I recently reviewed another book in this series, Teach yourself Borland C++ 4.5 in 21 days [1], which devotes about half its space to the Object Windows Library (OWL), Borland’s object-oriented approach to Windows programming. While the current volume is almost as long, fully 90 percent of it deals with OWL. As far as I know, it is the most comprehensive third-party survey of OWL 2.5 available. By the time this review appears, however, OWL 3.0, designed for Windows 95, will have been released. Also, while the current version of Borland C++ is 4.52, and version 5.0 is supposed to reach retailers’ shelves before the end of 1995, this book uses 4.0. It seems that in the fast-moving world of commercial software, the documentation never manages to catch up. On the other hand, new versions often add only a few new features and, most important, usually do not tamper too much with old features, so dated documentation is sometimes usable, depending on the programmer’s needs.

The book asserts that while the reader should know a structured language, no C++ experience is required. Since only about 60 pages are devoted to C++, this statement is probably false for all but the most talented. Although OWL generally shields the programmer from the Windows Application Programming Interface (API), Spencer admits that the book does not substitute for an API reference. In truth, to use this book effectively, a good knowledge of both C++ and Windows programming (in particular, API) is needed.

In this series, the reader is supposed to cover a chapter a day for 21 days. Here, there are 23 chapters, the last two labeled “extra credit.” The first three chapters survey the Borland IDE (Integrated Development Environment), object-oriented programming, and C++; the last two deal with OLE (Object Linking and Embedding) 2 and the Borland Database Engine; and the middle 18 focus on the nuts and bolts of Windows programming in OWL.

OWL programmers will likely find this book invaluable, if only because nothing comparable exists. I have some reservations, however. First, no one should expect to absorb this material in 21 days. Second, the author is affiliated with Borland, and part of his burden is to persuade his readers that Borland’s way with Windows is the best way. Having had some exposure to OWL, MFC (Microsoft Foundation Classes), and conventional API programming in C, I remain unpersuaded (although I am not prepared to argue for the other approaches either). Third, despite the generous space available to the author, gaps often occur in his presentation, which relies too much on example and not enough on exposition. This is a significant fault, since official OWL documentation (which frequently borders on the incomprehensible) tends, says the author, not to reflect the latest changes and fixes; therefore, it also forces the programmer to learn by example, a method best suited to young children.

Windows programs--even OWL versions--tend to be long, so the happiest news is that a disk containing all the source code accompanies the book.