Structuring your index

19 March 2010

An index is an index is an index, ne c’est pas? Well, actually, no. An index can come in all shapes and forms – though, as we shall see, in fact your options are reasonably limited.

However, what I shall briefly consider in this post is how many indexes you need, how many levels (or layers) are possible within the index, and the implications of these two factors on the internal structure of your index(es). Thinking through these issues must be done before you start creating any index entries.

Using the mind map

If you have indeed prepared a mind map of your study beforehand (discussed earlier), then its value will be immediately apparent at this point. The mind map is the structure of your index; its contours can be mapped into a skeleton index without a huge amount of additional work (though, as we shall see, if your mind map has a ‘deep’ structure then it will need to be reworked to accommodate a flatter one).

Indeed, all that you need to think about after this is the internal organization and consistency of your entries (plus the use of cross-references), the subject of my next post.

How many indexes?

A single, all-inclusive index may be normal but separate name and subject indexes are quite common. Indeed, in certain literary studies, for instance, it wouldn’t be unthinkable to have three indexes: the general index plus a separate index of authors and another of literary works.

That said, just because something isn’t unthinkable, that doesn’t mean that you need to do it. You will find it a far less complex task to create a single index. I suspect your readers will also thank you for it, too. Ultimately, however, if you are wanting to do anything out of the ordinary, it would be a good idea to talk to your production editor first; there is no point spending a whole lot of time and effort producing an index if it is then rejected.

How many levels?

Likewise, there is the issue of levels within your index(es) to consider. Theoretically, you could have as many levels as you want but in fact you are limited by the layout of the index. There are two factors here:

  • Page layout. Most indexes are laid out in two columns per page (I have seen indexes squeezed into three columns but they look awful and are hard to read). The width of each column will be no more than 6 cm (less than 2.5 inches) wide.
  • Arrangement of sub-entries. The convention with indexes is to indent each level of sub-entry. An index with three levels thus has a huge amount of indenting (and wasted space) to fit into such a narrow column.

In reality, then, the physical constraints of the index limits its depth to two levels. Such a flat structure is also much easier to work with and the resulting index is also easier to read.

Internal structure

Adoption of such a two-level structure has huge implications for the internal structure of your index, especially if you are working from a mind map (as undoubtedly, it will have entities with more than two layers of complexity). Flattening this structure will require:

  • creation of new related entries out of the old entry
  • judicious use of cross-references

Perhaps the best way to illustrate this point is by taking a portion of the mind map that I used earlier and translating this into a skeleton index.

Mapped as:

xxxxacademic standard
xxxxsee also Cultural Issues; Language
xxxxcorrect usage/spelling
xxxxplain English
xxxxtools. See Language tools
xxxxsee also English; Presentation; Style
Language tools
xxxxreference works
xxxxspell checkers
xxxxsee also Language

Note here that the mind map is not perfect (hence the index created is not a faithful reproduction) but it will have to do. Hopefully, you will get an idea of what I mean by this as a result.

At this point, you have all of the information needed to index your book. My next post shall be looking at the formatting of the actual index entries.

(Post #9 of the Indexing section of a lengthy series on the book production process, the first post of which is here.)

What to index

16 March 2010

Apart from the number of pages available, there is another important factor determining the length of an index: just what is to be indexed. This is not an issue considered in advance by most authors, I suspect, but your choices here will add to readers’ perceptions about the index and your book.

Does it look as if the indexing job has been thoroughly done – but not overblown – or is this a skimpy (even sloppy) affair? Although a threadbare index useless to its purpose will immediately signify that the rest of your book is equally inadequate, more is not necessarily better. There are limits.

So what material should be indexed in your book?

Obviously, your body text must be indexed, but what about notes that comment on the text? Citations and bibliography are usually not indexed; will you follow that practice? Some publishers think it unnecessary to index a glossary (indeed, sometimes an index becomes a kind of glossary), but consider it essential to index any illustrations and captions. What does your publisher say?

With luck, your publisher will have guidelines on all of these matters but otherwise simply use your common sense.

As you can see, decisions need to be made here but they are not life and death issues, nothing even to lose sleep over. Indeed, although what you chose to index can be significant, the ultimate quality of your index depends on other factors, not least the intellectual conceptualization of the index (or mind map, discussed here) and its ‘visualization’ in the index structure, the subject of my next post.

(Post #8 of the Indexing section of a lengthy series on the book production process, the first post of which is here.)

Index length

12 March 2010

Winter this year in Copenhagen has been cold and snowy (and I’m sick and tired of shovelling all that white stuff; the searing heat of Bangkok tomorrow should be a shock). Overcoats are seen here far more than is usual, some looking more comfortable than others. An index is like an overcoat: too small and its squeezes its wearer in a wrestler’s clamp, too big and it lets the cold air seep in underneath.

Knowing beforehand

A year or so ago, we hired someone to index one of our books. He was very good, very thorough, and also flagged up a number of typos and inconsistencies in the text that had been missed. There was a problem, however. The index delivered was 64 pages in length (about 23 typeset pages), the maximum amount of space available 12 pages. Drastic cuts were made.

A key lesson we learnt from this experience was to be (even more) clear about how long we expect an index to be. In your case, your production editor should know how many book pages are available for the index. Make sure that you are informed about this before you start indexing.


Is there any real limit to how long your index should be? Yes. In fact, there are several factors at play here.

  • Perception. Indexes are like books: if too short, they may be treated with disdain (a 2-page index may be seen as pathetic and laughable) but, if too long, they may be regarded as unwieldy and ‘over the top’.
  • What is physically possible. Books tend to be sized in multiples of 16 pages because that is how they are printed (on big sheets of paper holding 8 book pages on each side; more about book length here). If (say) the total extent of your final proofs is 276 pages including prelims, then no more than 12 pages will be available for the index (or 28 pages if an extra 16-page signature is to be used).
  • Flow-on effects. In the above example, no publisher will agree to an index that when typeset fills 13 pages and causes there to be 15 blank pages at the end of the book.
  • What has been announced. A 12-page index is quite reasonable, even generous. In the above example, it would be unlikely that an index longer than this would be permitted if the book has been announced as being 288 pages in length.

Calculating size

Of course, there is a confusion and potential trap for you when I talk about ‘pages’ above. The number of typeset index pages is not the same as the number of ordinary (A4 or US Letter) pages on which you prepare your index in Word or another word processor.

Typeset indexes are generally set quite tight in a smaller font size and laid out in at least two columns whereas more than likely your index will be prepared in 12-point Times Roman with generous line spacing. As such, your typesetter should be able to fit a bit more than two of your word-processed pages onto one typeset page in the book.

As a rule of thumb, then, if told that you have (say) 8 pages available for the index, then you can count on having 17–18 pages in Word for your index (though obviously not single-spaced in a tiny font size).

If the amount of space is too little or (in some ways, worse) too much, then there may be an issue of what (how much) you are indexing, the subject of my next post. If so, re-evaluate your situation and – if not resolved – contact your production editor urgently.

(Post #7 of the Indexing section of a lengthy series on the book production process, the first post of which is here.)

Indexing methods

11 March 2010

As intimated in my previous post, there are different ways of preparing an index. I have identified four main methods, none of them ideal. These are the:

  • traditional method,
  • mapping method,
  • mark-up method, and
  • quick and dirty method.

Traditional method

The traditional method is that you prepare a manual index at the same time that you proof-read the text. There are several ways of doing this, by:

  • recording the entries on the proofs (highlighting text and/or making notes in the margins),
  • writing them down on index cards or several sheets of paper, or
  • keying them immediately into a text document.

This method is simple. All that you need are the final proofs and a means to record the entries. The last way has the virtue of being an all-in-one method whereas, with the two earlier ways, there is a second step – to transcribe the entries – but they are visually much easier to work with.

If you have prepared the coloured mind map described in my previous post, then you can quickly ‘transcribe’ its text highlights over to the proofs and thus save a bit of time.

Arguably, the traditional method of indexing gives the best results because it allows a really thorough job to be done; everything is there before your eyes. It is, however, a painstaking (read: painfully slow) approach best suited to tortoise personalities.

Mapping method

A modern variation on the traditional method is to prepare a skeleton of the index (minus page numbers) beforehand and then fill in the page numbers by searching on a single PDF file of the book. (Obviously, this method requires that you receive the proofs as a PDF, not just in hard copy.) Acrobat’s search functions are very useful here and of course the text you are working with is the real paginated book (I find that psychologically useful).

This method can be done directly on computer (switching back and forth between Acrobat and your text file) but visually it is much easier to have the index skeleton printed out on paper (ideally with double line spacing), adding the page numbers to this and later transcribing them over to the text file.

As you can see, this uses the mind map to its fullest potential. The Acrobat search also allows you to find references that otherwise you might have overlooked. That said, this approach requires sophisticated and careful searching. For instance, in a study of nations and nationalism, if you only search for ‘nations’ then you won’t find ‘nation’ or ‘nationalism’ but doing a search on ‘nation’ might give you too many results to deal with. Better first to do a search on ‘nationali’ (to pick up ‘nationalism’, ‘nationalist’ ‘nationalistic’, etc.) then search for text often related to this term (like ‘ethno’) to find other entries.

In short, this can be a quick indexing method with the potential to give poor results but if used properly is a very powerful and fast tool. For this reason, it is the indexing method that I personally favour – also because its combination of broad-brush and nitty-gritty approach fits my temperament. This is the method best suited to impatient perfectionists.

Mark-up method

The mark-up method involves entering indexing tags in the book file itself.  It requires that you have a single text file generated from the typeset proofs (or, if you are brave, from the edited files delivered to the typesetter). Saved as a MS Word file, this must then be paginated to match the typeset proofs by playing with the font size and/or inserting hard page breaks. (The file need not be pretty; it simply needs to have the page breaks – every single one of them – at the same place as in the typeset proofs.)

The marking up process can be as slow a task as the traditional method (thought, again, much faster if you have prepared a coloured mind map in advance). However, when completed, the resulting index is instantly generated and with luck should not need a lot of adjustment. Index generation can even be re-run repeatedly in conjunction with adjusting the tagged entries until the index is perfect.

This indexing method has the virtue that you can regenerate the index as many times as you want until the results are perfect. Moreover, it is the best method to use if suddenly the pagination of your book is to be changed; this ‘merely’ needs to be reflected in the pagination of your base text file. Unfortunately, adjustments to index entries can be tedious. For instance, to divide a large number of single-level entries into groups of two-level entries requires that every single entry is manually updated to the new format. This is not something you want to do too much of. As such, this is the method best suited to organized, methodical people who already have the structure of the index perfectly clear in their mind beforehand (i.e. they have done some sort of mind map).

Quick and dirty method

My last method I call ‘quick and dirty’ but in reality it is not quick; perhaps ‘fast road to hell’ would be a better title. The version I know is done in MS Word but I know you can do the same in WordPerfect; quite possibly other word processors have the same feature.

Again, this requires that you have a single text file as your reference source (see above). What you then do is create a concordance file (a list of words to be indexed) then let Word automatically generate the index from your book file. Though quick to create, this is not something I’d recommend; the resulting ‘index’ will be full of junk entries that you can spend days (even weeks) weeding out and it may lack entries that later you realise are necessary. In the end, then, this method may save no time at all. Bluntly put, only serial losers would use this method a second time.

And there’s more

No doubt there are other ways of indexing a book but these four are the most common. Of course, an issue only touched on here is how your index is structured; this needs to be settled before you can start using one of these methods. Likewise, you must decide on exactly what you are going to index. These (and more) are topics for upcoming posts in this thread on indexing. (Just when I post these is uncertain as I’ll be travelling these next three weeks.)

(Post #6 of the Indexing section of a lengthy series on the book production process, the first post of which is here. This is a complete rewrite of an earlier post on the same subject.)

Mapping your study

10 March 2010

As I’ve said before on several occasions, an index is a mind map, the ‘visualization’ of your study as an alphabetical list. This map is implicit in the index but you can dramatically improve the coherence, balance and completeness of your index and the actual text of your book itself by making this map explicit.

In the act of plotting and drawing this map, often you will discover both repetitions and gaps in your text. Although you cannot at this stage record any page references against an index entry, nonetheless you will quickly see which entries are common (even over-represented) in your text and which are scanty or missing.

If these defects are found at the writing or even editing stages, there is little fuss in correcting the situation. Not so by the time of the final proofs, when it is virtually impossible to make such changes without the typesetter slapping a hefty fine on your publisher (a cost more than likely promptly passed onto you, as we have seen).

How you go about creating this mind map is a variant of the first and second indexing methods described in my next post, i.e. involving that you:

  1. Read through your text, highlighting words/phrases/paragraphs you wish to index and occasionally scribbling notes in the margin.
  2. Collate these colour dabs into an index skeleton (i.e. without any page references).
  3. Analyse this index skeleton in terms of coherence but also content, looking for repetitions and gaps in the entries.
  4. Rework the skeleton until it equates to a satisfying mind map of your book.
  5. Search your text again to see if any of the gaps in your index skeleton are truly missing or simply were overlooked in the initial highlighting of the text. Eventually, you will have reduced the problem to a core of gaps (and repetitions) in the entries.
  6. Analyse and adjust your text to deal with these gaps and repetitions.

