Indexing
Indexing is a difficult and skilled task, best carried out by experienced indexers. Unfortunately, financial constraints usually mean that the task falls to you, the author. This page offers some tips and ideas for making good indexes with the minimum amount of work.
Indexing in advance
The obvious time to compile the index is when you receive page proofs, but of course this is when you are least likely to have the time. You can make life easier by doing some of the work ahead of time. How? By using your word processor’s indexing feature.
Before the proofs arrive, go through your word processor files adding indexing codes. This is a comparatively straightforward job with modern word processors. You should be aiming for roughly one index entry for every 150–200 words in a typical book. If in doubt, index a little more heavily – it’s much easier to trim an index than it is to add new entries later.
When the proofs arrive, reformat your word processor files so that each page starts and ends at exactly the same point as the pages in the proofs, with each page having the same page number as in the proofs. You might have to reduce the font size to get everything onto the page, but you don’t need to be concerned with the actual page layout for the purposes of this task.
Once the pages (and accompanying index codes) match the proofs, you can generate the index. All the page numbers will be added for you, the alphabetical sorting should be correct, sub-entries will be formatted properly, and you’ll be able to deliver the index along with the corrected proofs.
Indexing from the proofs
Attractive as the preceding option may sound, it may well be that you never quite get round to doing it before the proofs arrive. You can still use the same method once the proofs arrive of course, but there are also some other options.
Indexing with a word processor
Go through the proofs, marking candidate index entries with a highlighting pen. Once that’s done, type each entry into your word processor, something like this:
trees 23
grasses 23
savannah 24
trees, deciduous 24
antelope, grazing 24
trees 25
Ignore alphabetical order or any issues of layout. You can use commas, as here, to separate sub-entries from the main entry, or you could use some other character.
Once all the entries have been typed in, use your word processor to sort the entries alphabetically. You may have to experiment a little to get the right kind of sort, especially when it comes to getting the page numbers in the right order, so work on a copy of the file while you experiment. After sorting, your index file should be in alphabetical order:
antelope, grazing 24
grasses 23
savannah 24
trees 23
trees 25
trees, deciduous 24
Finally, you can format the index neatly, collating page numbers and indenting sub-entries:
antelope
grazing 24
grasses 23
savannah 24
trees 23, 25
deciduous 24
Use a single tab for each level of sub-entry.
You’ll almost certainly find some misspellings during this tidying up process. You might also find that, as with the ‘antelope’ entry above, there is only one entry in the entire book. In that case, the entry might become either:
antelope, grazing 24
or even simply:
antelope 24
Indexing with a spreadsheet
This is an unusual way to compile an index, but if you are a regular spreadsheet user you might it easier to work this way. The process is similar to indexing with a word processor, but with the difference that your entries are typed into spreadsheet cells:
|
A |
B |
C |
D |
E |
|
|
1 |
trees |
23 |
|||
|
2 |
grasses |
23 |
|||
|
3 |
savannah |
24 |
|||
|
4 |
trees |
deciduous |
24 |
||
|
5 |
antelope |
grazing |
24 |
||
|
6 |
trees |
25 |
Ensure that you set the format for the spreadsheet cells to 'Text' before you start typing the entries; otherwise the program might try to interpret page ranges as formulae or dates.
You can then sort the spreadsheet and export it to a word processor file for tidying up. Sorting the entries may be easier using this technique, compared with sorting them in a word processor.
Indexing from a PDF file
Acrobat PDF files are in common use in publishing. Many typesetters create PDF files from which they print the proofs. A consequence of this is that you might be able to obtain a PDF file from the typesetter to help when indexing – ask your publisher.
How does this help with indexing? Very simply, it does away with the highlighting pen. As you read through the text, you can use the Acrobat Reader’s ‘Find’ facility to search for every occurrence of a term, and then decide whether or not to index it (as well as exactly how to index it). As you work through the book, you can type the entries into your word processor, but the rough draft that you produce via this process requires much less work than either the word processor or spreadsheet techniques described above.
If you use this technique, be very careful to ensure that you don’t simply include every single occurrence of a term - it’s all too easy to let the computer override your judgement. Nor does this approach absolve you of the need to create sub-entries and cross-references - a task that cannot be automated by this method. See the next section (Indexing technique) for more on this topic.
If you don't have a copy of Acrobat Reader (a free program), you can obtain it from http://www.adobe.com/products/acrobat/readstep.html.
Indexing technique
Indexing is all about helping the reader to find something quickly and easily in your book. This section covers some of the basic rules of indexing.
Detail of index
A typical book should generate two or three index entries per printed page. It might seem like a good idea to have many more entries, but this leads to a couple of problems:
- The index might become too long, resulting in excessive printing costs.
- Entries might generate long strings of undifferentiated page references, which make it hard for the reader to locate specific information.
The second point needs some clarification. It is not helpful to index every occurrence of a term – if you do, you’ll get entries like this:
trees 2, 7, 23, 34–37, 39, 43, 47, 98, 102, 121, 124, 138, 166, 168
Your unfortunate readers will have to wade through these to find the information they are looking for. If this happens, either prune the number of page references (a maximum of about six is reasonable) or break the entry down into sub-entries.
Order of entries
In almost every case, alphabetical order is required. However, there are two variations of alphabetical order: letter by letter and word by word. An example should illustrate this.
|
Letter by letter |
Word by word |
Letter by letter sorting ignores spaces, hyphens and punctuation; word by word sorting looks at the first word of each entry, then the second word, and so on. This leads to differences, as shown above. Both forms are acceptable, but only one should be used in each index. Your publisher probably has a preference; letter by letter is most common.
In both forms of alphabetical order, common words such as ‘and’, ‘in’ and ‘for’ are ignored when sorting.
There are some exceptions to the use of alphabetical order, usually restricted to sub-entries. For example:
architecture
sixteenth century
seventeenth century
eighteenth century
nineteenth century
twentieth century
is preferable to:
architecture
eighteenth century
nineteenth century
seventeenth century
sixteenth century
twentieth century
Sub-entries
It is rarely necessary to go beyond two levels of sub-entry in a typical academic textbook. Often, just a single level suffices.
If an entry has a lot of sub-entries, consider carefully whether the entry itself should be split into two or main entries, or whether some of the sub-entries can be grouped together as a single sub-entry.
Nouns, not adjectives
Entries should be nouns, with adjectives used in sub-entries or following the noun. For example, ‘cars, fast’ is an appropriate entry; ‘fast cars’ is not (think where ‘slow cars’ is going to appear). If you ever come across an index where this rule has not been followed, you’ll find it very difficult to locate related index entries.
Cross-references
Cross-references (‘see’ and ‘see also’ are useful, but should not be over-used. Ensure that the target of the cross-reference actually exists and that it is spelt the same way as in the cross reference. Most of all, avoid circular cross-references: ‘cars see automobiles’ and ‘automobiles see cars’ is obviously unhelpful.
Abbreviations
If you use abbreviations in the text, use them in the index as well. Use cross-references to help people find their way from the unabbreviated term to the abbreviation or simply repeat the page references under both versions.
Irrelevant entries
Just because something is mentioned in the text does not mean that it automatically merits an index entry. There has to be some useful information in the text about a concept, object or person in order for it to be indexed.