Content area
Full Text
As the capacity of hard drives grows bigger and the cost of hard drives goes down, many people want to maintain large files of text on their microcomputers. These text files can be the result of word processing, including original letters, memos, reports, manuscripts for books or articles, and the like. Or, they can be full texts and bibliographic records downloaded from various online retrieval systems, or from CD-ROM systems. Optical scanning combined with optical character recognition conversion provides yet another source of building huge textual files. Some textual files may be a combination of all of these data entry/document creation techniques.
Software for textual files have seen active development in the last few years as the need for such packages grows. In this article we review four such packages in-depth: Concordance, Concept Finder, Personal Librarian, and Topic. Before we describe the specific features of each package, an overview of the characteristics of textual files and the requirements for text retrieval software will help put the specific comments into perspective.
UNIQUE CHARACTERISTICS OF TEXT
Some software packages for text are enhanced versions of packages originally developed for bibliographic files, others are newly developed with full text in mind. In either case, the search and retrieval capabilities of these packages are necessarily different from those of software made for other types of databases or files. This is because of some unique characteristics of textual files:
* the data is primarily alphabetic and when numbers are included they are usually treated as text
* the files are often very large
* the file may consist of a single text (e.g., an encyclopedia) or may be a collection of many independent texts (e.g., journal articles on one topic downloaded from an online database)
* the size of the texts in one file may vary from short memos to entire books
* collections of texts in a file may be relatively uniform in format and size (e.g., a correspondence file) or they may vary considerably (e.g., correspondence, reports, contracts, articles, books)
In addition, there may be varying degrees of structure in the texts, including:
* records with fields, such as a typical bibliographic record
* amorphous text with appended fixed fields, such as a file of journal article texts...