Multi-character tries for text searching

摘要

Full-text searching is likely to become an even more important activity in the future as the amount of information in computer-readable form increases and our demand for information grows. The CD-ROM provides a cost-effective medium for storing large quantities of information. Unfortunately, its physical characteristics limit its retrieval effectiveness. As a result, new storage structures, especially those uniquely suited to the CD-ROM, need to be considered for improving retrieval times. This paper introduces the multi-character trie as an index structure for full-text retrieval on CD-ROMs. The structure offers several advantages when compared to other structures currently in use. Experimental results provide several comparisons with the widely used B+ tree. Comparisons are also made with other structures used for full-text retrieval.