Customer Reviews

Constructing the User Interface with Statecharts

5 star:		(6)
4 star:		(3)
3 star:		(0)
2 star:		(0)
1 star:		(1)

 

 

It is important to realize that this book is not about designing user interfaces. It states that quite clearly in the beginning and with a title like this it is not suppose to be. The book assumes that you know some user interface design principles and are applying them in the general screen designs.

Although it is fine that the book assumes you have the design skills, it is unfortunate that it in its examples does not use better design skills. Some of the screens/forms that are constructed in the book could also serve as good examples of how not to apply design knowledge - although there probably are worse. The author makes it clear that the book is not about the design of screens, but using well-designed screens as examples wouldn't hurt. It is however not big enough of a problem to degrade the star rating.

From a practical point of view the approach suggested in this book is very usable. Although one may design a complete system in this way, it will also serve useful in sub-systems and even individual screens. In essence this book deals with the problem that programmers started to face when event-driven programming became the norm. Before event-driven programming could predict the order in which pieces of program code are going to run. With event-driven program the events may happen in any unpredictable order. For the user interface designer this presents challenging problems regarding the currency and consistency of the user interface. This is the main problem addressed in this book.

The approach is based on splitting the presentation issues in two layers: user interface objects and control objects that manages the behavior of the user interface objects. In essence the modeling of the control objects are done using state charts.

A clear description of the approach is given. The handling of state charts is pretty much inline with state charts in UML although certain "customisations" are being introduced to cater for specific requirements. These are, however, fundamentally sound. The nicest aspect I suppose is that the book doesn't get stuck on the semantic aspects of state charts but provides very practical examples of how it can be used to model the behavioural dynamics of systems.

I found the book easy to read, especially as far as the basics are concerned. It is divided into four parts. Part 1, represented by chapters 1 - 5 (57 pages) deals with motivating the need for such an approach. Chapters 6 - 9 (Part 2) introduces the notational aspects of the approach in a reasonably condensed 50 pages. Part 3 consist of three chapters that provides the state chart designs for three case studies. Part 4 (Chapters 13 - 15) deals with how to take this from state chart notation to physical code - in my opinion the crux of the problem. Several appendices are used to give more detail on the case studies presented in the book - a fact that certainly helps with the readability of the book.

I managed to convert the "code examples" given with reasonable ease to a product such as Visual Basic. I believe that it is written in such a way that the technique could easily be applied to any event-driven programming environment.

The book certainly fills a gap in the body of literature surrounding user interfaces and event-driven programming. It is one of those books that I believe is worthwhile reading just because of the ideas behind it. To which extent you will want to apply the techniques depends on circumstances, in particular the complexity of the user interface, but the book will make you think about formalizing the user interface construction process. If you are constructing complex user interfaces regularly this book is a must read.

The book is essentially advocating a modfied model-view-controller architecture, and using statecharts to design the controller objects. The statecharts are being used to coordinate the behaviour of user interface components in the system. If you use the method described in this book then what you will get is a precise picture of the behaviour of your controller objects. The book is good because it actually shows you how statecharts can be used to model the dynamic aspects of your system rather than just describing the syntax of statecharts. It provides useful advice on how to code and test statecharts. It also makes some useful extentions to statecharts such as prioritised events, parametised states, and transient states. The book would be better if it emphasised the use of components more - but don't let this put you off.

The first real attempt to demonstrate how to construct a rigourous controller layer using UML Statecharts. Adopting the principles in this book will lead to much better quality software with fewer bugs and improved interaction. The only pity is the psuedo code examples. Some real Java or VB would have been better.

The technique described in this book is great. I used it to design a java applet recently and I found it reduced coding to a trivial task. The power of the technique lies in specifying the "syntax of interaction" before the user interface is coded.

I strongly recommend this book.

I really learned quite a lot from this book about statecharts, enough even to write my own book using the concepts to build device simulations (Flash MX for Interactive Simulation)!

The author gives compelling arguments for adapting the design methodology, and walks the reader through progressively more complex examples. The author's writing style is easy to read yet not fluffy or wordy.

If it were not for two points, I would have given this book 5 stars. First, I think he does not tie in the idea of the UCM (User interface, Control object, and Model layer) into the statecharts enough to show how they might fit together, or maybe I'm just a little dense. Second, his nod to coding statecharts should have been left out of the book -- it really didn't help me figure out how to code the charts by any means. In my book, I explore these areas pretty thoroughly.

All in all, though, a very worthwhile read that I recommend to all my colleagues who want to learn about the power of statecharts.

Most business systems are split into several layers, typically: Presentation objects, Business objects, and Data objects. Most object oriented design books focus on the objects in the business layer and the data layer. This book is different because it is entirely focused on the presentation layer. The book describes how the presentation layer split into two layers: user interface objects and control objects.

The book provides several case studies that show how to design the control objects using statecharts. The statcharts provide a precise description of how the user can interact with the system.

In modern systems more than half of the development effort - sometimes upto 90% of the effort - goes into implementing user interfaces. Remarkably, this is one of the very few books that provides significant coverage of user interface design and implementation. It brings a much needed rigour to user interface construction that other authors have ignored.

This book makes developing user interfaces incredably easy. Using statecharts to capture the behaviour of a user interface before any code is written is such an obvious idea that I can't believe it hasn't been written about before.

Forget about Use Cases, use statecharts instead because they can be converted directly into code.

I have lost this book, and re-order it.

What is in there is simple and very effective. Few computer books have this kind of impact. The code examples are not elegant, use a SateChart engine instead (built your own, or use something like xjCharts...).

Way of the future (but don't forget to provide your SateCharts diagrams with your code :-)

Great little book !

I don't know if the other reviewers have read the books carefully or not because the books have been over-rated. First of all, the book is filled with reference stuff to state theory. The useful stuffs are to help people HOW to get a gui done. In these aspects, the book did not do a very good job.

In conclusion, the book should be trimmed down in half and emphasizes more the practical writing of guis. A few lines of code make to little to understand if this tecnique is useful for REAL DEVELOPMENT.