Customer Reviews

Safeware: System Safety and Computers

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

 

 

This is an OK book that should have been an excellent book. Frankly, it was a big disappointment. The author is the grand poo-bah of software safety, and so I expected the book to be deep and profound. Instead it is rather shallow with lots of generalities. A lot of what is said is the type of common sense discussion that I would expect from a social scientist, rather than a scientist. I was particularly surprised to see but one page on independent verification and validation. This is a very important topic for ensuring software safety, yet one evidently not on Professor Leveson's radar screen.

This book is among seminal works on system safety and safety engineering and should be read by anyone who is entering the field. One may not necessarily agree with Prof. Leveson's views (e.g. her often unfair and unjustified dismissal of european approaches to safety engineering) but her views in the field are simply too important and influential to be ignored!
The book is a good "bed time reading" that gives an overview of the field, major problems and introduction to some major approaches. It is however not a handbook per se - you will not understand any particular method sufficiently to apply it.
Overall, very good introductory text with a good coverage and a very easy and clear language!

The key to understanding safety lies in the understanding that no one component failure or no human error ever occurs in isolation - an accident is a result of some systemic problem, and this is the fundamental theme of this book. No single book could ever address in detail all of the finer points of system safety analysis (indeed this would take a series of books) but Leveson is able to capture the core issues in an engaging manner. This book is a must read for any student of software and system safety, and will continue to be relevant in an age where automated technology is ubiquitous.

This book is a starting point of what is important to make a system safety.
This provide over all view of system safety not software safety.
The case studies in appendix are very important for safety engineers.
Appendix A is the therac-25 story on Medical devices.
Appendix B is Apollo 13, the DC-10 and Challenger, Aerospace.
Appendix D is windscale, Three mile island and Chernobyl, Nuclear power.
These are very bad accidents, so we should study more deep.
The naming "safeware" is very good wording, this is not safe software nor safe hardware.
Go ahead, after this book. There are many resources about sytem safety and software safety.
We have already many experiences about HAZOP.

ps.
2011/1/18, we have a Workshop on Critical Software(WOCS2011) with Nancy.

Interested in the development of high integrity software? Looking for a system safety engineering text? This is not the one for you.

This book won't describe when to use a sequence number or what size CRC is appropriate for your safety-related communication, or what level of code coverage is appropriate or what level of requirement coverage is appropriate for your identified hazard etc. It won't describe the actual versus perceived issues with OOA/OOD, what aspects of a hard real-time design require attention or how to meet the expectations of an independent assessor or certification agency.

The principles of software design assurance/software integrity can be garnered from the standards such as DO-178B, and DO-248-style discussion papers or from CENELEC 50128 and good old MIL-STD-882B to name but a few. Alternatively a free start can be achieved by downloading the Joint Software system Safety Handbook from the US DoD.

Save your money for the works of Moriarty (executive-level), Ericson (introduction/supervised practitioner), or Dunn (software controlled system safety intro) if you must buy something.

But if you're looking for some stories to put into your slick safety presentation or seminar then it might be just the right book.

A very good collection of case studys and commentary and analysis
on software safety, and software failures. Information like this is rare, and hard to find, yet invaluable.