Smart Search

Nowadays, almost any commercially available hard drive can store more text than a whole library. In the digital world, a traditional system such as a card catalog or a knowledgeable librarian is no longer adequate to help find the right shelf. Even software equivalents such as find or zgrep are not always fast enough to track a particular piece of information amongst giga- or terabytes of data.

The science that deals with this type of search problem is called information retrieval. Computer scientists have developed sophisticated methods for tracking down files that users don’t even know exist. The free Java library Lucene implements some of these methods. Doug Cutting published an early version of Lucene in 1999. Two years later, the project, which carries the middle name of Cutting’s wife, came under the auspices of the Apache Foundation when it joined the Apache Jakarta Project. Lucene has been available in Version 4.0 since October 2012. The index file structures are backward compatible, so the transition from 3.6 to 4.0 does not cause any problems. Over the years, Lucene has become one of the most widely used solutions for indexing and searching text. (See the box titled “Lucene In All Its Facets.”)