15. Searching and Database on the Web

15.4. Searching

search language and search engine
associative search (keyword matching)
boolean search

The searching module contains two parts: search language and search engine. The search language takes query from the user and transforms query into representation that can be applied to the index. Then the search engine applies the query representation to the index and derives a set of records that meet the criteria specified by the query.

Because computers cannot easily discreen the meaning of any language some types of search language or query language are needed. There are two basic types of search languages on the Web:

associative search (also called keyword matching),
boolean search.

Keyword matching is supported by almost all search systems. In the text keyword matching is really pattern matching of characters. For example, a search query of the string "comp" would yield all occurences of words containing "comp", such as "composer", "computer", "incompetent", and "composure". What if we are looking for information only on computer? The example shows that keyword matching search often yield a broad rang of matches, which may cover what we are not looking for. To narrow keyword matching search, one would take approach of boolean search.

Boolean search is a more sophisticated search language that narrow and refine keyword matching searches. If the users wants the set of results closer to what they ware looking for, boolean search may get better results than keyword matching search does. Boolean search allows the construction of sentence like queries with combinations of boolean operators. The boolean operators are : AND, OR, and NOT. They are used to build the search query, and can be used in combinations.

The objective of search engine is to applies the query to the database index, to find matches between the value of attributes in the stored records in the database and the attribute specified by the search query. If a match is found, then that record is retrieved. As a result of search query, the set of retrieved records is called result set.

WAIS (Wide Area Information Server) is the most popular indexing and searching method currently used on the Web. WAIS is a client-server internet service and was developed independently of the Web. WAIS offers a sophisticated searching language and search mechanism. It uses keyword search with optional use of boolean operators. Traditional databases are also very popular tools for indexing and searching Web resources.

Aixiang (I Song) Yao<ayao@csgrad.cs.vt.edu>
Last modified: Sat Oct 3013:15:51 1996