Thereafter, depending on which indexing method you use and provided there are not too many resulting changes to your text, you could use this highlighted version of your study or the mind map to speed up the final indexing process (more about this in my next post).

(Post #5 of the Indexing section of a lengthy series on the book production process, the first post of which is here.)

When to start on the index?

9 March 2010

The typical index is done at the time of the second (final) proofs and usually under a great time pressure. Not surprisingly, the indexes produced by first-time authors (and indeed many more experienced authors) are not always of a high quality.

Your index could be better.

Avoiding the last-minute rush

Let’s be honest, producing a great index is much more than a matter of timing. As important (if not more so) is that you successfully handle the issues raised in my previous post – issues like structure and format that my following posts will discuss in greater detail. Nonetheless, timing is important; indexing need not be a last-minute affair.

After all, what is an index? Superficially, it is an alphabetical list of cross-references to material in your text, each entry comprising a term and a page reference. (On a more sophisticated level, as described in an earlier post and in much greater detail in my next, an index is also a mind map.) So what is holding up your index, making its finalization a last-minute affair? Not the list of terms; theoretically, this could be devised before you even began writing your book. No, what’s holding everything up is having the correct page references – and there’s nothing you can do about that until your book is in its final, paginated form.

The secret to having a great index, then, is to ignore the lack of page numbers for now and to focus on all the other, more important issues. This means that you can start imagining the index much earlier than at the proofing stage, start even while writing your book.

Alternative approaches

There is a big advantage here if you do. It gives you time to create a mind map of your study and its index, if you so choose (more about mind maps in my next post).

Of course, you don’t have to create such a mind map. If you feel confident that the text sent for typesetting will change very little during the typesetting, you can instead simply start indexing your book straight after editing. To do so, all you need to do is create a single book file from the text in MS Word (or whichever word-processing program you use) and begin entering index tags. (This is indexing method 3, which I describe in much greater detail in my next post but one.)

That’s not all

Whichever approach you follow, you’ll need to bear in mind how much space you’ve got to play with, how the index is to be structured, etc., etc. – issues described in subsequent posts.

In the meantime, with a bit of breathing space before the index must be delivered, you actually may have time to enjoy its creation, time to unlock the intellectual puzzle of your study.

(Post #4 of the Indexing section of a lengthy series on the book production process, the first post of which is here. This is a complete reworking of an earlier post on the same subject.)

Who should do the indexing?

8 March 2010

You. The big finger is pointing directly at you. Just as probably it was you who had to key the copy-editing changes, you who maybe got to help design the cover, and you who definitely had to do the proofing, so too is it you who’s now expected to index your book – and in double-quick time.


This need not be so. No one is forcing you, personally, to do the indexing. After all, this is skilled work and you may not feel up to the task.

You could instead hire Anthony, an indexer we’ve used on several occasions when the author was unwilling and had the cash to hire a professional. Anthony is reasonably priced (surely he cannot live off these earnings) and not only does he turn out good indexes but also – in effect and free of charge – he gives the book another proofing; tacked onto his indexes is a page or two of comments about errors and discrepancies that he’s found in the text. In short, a professional indexer like Anthony could be just what you need.

But maybe not.

People like Anthony cost money, you cannot be sure you are hiring a good indexer, and they may not be available when you need them. Moreover, an outside indexer has no hope of ever knowing your book as intimately as you do. And, if it is you to do the index, then you can make an early start and refine the index as editing and typesetting progress. It has the added advantage that you can work with the mind map described in my previous post.

Definitely, this issue is something to think about carefully and to fully investigate in good time.

Hiring a professional indexer

If indeed you engage a professional indexer directly, then book a time slot early, and keep your indexer informed as the actual start date firms up. Most important, prepare a clear indexing brief that specifies what you want – issues such as these that will be covered in my following posts:

  • When will you deliver the proofs for indexing?
  • What is the indexer’s deadline for finishing the job?
  • What is the agreed price?
  • How long should the index be?
  • What should be indexed?
  • Is it only one index required or several?
  • How many levels should it have?
  • How should the entries be formatted?
  • Are there any special considerations to note?

Doing the job yourself

If, however, you decide to do the job yourself, then prepare a clear indexing brief for yourself, too. You’ll also benefit from taking note of the other issues and advice found in my following posts.

And whatever else you do, do not skimp on the job. A poor index signals to the reader that this is an inferior book. Do not fail your book at this last hurdle, mere days before it goes to the printer.

(Post #3 of the Indexing section of a lengthy series on the book production process, the first post of which is here.)