Enter your mobile number or email address below and we'll send you a link to download the free Kindle App. Then you can start reading Kindle books on your smartphone, tablet, or computer - no Kindle device required.
To get the free app, enter your mobile phone number.
Other Sellers on Amazon
+ $3.99 shipping
Clouds to Code Hardcover – January, 1997
While there are plenty of books on software-engineering case studies, most are big on theory and short on real-world detail. Despite its vague title, Jesse Liberty's Clouds to Code is anything but hazy. The author writes with real candor about his experiences with two large projects--one failed and one successful.
The book's most interesting description is of the demise of Ziff-Davis's Interchange, a never-released online subscription service. (It was a victim, the author says, of too much funding, too many players, too many meetings, and ultimately, the rise of the Web.) The author's proverbial two cents on management styles in software ("Keep teams small and let programmers program") is peppered with the clear vision of hindsight and experience and dispensed with a good deal of wisdom and humor.
The brunt of the book details how to design and implement an automated phone system (for calling thousands of users to disseminate weather warnings and other information) that was developed for a client company by the author's consulting firm. The author chronicles the entire project, from meetings about initial requirements, analysis, and design to implementation and deployment over a ten-month period. Instead of relying on fictitious projects, the guide provides excerpts from actual design documents at different points in the software engineering cycle, along with excellent commentary on key design decisions (such as what platform, what language, and what components to use). As it turns out, the automated phone system you have built uses Microsoft products throughout, such as NT, Internet Information Server (IIS), and SQL Server, as well as Visual C++ and Microsoft Foundation Class (MFC) for the coding. All of this provides an exceptional real-world picture of the software design process in action, along with compromises and imperfections in the final "initial" release. Any software manager could benefit from this often engaging and candid text.
Top customer reviews
It is very rare that you can find a real life project completely documented in one book. This by itself makes the book valuable. The only criticism I can give is that this is a fairly small-scale project (one customer, one location). The author does reflect on this subject and does consider scalability. However, I fear that the reader gets a false impression that this project can be easily scaled up. In this respect I disagree with the impression the writer gives. Nevertheless, a brave attempt and valuable book.
By the way, some others critics have some tough criticism. To me most of them are biased. His choice of an 'all Microsoft solution' might not be yours but it is a solution! IT is not science, there is no single right way of doing things.
I am biased too, that is, I would have chosen a different architecture, design, tools, and implementation model. But that is not the point!
Jesse is very savvy on the business of software, like no one I've ever seen--or at least not as bluntly. He also understands the essence of true free enterprise which I applaud. Some of the reviews of this book actually prove one of his points. I've been amazed for years about how many technology bigots are out there. "I'll do anything as long as it's not Microsoft!" How idiotic. They've produced some good software and they've produced steaming electronic piles of crap. These days I'm quite technology agnostic--I use whatever I can to make the customer happy! I do not see that Mr. Liberty (great name, BTW) gives too much nod to MS...he even suggests his hesitation about using VSourceSafe. I concur, it's not one of the better SCC systems.
My suspicion is that most so-called developers really don't know how to build projects, start-to-finish but how many of us will admit it? I was surprised, and refreshed to know that I was basically on the right track, though Jesse congeals some things and illuminates others in a way that most people can comprehend.
Summary: if you want thick over-inflated theory, there are tons of Rational publications to jam up your synapses. If you want simple, solid principles on how to build reliable systems the red book's for you!
While the first third of the book lived up to my expectations, the last two thirds were heavily bogged down in the implementation details of his project. These details, while important to documenting Liberty's own project have very little value added to the case study.
I also have extreme doubts about the software development process the author advocates. Liberty claims the project was successful because he delivered something on the due date. In fact the delivered product probably had only 2/3rds the desired functionality and (by back of the envelop analysis) was 50% overrun.
The "give it to a single guy and let him hack like a crazed weasel" school of software advocated by the author leaves me cold. Software development has a strong team dynamic element. The author proposes dealing with this element by not having a team at all. Perhaps fun for the individual, but not realistic in "the real world." In the end, Liberty has to engage several other people to produce significant parts of the project that he didn't have the resources to do himself.
It is good that case studies are becoming available, but given it to do over again, I would not have spent the money on this one.