design and development

First off, if this book deserves 5 stars just as a recognition of the depth of the accomplishment (given the breadth of the undertaking) in getting it written, and written extremely well. In an industry where the most successful authors are hacks that put out paper thin salvos to ride whatever the new gauche rave is, this book represents the far opposite end of the spectrum. Clearly, the long history of CASE and modeling and code generation is well in hand as the author pushes through a detailed elaboration of where MDA stands and where it's going. And the news is balanced and delivered honestly, this is not a call to Kool Aid coming from a deluded cult leader. There are open issues, but the general direction of MDA is so overdue and important it's silly. [For a good means of achieving parallax on this issue, consider looking at Cheeseman and Daniels' excellent UML Components, which makes a strong argument for a level of indirection between logical and physical models. If MDA delivers no more than that in the short term, it will still be a huge contribution.]

The reversibility of those models and the degree to which synchronization and editing can be simultaneously supported is perhaps the key issue that remains. Unfortunately, it is something of a deep bind because there is no doubt that the dream of complete reversibility of code and model, which has been hyped a lot over the past 5 years, is a shallow dream (reducing the 'model' to a mere visualization), and yet the idea of not being able to touch what was generated has a similarly stultifying unappeal to it.

As the development world is becoming more polarized between those who would model solutions and the band of hackers who claim to be able to do anything with a scripting language and a database, MDA is a key focal point for the former that is long overdue. Way too much attention has been spent developing tools that make it easy to lay out user interfaces while very few places have achieved even a basic ability to keep a serious domain logical model in tact through a single version, let alone a number of generations. Hopefully the tremendous consensus that has crystallized around MDA is an indication that that balance will soon start to change.

I have had the pleasure of working with David Frankel for a number of years on Object Management Group efforts, so I expected a lot from this book. I wasn't disappointed.

Dave has written a clear, pragmatic guide to what MDA is and, more importantly, what really can (and cannot) be practically accomplished with MDA today. He unerringly focuses on the highest payoff areas for most projects, such as the generation of code from data models. He also brings his years of experience in developing enterprise systems to bear, clearly describing the specific issues involved in applying MDA in this difficult area.

The book gives a thorough presentation of the concepts behind MDA -- including the clearest discussion I have seen anywhere of OMG's Meta Object Facility, a perennial topic of confusion. Nevertheless, I don't really consider this a book on "MDA" as such. It is, indeed, a book on APPLYING MDA, as the title states.

If you are looking for a more theoretical presentation or a grand vision of how MDA will work someday, you may be disappointed. But if you are looking for techniques you can start applying the week after you finish the book, this is the book you want to be finishing.

I am currently Chief Architect at a company that is in the process of making the cultural and technical shift to model-driven development. I found this book so relevant to where we are and the next steps we need to take toward MDA, that I had the company buy copies for all our architects, plus a few extras to circulate among the developers. I even had my boss (the company president) read Part One, which provided just the right level of overview for him (plus Michael Guttman's forward, which is a fun read in itself).

If you are in a similar situation where you work, I couldn't suggest a better book as a helpful change agent. And if you simply want to know how to start applying MDA techniques for enterprise development, this is where to find out.

I read this book after reading MDA Explained by Anneke Kleppe and found it wandering into many areas but not deep enough in any one of them. While this book is a decent attempt to bring forth the impact of MDA in enterprise computing, a reader looking to understand "what" MDA is and "how it" works would be thoroughly disappointed. MDA Explained is a far more useful (and thinner) book that is not only more readable but also lucid in explanation.