As a Society of Indexers trained indexer I work under the tenets of BS ISO 999:1996 Information and documentation — Guidelines for the content, organisation and presentation of indexes. This states, among other things, that the function of an index is to provide the user with an efficient means of tracing information. The indexer should therefore:
a) identify and locate relevant information within the material being indexed;
b) discriminate between information on a subject and passing mention of a subject;
c) exclude passing mention of subjects that offer nothing significant to the potential user.

Passing mentions are an item or concept mentioned incidentally in the text but lacking worthwhile information about the item or concept itself. Mere mentions of the existence of something that does not provide at least one fact should be avoided in an index. Generally I try not include locators where no substantial information is provided.

Passing mentions typically fall into four main types:

• examples (Many marsupials, including possums and bilbies, are nocturnal – entry would be marsupials or nocturnal animals, not possums or bilbies),
• lists of things or people (the group subject is the entry if it needs one)
• asides (as my predecessor, Dr Jones, might have done – no entry for Jones)
• scene setting may include passing mentions that are not followed up with what follows.

In the age of Big Data and ctrl+f searches it can seem quaint and outdated in that as an indexer I have to try and assume the role of reader and what they might be looking for and make judgements. But that’s the advantage of having a human being do the job, not a computer. However, I also have to ensure the terms used are appropriate and will

• quickly establish the presence or absence of information on a specific subject in an unfamiliar work [the new reader or browser],
• quickly retrieve information on a remembered item in a known or partially known work [someone who has already read all or part].

The first will let someone know if the book is worth reading, the second helps them find things when they return to it.

So I try to find relevant information, concentrating on the ‘aboutness’ of each section, and generate ways of readers finding material they might want. Indexing is an art, and as such every indexer will without doubt produce a slightly different take on any particular text, they might choose different ways of saying things, they might select different pages. But each would produce a usable index.

However, there are occasions when including every mention might be appropriate, such as family or local histories that include lots and lots of names, and deciding when it is appropriate is what sets indexers apart from mere machines.