Customer Reviews

Modeling Software with Finite State Machines: A Practical Approach

5 star:		(2)
4 star:		(2)
3 star:		(3)
2 star:		(0)
1 star:		(1)

 

 

This book is excellent for making the reader understand how a complete software application can be easily controlled using state machines, simplifying the architecture. The writing can be hard to follow sometimes and the authors make claims about areas they are not expert in that are clearly false but when it comes to state machines and developing applications using them, they clearly know what they are talking about. Just skip the non-state machine parts of the book.

The title "Modeling Software with Finite State Machines", and the accompanying web-site [...], proclaim a new and fabulous way to design, develop and implement software-solutions. This method, so the authors state, will bring the engineering back into software development.

To my mind, those statements are ludicrous, very biased and uninformed and so it made it really very hard for me to read the book. The tendentious style penetrates pretty much the entire first third of the book. As however I needed a good overview of practical use of state-machine (for a non-hardware problem), I persevered.

Chapters 4, 8 and 9 gave me what I needed, though again the material is presented in a pseudo-academic, know-it-all style. The academic material on Finite Automata I have read tends to be pretty impractical, whilst this book takes a practical approach and as I get the impression, that the authors are experienced in their fields, that was good enough for me.

The last third of the book (Chapters 10 to 17), focus on StateWorks. As I do not intend to buy that product, this gave me little information I could use.

As others have added, the first 1/3 of this book is a very opinionated editorial on the software industry, the middle 1/3 a pretty good presentation of state machines, and the final 1/3 simply a sales pitch for Stateworks.

My biggest disappointment is that there is little in the way of code examples demonstrating their principles in this book. Obviously, they would rather you buy Stateworks instead.

Stateworks - I think you'd be crazy to stick a proprietary black box bunch of code into a major project, but that's just me. No idea what the license costs, but the price of this book as the the main sales literature isn't cheap.

Suggest checking out the SMC compiler at: smc dot sourceforge dot net. This open source code combined with some of the general design principles found in this book set me on my way. If you can't afford the book, you'll probably do pretty well with just the SMC material.

The subject of this book is probably one of the most written about in software industry.

The first three chapters is authors' view on just about every thing in software development including a brief rundown of programming languages (with healthy dose of authors' anecdote), and (somewhat biased) a view of development methodologies including Agile.

The rests are introduction to FSM, discussions about Moore and Mealy machines, design considerations and case studies. Finally in Chap 10 is about the StateWorks, the software they're selling (By purchasing this book you're entitled to a copy of their light version of StateWorks software. I didn't evaluate the software)

The book takes pragmatic approach to design and analysis of state machine. There are a lot of practical advises, considerations that could only have come from many years of experience applying the principle in practice. On the other hand, I don't think the first three chapters serve any purpose to the entirety of the book. Thus I'm giving a four star.

I completely agree with the review by John Nash, these authors are experts in state machines and they have some very good info and shared experience regarding modeling problems using FSM. However, the authors are extremely opinionated and biased, they frequently trash any method that doesn't 100% comply with their own thinking even if the method is almost exactly the same as theirs save a few small details.

The first few chapters of the book are an overview of the programming landscape in general, and it is clear that the authors know a great deal about some subjects and have good insight to share. On other subjects however it is very obvious that they have not really delved into what they are talking about yet they make broad generalizations, often negative unless it is a technology employed by their product, StateWorks.

Also, they present StateWorks as a panacea for all software ails, not just embedded development or control systems, while it is plainly obvious that this is not the case.

This book has 17 chapters, only 3 of them (7-9) and a few of the appendices are worth reading for the general reader. If you are intending to use StateWorks then a few more of the chapters would be useful.

Extremely difficult read:

This book appears to have been written by four world class experts in the field of Finite State Machines; apparently only one of which was born in a first-language-English country. Their target audience seems to be the other four world class experts of Finite State Machines.

After 20 years designing embedded systems (with no OS), I faced several thoughts in how to implement a serious and flexible way to deal with small, medium and big size softwares.I'm sure the one of most important disconery I got is the use of FSM/FSA.It makes your sw clear and visible, without tricks wich only the author can undesrstand them.This book covers this subject with very simple and direct approaching, making clear that there are not a MAGIC and COLLORED way to develop a safe and robust software, unless you focus and udesrstand the behavior of you are controlling, wich is tottaly feasible and confortable if you apply FSM to accomplish it.The desmystification of ilusions based on new ideas wich are not new ideas, leads the programmer to see software as it is: a not easy task wich must be done by programmers and not hobbyists.Of course', I'm not talking about to control bips and leds. Very good book...very good and experienced writers...excelent for "embedded guys", like me...wich develop Ethernet TCPIP stack , GPRS TCPIP stacks, applications tasks based systems, only using the good and powerfull C language

I am not a computer scientist by profession but have worked in the systems and networking area for some years now and always had an interest in software engineering. The book opens up a whole new perspective for a hobbyist programmer who is self taught. I will never look at software the same and have changed my way of thinking when it comes to developing programs. This will be a reference book on my bookshelf for a long time.