The first practical guide to Microsoft's .NET Common Language Runtime!
Compiling for the .NET Common Language Runtime (CLR) is the first book to give software engineers all the details and guidance they need to create full-featured .NET compilers or implement programming languages for the .NET environment. Renowned compiler expert John Gough presents an unprecedented "under the hood" inspection of Microsoft's .NET CLR, illuminating the key concepts underlying Microsoft's new technology and demonstrating exactly how CLR maximizes interoperability across multiple programming languages. The book uses a running example to illustrate its key points(Gough's own open source Gardens Point Component Pascal compiler(which can be used as a basis for sophisticated experimentation.
If you need absolute mastery over the .NET CLR, only one book delivers what you're looking for: Compiling for the .NET Common Language Runtime (CLR) by John Gough.
As a young person, I thought it quite reasonable that one should try to understand every detail of the technology that underpinned the artifacts of everyday life. It seemed entirely appropriate that as I drove along in my automobile I should appreciate the relative merits of overhead camshafts and the alternative design based on pushrods and rocker arms. Alas, as the catalog of artifacts multiplied I realized that it was simply not possible to understand everything. I must now cheerfully confess that I am slightly vague on the difference between the NTSC color system used by television sets in the United States and the PAL system used throughout most of the rest of the world. In spite of reality curbing my more extreme ambitions, I still want to understand the details of the technologies that are central to my life as a software producer, or at least intrinsically interesting.
At about the same time that I realized one cannot understand everything, I made an even more startling discovery. There are two kinds of people: There are those who want to understand their technologies, and those who frankly don't give a damn. This book is written for people who are of the first kind.
.NET stands at the convergence of three of the hottest technologies for the future of computing: component technology, web-based service provision, and language implementation based on abstract machines. Component technology will finally deliver on the promises of object-oriented programming. Web-based service provision using industry standards such as XML and SOAP will power the new world of pervasive and mobile computing. Finally, program realization by means of abstract machines allows unprecedented levels of program portability as well as enabling verifier-based guarantees of runtime behavior. Add to this the broad industry support for standardization of .NET, and it becomes clear that .NET is set to dominate the immediate future of computing on the desktop and beyond.
The technology that underpins .NET and enables all of the functionality that other books in this series will describe is the Common Language Runtime (CLR). The CLR is the engine that powers the machine, the Bernoulli theorem that explains why the whole thing flies. For those who wish to understand how .NET works, this is the place to start.
In the final months of 1999 a small group of programming language implementation teams from around the world were invited to implement programming languages on what was to become the .NET runtime. The project produced a number of alternative languages to the ones that Microsoft themselves were developing, and showcased the multilanguage capability of the CLR. Paul Roe and I produced a compiler for a Pascal language variant. I learned several things from the experience. I learned that the CLR is compiler-writer friendly, and that a small team can implement a real language for this target in a short time. As well as the learning, the project turned out to be great fun and I developed an enthusiasm for the CLR that I hope will prove to be infectious.
Since the public announcement of .NET and the public availability of the pre-release software, I have been involved in some experimentation with the use of the CLR in education. Wayne Kelly has pioneered the use of the CLR as the target machine for an introductory compiler course at QUT, with encouraging results. For those colleges that advocate a practical approach to the compiler subject the CLR has very much to recommend it.
Since the technology described here is so new, it was difficult to find reviewers who were able to provide feedback during the development of this book. I owe a special thanks to the following .NET pioneers: David Bennett, Simeon Cran, Bertrand Meyer, Erik Meijer, Christine Mingins, Paul Roe, Emmanuel Stapf, and Clemens Szyperski. As usual, any remaining errors are entirely my responsibility.
The compiler that forms the running example for this book grew out of a project jointly sponsored by the Distributed Systems Technology Center (DSTC) and the Programming Languages and Systems Research Center (PLAS) at QUT. Much of the underlying technology for the compiler was developed in an earlier project supported by the Australian Research Council. Diane Corney and Siu Yuen Chan, both doctoral alumni of PLAS, wrote some of the associated tools and participated in design reviews of the compiler.
Writing a book to a tight deadline plays havoc with one's social life. My family has been tolerant of the occasions when I have been absent, and other times when my body was present but my mind was far away. I owe a special thanks to my partner, Noeleen Atwell, for her love and support throughout the whole process.