June 11, 2009


Posted in books, Design, indexing, production, Tools, Typesetting at 4:45 am by India

talk to your kids about indexing

Via e-mail, Lars R. asks, “Would you consider doing a write-up on your blog on the production of indices and how indexing relates to the design process as a whole?”

Some topics I’m interested in include

– The usefulness of InDesign’s indexing feature (as opposed to third party programmes if they exist, or simply manually typing in numbers)
– The practicalities of the designer being involved with the nitty gritty versus any sort of indexing specialist working independently)
– At which stage in the production process indexing begins and ends
– Differences between independent/inhouse publishers and large commercial affairs
– Does the designer generally have any input to level of detail, extent etc, or is it exclusively a case of matter having priority over form? How does the index influence castoff?

Thanks, Lars! I love trying to answer specific questions. Here’s what I’ve come up with.

On “The usefulness of InDesign’s indexing feature (as opposed to third party programmes if they exist, or simply manually typing in numbers)”

Most of what I know about the first question is contained in the recent post Big Is Beautiful and the comment that follows it. To recap, I’m not an indexer, so I can’t speak from experience, but my impression is that most professional indexers prefer to either type indexes up the old-fashioned way or use tools that are dedicated to the task, rather than employing the indexing modules that are built into page layout programs. Carol Roberts kindly corroborated this in her comment on that post:

There are a few indexers who will work in layout programs, placing tags. . . . indexing that way means the indexer doesn’t necessarily have to wait for that final set of page proofs before starting the index. The drawback, . . . is that indexing features of layout programs (or of Word, for that matter) are not very powerful or versatile. Working that way is painfully limiting, I’m afraid. Most of us professional indexers use one of 3 stand-alone indexing programs: Cindex, Macrex, or SKY.

Furthermore, it’s unlikely that many indexers have experience working with the layout programs’ tools at all, since an indexer usually receives a PDF or hard copy of the book, not an application file.

On “The practicalities of the designer being involved with the nitty gritty versus any sort of indexing specialist working independently”

As I said in Big Is Beautiful, “Indexing is a highly specialized skill—a great indexer is born, not made, I suspect—and most indexing is done by professionals.” I’ve written two indexes, but that was while I was making my living mostly through editorial work, not design; and writing two indexes was enough to convince me that it’s not something I’m good at. It would be extremely unusual for a designer to have anything to do with indexing a book, even at an independent press. When a publisher is too cheap to hire a professional indexer, it’s usually the author who gets saddled with the task; and if that isn’t appropriate, I’m sure an editor or even an intern would get stuck with it long before anyone would dream of asking the designer to do it.

What I have done a lot of, as a designing-and-typesetting managing editor at a small press, is editing indexes. I did this sometimes to make them more coherent but more often to make them fit. It’s very handy, when you’re required to cram an index into a space that’s too small, to have the liberty to change the format of or otherwise condense entries. You stand a much better chance of achieving a good balance between readability and comprehensiveness when you can jiggle both variables at once than when you have control over the typography or the content alone. Such flexibility is rare, though.

“At which stage in the production process indexing begins and ends”

The index is typically the last part of a book to be written and typeset. According to the Chicago Manual, “Most book indexes have to be made between the time page proof is issued and the time it is returned to the typesetter—usually about four weeks” (18.3). I believe when I was commissioning indexes for Scrappy Independent Publisher X, the turnaround time was more like two weeks. And as a typesetter at Book Packager Y, I’d usually have to set the index within one day of receipt. That task would jump the schedule ahead of just about anything else, because a book whose index is ready to be set is, almost by definition, a book that’s due at the printer tomorrow.

On “Differences between independent/inhouse publishers and large commercial affairs”

I’ve worked at only one large commercial affair, and as far as I can recall, none of the books that crossed my desk there contained indexes. Had an index been required, however, I can guarantee that there would have been several more days of padding in the schedule to accommodate it than we would have had at a small press. For one thing, the index would have to be sent out to a big commercial typesetter, where rush work costs extra. For another, the pages would have to pass through several people’s hands for approval.

At Scrappy Independent Publisher X, I’d receive an index via e-mail, give it a quick skim, shoehorn it into the back of the book, and send files to the printer that same afternoon. Nobody besides me ever so much as glanced at it.

When I worked at Book Packager Y, I’m sure our clients usually set aside a few days for editing and setting the index, but they knew that in a pinch we could probably turn it around on the same day. It depended on who was available, though—indexes were treated as somewhat magical and mysterious in that shop, and not everyone was allowed to set them. I’m not sure why; it can be fiddly work, to be sure, but it’s hardly rocket science. Still, I enjoy setting indexes, so I didn’t mind getting stuck with most of them.

