Superdistribution is the most important software engineering book of this decade. It is controversial, because it locates the difficulty of software engineering not in development processes or tools---the focus of 99% of the software engineering community---but in the way that software is bought and sold.

Cox's claim can be summarized in four points: 1. The reason that software is costly, of low quality, and difficult to construct is that we build it rather than assemble it from prebuilt components, the way that every other engineered product is constructed. 2. the reason we build rather than assemble is that there is not a robust market for buying and selling components. 3. The reason there is not a robust market for components is that there is no standard mechanism for pay-per-use of components. 4. The reason there is no standard mechanism has to do with the difference between information and atoms

Get it? Neither did I at first. But I am conviced he is right about all four points.

Cox also offers a solution to this problem, a "superdistribution" mechanism that provides pay-per-use. But I think the real value of the book is its compelling explanation of the problem.

David Bridgeland

Powersim Corporation

The first three chapters are conceptual, and compare various approaches to object-oriented programming. Chapters 4 until 8 are highly technical and give detailed information on the Objective-C runtime and class libraries. I find the chapter on user interfaces a bit sloppy. Interesting, on the other hand, is the final chapter with projects for extensions.

The book is kind of small.. but its packed w/ info onobjective-c. It gives complete info on how to use the objective-c OOPextension as well the objective-c internals.

This book is a must read for anybody working in Objective-C (NextStep/Rhapsody). It basically describes Objective-C, how it works, and future directions for the language. It also has a good comparison of other OO languages. Overall, it's a great book