Articles indexed in a database

Book index

"Open" structure: new articles are constantly added to database (expanding universe of information)

"Closed" structure: finished book is a finite universe of information

Index language is created from terms generally used in the subject area of the articles (e.g., medicine); index terms are assigned to articles

Index language is derived from the information in this book only

Indexer assigns index terms from an existing index language, and does not create or change the terms

Indexer creates the entire index language for each book, including the cross-reference and heirarchical structure

Indexer assigns terms based on her/his knowledge of the subject area

Indexer must be generally familiar with the subject but need not be an expert

Index terms may not appear as text within the article; users must trust the indexer's subject expertise

Index terms should closely resemble terms actually used within the book, so users can quickly find information on the page and know they've come to the right place

Index terms are often part of a thesaurus of preferred terms, scope notes, cross-references to broader/narrower terms, etc., which is available to database users

Index must be its own thesaurus: scope of each entry should be self-evident; educational cross-references lead users from "commonly used" synonyms to index terms, and to narrower/broader terms

Index term pertains to all or a major part of the article; small, limited number of index terms per article => more generalized indexing

Index term may pertain to a single phrase; unlimited number of index terms per page (technical reference book averages 8-10 entries per page) => more specific, detailed indexing

Preferred term used in one form only; thesaurus gives cross-references to it from other forms of it; saves space in database

Preferred term is often multiple-listed, by rotating all pertinent words: helps users find information quickly, even if it uses more space

Unlimited number of articles per index term; articles are differentiated by titles or other phrases

If there are more than 7 locators for a main entry, it needs subentries to differentiate its contexts => the index language's phrasing depends on the amount of information on each subject

Users often use same database repeatedly => they become familiar with its index language

Users use many different book indexes, and may use this one infrequently => they often remain unfamiliar with its index language

Users are often subject experts, and/or searching experts => they become familiar with the index language

Books are often aimed at a general audience => users may be unfamiliar with this book's subject and index language

Many authors, and articles often aren't edited consistently => extremely varied terminology used between articles

Fewer authors (often only one), and more consistent editing => more consistent terminology used throughout the book

Databases are usually published by professional information clearing houses, with professional indexing staff: selling information access is their business, as users can't easily "browse through" the articles => more consistently professional index languages

Book publishers are out to sell a particular book, not information access; they may see an index as a "value-added" luxury, or unimportant, or even unneccessary, as users can just "flip through" a book => indexing quality ranges from pathetic to professional

© Lise Kreps 1998