Posted on by & filed under ebooks.

Mike Cane hates ebook indexes. He’s right — that kind of index is worse than none at all.

I happen to believe a good search feature is valuable and in some cases can replace an editorial index (for example, if the requested term is infrequent in the text). But if a publisher spent the money to create an index in the first place, it’s worth preserving.

I recommend three approaches based on time and cost, in descending order of preference:

Update all the indexes as direct hyperlinks to the term that was originally referenced by just a page number.

In this way the ebook is actually superior to its print counterpart, and is a nice value-add for the reader.

Since an ebook doesn’t really have page numbers, it’s best to re-write the index to eliminate them; one approach might be:

greyhounds – Chapter 1: [1, 2], Chapter 2: [1]

The numbers are incremental instances (“first appearance,” “second appearance”) rather than page numbers.

Another more interesting possibility would be presenting the index in search result format, with context:

In Chapter 1:
…certainly the fastest, if not also cutest, breed is the greyhound

This is radically different from traditional print indexes, but it’s more “digital-native” and better than even a directly-linked term, as the small amount of context can steer readers exactly where they want to go.

Target the original print page on which the term appeared, but obscure the page numbers as above.

Some conversion vendor workflows allow for marking the original page boundaries in the epub file itself, as hidden anchors. They can then automatically target hyperlinks to those pages, but not the actual term within the page. This is not fool-proof, as many times the user won’t see the term when they arrive at the target page — consider how much less text can fit on a smartphone screen. But savvy users will adapt quickly.

If you can’t hyperlink the terms at all, drop the index.

It provides no value and will only alienate readers.

Semantically, I’d probably recommend marking up indexes as an HTML definition list, but I’d be interested if there were other suggestions or even microformats that might be more appropriate.


2 Responses to “Improving indexes in ebooks”

  1. Dave Cramer

    Using the “incremental instances” rather than page numbers can lose information if the original index entry spanned a range of page numbers:

    Greyhounds: 10-16, 19, 26

    may be more helpful to readers than

    Greyhounds: Chapter 1[1, 2], Chapter 2[1].

    Every idea I’ve heard for eBook indexes seems designed for well-defined index terms like proper names. Indexing concepts and ideas makes it incredibly harder to automate the markup and links. Such an index may be closer in spirit to an incredibly-detailed table of contents, but sorted differently… I’d also encourage us to explore expanded tables of contents. Some books have lists of figures and maps, but how about a list of epigraphs and quotations, sorted by author? I can imagine all sorts of humanities books where that would be really useful.

    Dave Cramer

  2. Mr. Kramer

    I love the hyperlink idea! I haven’t gotten into ebooks yet, but it would definitely make it way more appealing to me!
    I don’t know how many times I’ve just skimmed over paragraphs in books because I don’t know what they are talking about.

    Great post!!