Make your own free website on Tripod.com
 home > indexing
How Deep Should an Index Be?
.
.
By Seth A. Maislin

Editor Note: This essay, originally published in 2000, was rewritten in March 2003.


  What is index depth? A good question. The depth of an index refers to how many entries and subentries there are, relative to the number of page numbers in the index or documentation. An index that has subentry after subentry after subentry is a deep index. An index that has only primary-level entries and page ranges (and thus little detail) is a “shallow” index.

  Index depth is subjective. The documentation by itself might be a very “shallow,” in which case index depth is not going to be high anyway. On the other hand, the documentation might be quite detailed, in which case the indexer has to decide how detailed (deep) the index should be.

  There is no ideal depth for the index, but there are certain rules of thumb that I follow in my own indexes. In each of these guidelines, it’s important to consider the book’s audience. Are these people looking for a quick reference or for some in-depth information? Are they highly educated on the subject matter, or are they novices who want decoration for their coffee tables? This knowledge needs to be provided in advance by the author, editor, or publisher.

  Below are four rules of thumb, with examples.

1. It’s almost impossible to provide too much access.

  If you’re even remotely concerned that your readers won’t know the terminology, you must do something about it. How many different ways are there to say “software”? Programs, applications, code, and so on. These aren’t synonyms, but they’re close enough for some readers. Then come the kinds of software, like utilities, games, and drivers. Software-related information could be made accessible through any of these terms. Choose your favorite term (in this case, software), and for the rest either use cross references or repeat the page numbers.

 
	applications. See also software
	code. See software
	programs. See software
	software, 10–20
  The above example is simple, but information can be more complicated. In the example below, the term drivers is important enough to merit additional access. Notice how the page number for 15 is repeated under the software entry.

 
	applications. See also software
	code. See software
	drivers, 15, 180
	hardware, 1-10
	   drivers for, 15, 180
	programs. See software
	software, 10–20
	   drivers, 15, 180
  There are two limits. One of them is absurdity: “zeros and ones. See software.” The other is caused by alphabetization, such that two entries are so close to each other in the list that one is enough:

 
	World Wide Web (WWW), 41–46
	WWW (World Wide Web), 41–46

2. Don’t use cross references for entries with one or two locators.

 
	drugs. See medicine
	medicine, 5
  With a construction like this, you’re wasting the reader’s time. Avoid the cross reference in favor of the more direct “drugs, 5.”

3. If there are enough page numbers to be annoying, use subentries.

  If a page range is too huge, readers are forced to go digging among those pages to find more specific content. Similarly, if you create long lists of undifferentiated page numbers, readers have to check each number for their desired information. Here are examples:

 
	coffee, history of, 133–178
	tea, medicinal effects of, 93, 134, 149, 206, 211, 217, 355
  Both of these call out for the use of subentries. For long ranges, create subentries you’re your subsections; the result might resemble your table of contents, although in alphabetical order. For long series of page numbers, use subentries to provide at least some differentiation. In both cases, some page number repetition may occur.

 
	coffee, history of, 133–178
	   agricultural research, 169-171
	   beans as currency, 140-149
	   caffeine as medical stimulant, 172–178
	   international trade, 150–157
	   smuggling, 157–169
	tea, medicinal effects of, 93
	   caffeinated tea, 134, 149, 355
	   decaffeinated tea, 149, 206, 211, 217
  It is important not to overdo this. Creating too many subentries is just as bad. In the example below, you’re forcing the reader to make all sorts of choices even though the information is clearly within pages 15, 30–33, or 101. The indexing effort required to create these subentries ends up putting the burden back on the reader.

 
	blood, drawing
	   disposal of needles, 32. See also biological waste
	   eating before, 30
	   equipment for, 30–33
	   fainting behavior, 33
	   fear of, 30-31
	   HIV risk, 32
	   from left arm, 30
	   phlebotomists, 15, 101
	   right- vs. left-arm drawing, 30
	   vampire myths, 15
	   volunteer donation, 15, 30
  Consider how much easier this is to read and understand:

 
	blood, drawing, 15, 30–33
	   phlebotomists, 15, 101
	   See also biological waste
  ven this variation below, which has placed greater importance on several concepts, is less burdensome.

 
	blood, drawing, 15, 30–33
	   disposal of needles, 32. See also biological waste
	   fainting behavior, 33
	   HIV risk, 32
	   phlebotomists, 15, 101

4. When page numbers are almost sequential, use ranges anyway.

 
	BAD:        macroeconomic precepts, 3, 5, 8-10, 12
	BETTER:     macroeconomic precepts, 3-12
	OPTIMUM:    macroeconomic precepts, 3-5, 8-12
  I know something feels wrong about doing this, but it demonstrates one kind of compromise between accuracy and usability. Trust that your readers know how to turn the page.

Copyright 2003 Seth A. Maislin

Top


HOME  |   ABOUT  |   INDEXING  |   WEBSMARTS  |   FUN & WACKY  |   EMAIL
Site design by little graphics studio.
© 2002   All rights reserved.