Software Architecture in Practice (SEI Series in Software Engineering) 3rd Edition

Muy buen libro

Used this text book for a Master's of Software Engineering course on System Architecture. I would say the book is more of a reference than a read and learn type of book. I say this as the first group of chapters (1-12) covers Engineering Objectives and Quality Attributes. For each Quality Attribute, there are a few examples of tactics to solve them (for instance, for availability they discuss hot-spares for swapping out a failed component with an active replacement). Throughout the course, these chapters were extremely useful when building the architecture of an online collaboration system. In addition, the book discuses how tactics (and specifically which ones) cause trade offs and how to analyze them for your unique use. I now use this book at work when discussing the structure of new/existing products and it comes in handy when needing to make a tough decision. It also helps the book is an easy and understandable read.

As a software designer and engineer for over 1/4 century, I am always on the look-out for a process which claims to separate high-quality from low-quality. Software is not a tangible, 3-dimensional, manufactured product, but software has "quality". Initially someone at work started pushing this new "quality" rating methodology for software......distributed or centralized.I have dedicated the last several years to wrestling with "quality".....everywhere in my life but particularly at work with software. When I heard "Architecture Quality Attributes" I was intrigued and bought the book. Each AQA is broken down into Concerns and Common Tactics. Thinking back over the last 1/4 century....they all rang true. Quality in technical design and quality in technical management all come together here. Buy this book.My only peeve......why such a half-hearted writeup Probabilty Statistics, Fault-tree Analysis, and FMEA ? There is a dis-organized and scattered treatment through out the book. Page 84 (3rd edition) discusses AND gates, OR gates, etc etc...and then statistics shows up again from pages 255 - 259. In today's distributed environment, probability statistics is very revealing. The more components bolted together to create a network, the more OR-gates you have that could fail and make your network inoperable. This is the Achilles heel of distributed environments.Ironically, that entire concept is hidden from the reader of this book. Even on page 256, at the very bottom, the authors give a formula for "probability of failure". In fact, the formula is EXACTLY WRONG !! The formula shown is the formula for "probability of success". ( probabilty of failure + probability of success = 1 ). Oh well.Perhaps, I thought chuckling to myself as I put the book down, this is why the young kids keep buying more and more servers whenever they run into capacity constraints.......page 256 assures them that each component approaches zero chance of failing as time approaches infinity and the component is left in service longer and longer and longer....... DOH !! :-)

A must have for software architects

I teach software architecture and I have been using this book as a base text. It reflex the core of the (now maturing t.G.) practice of sw architecture and multifaceted hurdles to overcome in the daily life on an architect. The first 3 chapters are rather dense because they basically resume the whole book. Do not be discuraged, everything will be fine chapter 4 and on.

This is the best book about architectural principles and practices. Clearly described all mandatory artifacts for the successful requirements gathering and transforming into an appropriate design. The instruments for organizing proper orchestra mentioned in the book already became my daily tips and notes to follow. Highly recommend, and I would call must-have book to read for all who involved in Software and Enterprise Product lifecycles.

This book is certainly a good reference for already experienced architects, and maybe a god overview for less senior people. However, I find it far too verbose and far to thick. It would be good if the authors shrunk it to half of its size. Whenever they go into more details (e.g. when they explain concrete patterns or tactics), they actually stay rather shallow. If you want to know more, you have to consult other books. Therefore, these parts could be shrunk even further.On the other hand, whenever they walk the reader through more abstract concepts and thoughts, the text is full of useless contemplation and even bare repetition. Maybe, for these more difficult and dry topics, the authors were hoping that the reader would understand eventually after the 5th bad explanation. Sorry, but I would prefer one or two really good explanation.

So far I liked the book. It really covers everything in a good level of detail. If you’re new to software/solution architecture you can use this book as a strong starting point.