If you want to keep old ANSI behaviour, just delete _UNICODE and UNICODE macros from prepocesor definitions, altough new applications should use only Unicode these days.
The recommended way is to include <tchar.h> and use TEXT() macro for string literals and TCHAR instead of char, this way your application will work in both Ansi and Unicode mode without changes to source code.
I don't know about VS 2010, as I have VS 2008. But in VS 2008 you can go into the project properties and set what modoran above suggested using the list box they provide. In mine the choices are multibyte, wide char, and not set. Using the GUI like that just changes the macros I imagine - like modoran suggests. In the end though, you need to fully understand this character set issue and learn the TCHAR macros. In the short run though, so as to help you use the ansi book, I'd just change the char set in use.