Users performing a search are allowed to supply strings that the registry needs to parse according to the search algorithm described in the TAD. In order to convert a user-supplied string into a meaningful search, the application should convert it to single- or multi-word, normalized tokens of specific types.

For example, a search string including a five or nine digit number which is separated from other words probably indicates a US postal code. During the parsing and tokenization process, that portion of the string should be classified as a postal code, so that the higher-level search function can make decisions based on which types of token appear in the user's search string.