Well, I've finished (more-or-less) compiling an index for a book. This is my first book and my first index. I've learned a few things, mostly that it's a long, tedious process. But there are some other things that I'd like to remember for next time.
First, I should say that my book is the published form of my dissertation completed at Hebrew Union College under Adam Kamesar and Richard S. Sarason. If everything goes as planned, it will appear in May within the series Vigiliae Christianae Supplements (Brill). See here for a preview.
I compiled two indices--an index of ancient sources and a general index (subjects and modern authors). I did these in reverse order.
For the general index, I used the instructions provided by David Instone-Brewer at Tyndale House. It was supposed to be an easier way to index (I suppose it was), as the computer did some of the work for you, but it was still a bit more complicated and time-consuming than I anticipated. I'm sure this is partly because I was inexperienced, and partly because I am far from being a computer whiz.
You first have to make a Word document of the PDF that the publisher sends you (the proofs), and then get all the page numbers in the Word doc to be right, and then use a program to make a concordance of the Word doc, and then delete all the irrelevant words (words you don't want indexed--this took a long time), and then clean up the resulting word list, and then have Word provide the page numbers.
But, after all that, I still had to go through the index and eliminate the page numbers that weren't right. This is because sometimes people have the same last night, and Word provides the page numbers for both guys, so you have to eliminate the page numbers that don't apply. Or, when I look up Scholar X in the index, and I don't want to find references to articles written by someone else that happen to appear in books edited by Scholar X, so I eliminated those page numbers, too. And then also, for the subjects, I didn't want to include references for pages that included the term but were irrelevant. For example, I mention Homer a few times in my book, and those pages are referenced under the key word 'Homer'. But I also cite a book or two with Homer in the title, and I cite that book in contexts having nothing to do with Homer (like citing Honigman's Septuagint and Homeric Scholarship while talking about the Letter of Aristeas). I didn't want to include those references under the key word 'Homer'. So I found that I still had to go through the index and check every single reference to make sure it was one I wanted to leave in there.
If I had known at the beginning what I know now, I would have skipped some steps. I suppose I still would have the concordance program produce a concordance for me, and then I'd go through and delete all the many (thousands, it seemed) words that shouldn't be included. But, since I would go through and check all the page numbers, anyway, there's really no reason to have Word automatically provide the page references, and therefore there's no reason to make a Word doc based on the page proofs in PDF supplied by the publisher, and then ensure that all the page numbers are right in my newly-created Word doc. Next time I think I'll just make my concordance from the Word doc I sent to the publisher originally (from which the proofs were made), and then supply the page numbers myself by searching the PDF proof pages. This sounds incredibly tedious, but I think it will actually save me some time.
There was also an unforeseen problem with using a Word doc made from the PDF proofs supplied by the publisher. In the transfer between formats, for some reason, many (but not all) of the 'f's' did not come over. So, in the newly-created Word doc, the name 'Rufinus' appeared as 'Ru inus', and 'Niehoff' became 'Nieho '. This obviously created problems for the concordance program, which produced no occurrences of Rufinus but instead had 'Ru' and 'inus'. Making the concordance from my original Word doc would avoid this problem.
For the index of ancient sources, I knew of no better way than to simply go through the document and find who I cited, and then run a search in the PDF for all the places that I cited that author or work. Instone-Brewer does provide instructions for an 'automatic' way of doing this index, but the directions were too complicated for me.