in places, something I enjoyed. Jackson's analogy between predicate logic and assembly language is most apt.Highly recommended.

This book approaches to attack the essence of Software Construction rather the accidents. It is arranged as a lexicon(dictionary) so browsing for a specific topic is simple. It also provides a nice vocabulary for specifying software systems.Main themes seem to be: 0. People who make software should become good at describing the domain, the domain-system boundary and the system itself. Talks about partial descriptions, multiple descriptions etc.1. Strong problem focus. Knowing what to solve is more important than how to solve the wrong problem. All in all well worth for the language, insights and original content.

Each of Jackson's books has expanded on the previous one. PRINCIPLES OF PROGRAM DESIGN was about program design. SOFTWARE DEVELOPMENT was about systems development (specifically, for information systems). Now, SRS takes it further, and considers techniques for specifying different kinds of systems, and systems that require multiple viewpoints to specify them. One of the main points of the book is to promote the idea that many systems must operate in multiple "problem domains" and so require multiple techniques in their specification. (Does your information system also have security/access requirements in addition to its functional requirements? Then you have a "multi-frame" problem, and you should read this book.)The second major emphasis of the book is on describing the "application domain" -- the real world where the system's users conduct business -- before embarking on designing the system. That is, on understanding the problem befor you jump into trying to solve it. If you've ever worked on a huge system development project that turned out to be a catastrophy because half way through the project you discovered that there was a whole area of user requirements that you didn't even know about, then you should read this book. Find out what went wrong and get some ideas about how to avoid it in the future.This is an excellent book about specifying computer systems, and I can recommend it to anyone. It is especially highly recommended if you enjoy thinking about software development at a fairly high, abstract level (a level above the one where you are designing code and database structures).

I loved this book. It is one of the rare books that can be read for education or for entertainment. The book is written for software engineers and architects. Jackson's thesis is that most of the battle in software engineering is in describing the problem to be solved, not in solving it. When the problem is clearly represented, then the solution often falls out. He presents universal techniques that are independent of analysis and design method. He provides systematic, disciplined methods for thinking about and communicating about problems of any sort. Thoughtful readers will enjoy it greatly.