An index is an ‘ordered arrangement of entries ... designed to enable users to locate information in a document or specific documents in a collection’ (ISO 999, 1996). A document may be a book, an issue or volume of a magazine or journal, audiotape, film, computer file, or any other information source.
Some indexers work from page proofs and like to mark the text (with a highlighter or by underlining) as they read through, before making entries; others read right through first but don’t mark as they go. Many indexers no longer work form printed page proofs, preferring instead to read pdf proofs on-screen, often using a separate monitor. Indexers use specialized software, which helps to automate the more mechanical aspects such as sorting index entries, enabling them to concentrate on the intellectual task of anlysing the text.
Yes, although for some simple texts you may not need to read right through before starting to index. It’s preferable to read a book once straight through to get an idea of the topics, then again when indexing. Production schedules rarely allow enough time to read a book thoroughly before starting to index.
It's essential to understand the subject matter, so that you are familiar with the terminology and concepts and more aware of the nuances and relationships between topics. If you have a wide general knowledge you may be able to index books on a variety of subjects aimed at the genral reader, but specialist books require specialist indexers.
Indexing is not merely a listing of words but a creative process interlinking concepts and constructing a logical hierarchy, sometimes with many levels. It involves considerable and sustained mental effort, so it is certainly not mechanical and therefore not boring.
Generally you can pick out important words or phrases used in the text on a first reading. Part of the indexer’s skill is to recognize general topics and find appropriate wording that may not be obvious from the text. Minor topics can be included in more general ones with a ‘see’ reference to the indexed term. This can be done as the index progresses, or a fuller index can be condensed in this way at a later stage.
There shouldn’t be. The entries should be broken down into subentries so that different aspects of the topic within an entry can be selected, thus giving more targeted information.
It can be. The difficulty lies in the different densities of text, inconsistencies of terminology, and repetition or overlapping coverage. One chapter may be very factual and yield lots of indexing terms, whereas the next could be full of theories and concepts. Firm briefing at the commissioning stage and careful editing can minimize the use of different language for the same concepts, and different or repetitive treatments of a subject, but there may still be problems for the indexer.
If the subject is suitable, a thesaurus (a set list of terms) can be compiled and consulted as work progresses. By this means it is also possible to compile a cumulated index where terms are consistent over a long period (say 5 or 10 years) or over a large work (say 25 volumes). Many indexers may be involved in such a large project, and detailed guidelines are necessary to enable them to work consistently.
Yes. Illustrations such as photographs, diagrams and tables should all be indexed because it is useful for the reader to know that something is pictured or listed in a table, and sometimes the illustration or table includes extra information that does not appear in the text.
Even quite young children can use an index to find their way around a book, and should be encouraged and helped to do so. But the index should be short and set out clearly and simply. For more information see Indexing Children’s Books and the article by Valerie A. Elliston (reproduced from the Children’s Writers and Artists Yearbook).
Computer software can easily produce a list of words used in the text (a concordance), but this is only a word list with no structure: it does not include concepts implied in the text, nor does it link related terms.
Word processors or page layout software can be used to produce an index, but not automatically. The indexer (who in this case may be the author or editor of the document) has to insert suitable index entries at the appropriate points in the file and then generate the index from these when the document is complete. This process is known as embedded indexing.
An index can use cross-references to alert the user to wider or allied concepts, whereas a computer search can only find designated words or phrases and can’t identify text that may convey the same meaning but uses different words. The searcher has no way of knowing what has been missed in this way. Another drawback is that a computer search cannot single out the most important references to a topic, or ignore trivial passing mentions. For more detailed discussion of this topic, see the article on Human or computer produced indexes?