“Does the designer generally have any input to level of detail, extent etc, or is it exclusively a case of matter having priority over form? How does the index influence castoff?”

Short answer: No; index pages are figured into castoff calculations as an indigestible lump.

Long answer:

The density of an index is an editorial concern, and, as Nancy C. Mulvany explains in Indexing Books (Chicago: University of Chicago Press, 1994), it’s helpful to express it as a percentage of the book’s total length.

Assume that a book with 200 indexable pages has 10 extra pages reserved for the index. If we divide 10 by 200, we end up with 5 percent:

10 ÷ 200 = 0.05, or 5%.

We can refer to this as a 5 percent index. As a general rule of thumb, a 5 percent index is not an exceptionally dense index. To an indexer, this means that a moderate amount of indexing is expected, perhaps averaging five to six entries per indexable page. . . .

The ratio of indexable text pages to index pages varies for different types of book. . . . When an indexer is told that 6 percent (or more) of the indexable text pages have been reserved for the index, it is reasonable to assume that a fairly detailed index is desired. When the percentage drops below 5 percent, it is reasonable to assume that light indexing is desired.

In the case of technical manuals, a 10 percent index is not uncommon. Some complex manuals have 20 to 25 percent indexes. . . .

At the other end of the spectrum, in general trade books, we often find 2 percent or 3 percent indexes. (64–65)

When I designed for Large Commercial Affair Z, I would receive from the production editor a castoff worksheet that specified the number of pages to set aside for an index, if one was required. I would then design around this number, as if it were an indigestible lump—I could not make it smaller, nor could I add more than a handful of blank pages around it. I can’t say how I’d spec the index design for another typesetter, since I never needed to do so. I guess I’d have asked someone else for advice . . . and probably received a blank look. Whenever I asked what the design conventions were at Large Commercial Affair Z, I’d get the answer that there were no conventions.

At Scrappy Independent Publisher X and Book Packager Y, since I knew I’d be setting the indexes myself, I could be more vague about length. I was almost never told that I had to hit a more specific page count than as short as possible and as long as necessary, and I designed using the live text, so I’d usually make the book the length it wanted to be and then tack on however many pages would both accommodate a reasonable index and round out the last signature.

At Scrappy Independent Publisher X, as managing editor I would commission the index based on the type of book and the space available—usually about a 4 percent index, I believe. Mulvaney’s book has a handy “Index Specifications Worksheet” among its appendices that loosely inspired the cover letter I sent to the indexers I worked with. My letter was not nearly as detailed, but I’m pretty sure I stated in it my preferences for alphabetization, punctuation, cross references, and page numbers.

When the index came in, I’d design it to fit—add or remove letter headings (A, B, C, etc.); fiddle with the text size, leading, and indents; run in entries if I needed to shave a lot of space. At Book Packager Y, I would have to ask the client (in the form of the production editor) if it was okay to run the index in; the default was for entries to be set indented. At Scrappy Independent Publisher X, I was the production editor, so I had the authority to do whatever I had to to make the index fit, including rewording entries or cutting a few lines. A good index can make the difference between a book’s being adopted for university courses or not, though, so my goal was always to make as complete an index as possible, as legible as possible.


Most of my experience in dealing with indexes is, I believe, not typical of U.S. publishing, so it would be helpful if readers who’ve designed or indexed in more traditional contexts would weigh in. The gist everywhere, however, is, I’m pretty sure, that an index is usually generated as a low-tech document, and the designer usually has very little control over its formatting besides fiddling with the typography.

This will change, I suppose, as more publishers start practicing platform agnosticism. When you’re issuing not only a print edition but also an array of e-formats, suddenly it becomes important to have an index that’s electronically anchored to words in the file rather than keyed to page numbers. My guess is that this work of anchoring will fall into the lap of the designer, rather than into that of the indexer or the production editor, simply because designers tend to be the biggest computer geeks in the production food chain. However, maybe if software like InCopy starts supporting the creation of index entries (it’s astonishing, really, that it doesn’t), that current division of labor will be maintained.


Photo: If you won’t talk to your kids about indexing, who will? by broken thoughts / Mark Lindner; some rights reserved.


  1. Lars said,

    India, thanks for humouring me :-) I look forward to any comments!

  2. Monica said,

    This blog is so unbelievably informative, I feel I should pay you. Thanks so much, India.

  3. India said,

    What a nice thing to say! Thank you, Monica.

Comments are closed.

%d bloggers like this: