A fast parallel search method for large dictionaries

The paper presents a simple but highly efficient parallel approach to the problem of searching on large dictionaries. It is based on the concept of direct addressing a table where the words are stored. To obtain a small table, successive reductions of the number of bits used to represent each word are done by using look-up tables (Translation Tables). Search time is virtually independent of the size of dictionary. Unsuccessful searches always take less time than successful searches. The data structure is very flexible, and can be used not only for English dictionaries, but also for those with larger data sets like Japanese or Chinese