Customer Reviews

Trustworthy Compilers (Quantitative Software Engineering Series)

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

 

 

Again, Dr. Safonov has produced an excellent book on computing. His ability to mix in the history and examples to support area being discussed and build understanding are excellent. This mixture makes the book easy to read while providing depth of understanding for the student or interested reader.

Vladimir O. Safonov's approach to this complex topic is superb. Developer's who are seasoned as well as those who are building experience will find this text rewarding. I appreciated the gradual and logical approach starting with basic concepts of Trustworthy compilers and development principles. This text should be included in computer science curriculum as it gives a superb backdrop/history and exposes new concepts and approaches within the Trustworthy framework. The questions and exercises after each chapter are detailed and when completed will give the student a firm grasp of the applications and theories presented. I highly recommend this book for new developers, experienced developers, and most importantly, computer science and computer engineering students.

I recommend this book by Vladimir Safonov to fellow software developers having any interest in compiler technologies because of its unique combination of the following traits:

-- "Trustworthy Compilers" provides a thorough introduction to industrial-level compiler technologies of all stages of compilation process. Safonov introduces the concept of trustworthy compilation very early in the book and then the material is given with the focus on trustworthy, or as I would phrase "industrial-strength" technologies.

-- Consequently, since the book is adjusted with author's course on compliers and is rich with references to other books and web resources related to the topic, it can be used as a textbook, valuable as for newbies, as for professionals interested in brushing up their knowledge in the given area. One can find it convenient (as I did) downloading author's compiler course slide decks from his site and using then in parallel with reading the book.

-- Due to the book's special attention to the aspect of software trustworthiness (and compiler trustworthiness in particular), as newbies, as professionals can find insightful considerations on particular details of compiler implementation, drawn from author's wide personal experience of a leading specialist of a number of top-notch compiler projects like Sun's Pascal programming system. This experience spans over several decades in past and includes little known to Western readers Soviet era projects like a compiler for Barbara Liskov's CLU language implemented for Elbrus series of supercomputers.

-- Finally, besides abovementioned, the book features a couple of in-depth considerations of particular subarea of nowadays compiler technologies. One is graph compilers; I have read through this part with interest (though it's unlikely that I will ever deal with technologies of that kind). Another one is of a greater practical importance in my case: it is devoted to Microsoft's compiler toolset named Phoenix which is used (accordingly to some articles I've found on Microsoft's sites) in Midori, a new OS incubation project. As far as I know, to this date Safonov's book is the only one describing and discussing this toolset (available for download from Microsoft Research's Connection site) in depth -- so computer scientists and software developers dealing with Phoenix will certainly benefit from a thorough reading of that chapter.

We are teaching one-term undergraduate course on Program Languages Processing at Faculty of Information Technologies (FIT) in the Novosibirsk State University (NSU), Russia. The course comprises 28 hours of lectures, the same hours of seminars and 3 laboratory classes. Since we are theoreticians, the course is very much oriented on mathematical foundations of the syntax analysis, basics of formal semantics (structural operational and denotational), virtual machine and translation to byte-code, formal verification (Hoare-style and model-checking). But due to this professional specific, our course has some deficit of modern compiler technology and tools.

To compensate this deficit, we decided (in the very beginning of the spring semester 2010) to use some lecture hours for abstracting and reviewing a book of Vladimir Safonov "Trustworthy Compilers". The choice was motivated by 3 simple arguments. First, the author is well-known in Russia compiler community for his contribution to several compiler projects. Next, the book is published abroad Russia (by John Wiley & Sons), i.e. it has passed internal review process successfully and and has been accepted by Larry Berstein, an expert in software engineering and the editor of Wiley book series on qualitative software engineering. Finally, the book was "brand new" (it was published just a month before start of the spring semester).

Nine "major" chapters (from 2 to 10) of the book were assigned to volunteering students after two months of our studies (i.e. when syntax analysis, formal semantics, virtual machine and translation to byte-code have been taught already). They prepared short summaries of assigned chapters and presented them during the very last 5 lecture hours of the course. We had 3 sessions when chapters 2&3, 4-7, and 8-10 have been presented by students in 20-25 minutes each. (Please be aware that lecture hour in Russia is 45 minutes, usually we have 2 lecture hours in consequence once a week.)

We can report that these presentation sessions have been very useful for us and all students (for those who reviewed the book, and those who listened them). The content of the book really compensates lack of tools and technology component of our course. For sure, we will use the content of the book in the next couple of years as "technology and tools supplement". Moreover, we can recommend the same for colleagues who teach similar undergraduate courses "from theoretical perspective". We recommend to include in a course the material of chapter 4 about error recovery, chapter 5 - on semantic analysis, chapter 7 - on runtime data representation, and chapter 8 - about JIT compilers. In contrast, we are a little bit skeptical about educational importance of the two last chapters 9 and 10 of the book for undergraduate students. In particular, we think that graphical languages and compilers (chapter 9) are out of scope for undergraduate studies, but can be taught/learnt as an elective topic during master studies. Regarding introduction to Microsoft Phoenix (chapter 10), we think that it will be not easy to master a complex and complicated industrial software optimization and analysis framework with aid of the book, maybe more comprehensive tutorial will help better.

Let us summarize. The book "Trustworthy Compilers" by Vladimir Safonov is a high quality textbook on compiler technology and tools, that can be used (and is recommended to be used) at undergraduate level after course on foundations of program language processing. It is written on base of Vladimir Safonov expertise in compiler construction, in good language, contains very many examples with detailed comments. It makes the book helpful and useful for university lecturers as well as for students who want to teach/learn modern compiler tools and technology better and deeper.

Eugene Bodin and Nikolay Shilov

Novosibirsk State University

I wish to cast a perspective look at a new and progressive state-of-the-art text by Professor Vladimir Safonov (with whom we befriended through Microsoft's Trustworthy Computing Curriculum-TWC grant competition in 2005 as two of the 14 global grantees while we met at Redmond, WA to attend Microsoft scholars' gathering in April 2006)is another precious piece of pearl added into shaping the necklace of the Trustworthy Computing curriculum initiaves after my 2007 Wiley book titled "Trustworthy Computing: Analytical and Quantitative Engineering Evaluation" and Dr. Safonov's 2008 Wiley text titled "Using AOP for Trustworthy Software Development". I do not wish to repeat here what other reviewers have already stated; that is, this book is in very sense of the word a well-deserving antology and full menu all-in-one covering a multitude of aspects of trustworthy compilers toward a more credible software engineering planet. Other than a rich facade of techniques, lexicons, optimizations,semantics, case studies, mechanisms, models, theorems and examples offered in the book, at the Conclusions whereas Dr Safonov modestly confesses that "...We have not considered all topics related to compilers- it would require a series of books... In particular, we did not cover topics such as interpreters.." So there is much more to it than meets the eye, however this book will be an incentivizer and a catalyst for the new generation software engineers to move forward in the right direction of the common goal defined to be the "Trustworthy Computing"as opposed to flawed Haphazard Programming. Once again congratulations to Dr. Safonov for a quality book well-engineered, practically useful both for the academia and software industry as well.

-by Dr. M. Sahinoglu, Director, Informatics Institute, Auburn University Montgomery, Montgomery, AL 36124-4023, USA; EMAIL: mesa@aum.edu; URL: [...].