Summary

• Windows 95 encodes characters using the code page model inherited from Windows 3.1, though Windows 95 can support more than one code page at a time. (See Chapter 6.) Single-byte code pages are limited to 256 characters. Each double-byte code page, which is what Far East editions of Windows 95 use, covers several thousand characters.
• Double-byte character sets on Windows include a mix of 1-byte and 2-byte characters. When parsing strings in a DBCS environment, always treat 2-byte characters as a unit. Never separate a Lead Byte- from its trail byte.
• Windows NT encodes characters in Unicode which is a fixed-width, 16-bit industry standard that encompasses most of the characters currently used on computers. Unicode can simplify the process of sharing multi-lingual data if your goal is to create a single, international-aware code base for Windows-based applications.
• Because it is only a character encoding, Unicode does not directly address sorting, font, or layout issues.
• Programs that create Unicode-based documents can use compression algorithms to keep files from doubling in size.
• Commercial products based on the Unicode standard are becoming available. Even if you do not choose a full Unicode implementation for your Windows-based software, add support for converting Unicode data.
• In the long-term future, all Microsoft operating-system technology will be based on Unicode. At present, the Win32 API string-handling functions exist in two forms: one that expects string parameters to be in Unicode (-A entry points) and one that expects string parameters to be expressed in Windows code pages (-W entry points).
• Windows NT supports both -A and -W API function calls, whereas Win32s and Windows 95 support only -A API function calls. You can create a single source-code base for both types of function calls by using generic prototypes.
• The Win32 API contains two functions, WideCharToMultiByte and MultiByteToWideChar, for converting between Unicode and Windows code pages. Windows also supports clipboard formats for converting between Unicode and other encodings.
• Both the Win32 SDK and Visual C++ 2 ship with sample code that can help you learn techniques for programming for Unicode.
• The Visual C++ 2 run-time libraries contain functions for processing wide characters. The same functions also exist in versions that process single-byte characters and multibyte characters.
• By following several basic steps, you can easily change existing programs so that they can use Unicode. You need to convert code to use wide-character function calls and data types, revise code that assumes that characters are 8 bits wide, and add support for special Unicode characters such as the byte order mark or combining characters.