Windows 7 / Getting Started

---

Initial Configuration

When Windows 7 is installed on a computer, the Windows Search engine configuration code performs the following steps:

1. The Windows Search service is started.
2. The system catalog is created (the first time).
3. The file, MAPI, and CSC protocol handlers are registered.
4. Predefined system exclusion and inclusion rules are added (the first time).
5. Predefined start addresses are added (the first time).

Search in Windows 7 and Windows Vista is language agnostic, but the accuracy of searches across languages may vary because of the tokenization of text performed by features called wordbreakers. Wordbreakers implement the variable tokenization rules for languages. Wordbreakers will break down both the language of the indexed text and the query string; a mismatch between query and indexed language can cause unpredictable results.

Windows ships with a well-defined set of wordbreakers and includes new tokenization code based on the Lexical Service Platform (LSP) for some languages. For others, LSP will delegate to the class wordbreakers. If no wordbreaker is installed for a language, a neutral white-space breaker will be used. (Windows provides wordbreakers for 43 different languages.)

At index time, the IFilter should determine the language of a property or chunk. If a file format does not encode language information, the language auto-detect logic is used to determine the language. At query time, the calling application (the shell, for example) designates the locale of the query.