Written for the advanced C++ developer, Learning DCOM looks beyond wizard-generated code to teach you how you to build most of the C++ code yourself in order to create effective distributed Component Object Model (DCOM) components. The book commences with a tour of distributed computing, from the early days of terminal emulation, to the Remote Procedure Call (RPC) and onward to today's DCOM. The author includes a rich introduction to COM, from objects and built-in and custom interfaces to important concepts such as containment and aggregation. Though somewhat densely written, these chapters on the details of DCOM expose its real inner workings with standout material providing a full treatment of the different thread apartment models.
The second half of the book focuses on existing wizard-based code (using Visual C++ tools for the Active Template Library [ATL] and Microsoft Foundation Classes [MFC] COM components) starting with an ATL server-side optical character recognition (OCR) component. The author then presents client-side programming strategies for COM, far beyond tapping built-in Visual C++ capabilities, that incorporate custom template-based smart pointers for calling COM objects.
The book then turns to Web development where an OCR example is used within an Internet Explorer Web page. Sections on security demonstrate how to cooperate with Windows NT, as well as auditing and administrative options. Unfortunately, coverage of Microsoft Transaction Server (MTS) is omitted here, arguably one of the most critical aspects of writing scalable Web applications. A final chapter on event handling (and connection points) shows off how to process events with distributed components.
Notable for its considerable technical depth and detail, Learning DCOM gives the advanced developer the inside track on creating state-of-the-art DCOM components. --Richard Dragan
