My First Embedded Index Job: WordEmbed
A client approached me about embedded indexes in manuscript Word documents. I replied that I had played with James Lamb’s WordEmbed macro but never used it for a job. They then asked me to embed an index using WordEmbed.
WordEmbed on a Macintosh
I work on a Macintosh. If you work on a PC then skip down to the next section, “Version of Word.”
There is a version of Macintosh Word that runs the WordEmbed macro, but I do not own it; I own Word:Mac 2008. So I installed Parallels on my Intel Macintosh; Parallels allows a Windows operating system to run as another window on my Mac. Parallels requires an Intel processor; to see if you have an Intel processor, look under the Apple menu, About This Mac, after the Processor label.
I needed an OEM version of Windows to install; an Upgrade version would have asked me to provide proof of ownership of another copy of Windows. I picked up a copy of OEM Windows XP on eBay for not much money. I also bought some copies of old versions of PC Word and/or Office on eBay for not much money. Buy from a reputable dealer to increase your odds of receiving a valid validation number. I haven’t been burned yet.
This setup works great: I have my Cindex window and my PC Word window side by side. WordEmbed throws its page reference on the clipboard, which I can then paste into my Mac version of Cindex.
Version of Word
I have played with WordEmbed under Word 2003 but this job had Word 2007 equations, so I used my Word 2010 version to maintain the integrity of the equations.
Remember: The index is embedded into live production files. Don’t mess up the files.
Preferences Export Setting
At the end of this embedding process, I’m going to export the index to an RTF file. WordEmbed requires a specific Preferences setting for exporting to RTF. Remember that this has to be set before exporting an embedded job AND it has to be turned off before exporting an index for a non-embedded job.
Under Cindex Preferences on the Export tab, the “Define indent by:” radio button needs to be clicked to Other with an equal sign = put into the little box. No spaces; just an equal sign.
And when you start your next non-embedded indexing job, remember to click the “Define indent by:” radio button to Style (most likely).
Multiple File Job
The book was supplied as 15 separate chapter files, plus an appendix and the front matter. My first task, therefore, was to give each file a unique page range. I needed to do this because the reference number used by WordEmbed to make each entry unique starts with the page number, followed by a period and the line number plus any needed additional numbers. When generating a full-book index, I needed each chapter’s reference numbers to start with unique page numbers.
It wasn’t enough to simply specify Roman versus Arabic numbers for the front matter page one versus the first chapter page one. I ended up giving the front matter the highest page range (just because the lower ones were already taken).
The actual page numbers do not matter; they won’t be used except for generating unique index entry reference numbers. The page numbering doesn’t even have to be sequential; the referenced document built at the end of the process determines chapter order in the book. In this job, I received a chapter late in the process and simply gave it a higher page range than any other chapter.
To set the starting page number, put the cursor at the beginning of the chapter file, use Insert, Page Number, Format to change the starting page number. Don’t insert an actual page number into the file.
The Process
WordEmbed is wonderfully easy to use. Select the text associated with the index entry, then hit the WordEmbed key combination Ctrl-Shift-\ (hold down the Ctrl and Shift keys then hit the backslash key). This turns the highlighted text into a comment consisting of a unique WordEmbed reference number. This same reference number is now on the clipboard. So create the new index entry in Cindex and paste the reference number into the page field.
You can use WordEmbed reference numbers as part of a range. The appendix to this job consisted of three long lists: prefixes, suffixes, and word roots. I started a new prefixes entry in Cindex then built my page range by placing my cursor at the beginning of prefixes and hitting Ctrl-Shift-\ then pasting that number into the page field, followed by a dash. Then I placed my cursor between prefixes and suffixes, hit Ctrl-Shift-\, and pasted that into the page field, finishing the prefixes range. Then I created a new suffixes entry, pasting the first reference number (the last reference number of prefixes, still on the clipboard), followed by a dash. Then I put my cursor between the suffixes and word roots, and so on.
The WordEmbed toolbar has a button for finding reference numbers, which are presented with their associated text highlighted. The visual aspect of WordEmbed makes it pleasant to use. I thought that the editing stage was faster because of it also.
I indexed all of the chapters into my Cindex file. When I finished, I exported my index to RTF. I copied that RTF file over to my PC directory so that my PC Word had access to it.
Then I backed up all those Word documents, just in case.
One by one, I opened the chapter files and clicked on the Embed Index button on the WordEmbed toolbar. Because I’m using Cindex and set the RTF export preferences to indicate indent with an equal sign, I needed to tell WordEmbed that I’m using the equal sign with each chapter file that I embedded.
The first chapter that has its index embedded has the cross-references written to its embedded index. All subsequent chapters then need to have the WordEmbed option to suppress cross-references check-marked. If this isn’t done then duplicate cross-references will appear in the final full-book index.
Troubleshooting
Out of the 17 files that I embedded indexes into, three did not work because “Word encountered a problem.” You might not get an error message; the embedding process might just hang.
I used Word’s Open and Repair on each of those files (File menu, Open, select the file, click on the dropdown arrow on the Open button and select Open and Repair, then save the file). One file did indeed have a problem, but that probably wasn’t its main issue in the embedding process.
I tried the usual corrupt Word document fix of selecting the entire document except for the final paragraph mark and copying into a new document. While this certainly didn’t hurt, I don’t think it helped in this situation.
I finally realized that the chapters were hanging at the same spot in the chapter file each time I tried to embed the index. I decided that Word didn’t like my selections. For example, one bad selection had a range that was partially within a table. In that case I reselected to grab the text on either side of the table, including the entire table within.
Then I replaced the problem page references in my Cindex index. I searched for both ends of a bad selection range within the index file so that the bad selection wasn’t referenced at all. Then I re-exported the RTF, re-copied it to my PC area, and was able to embed the indexes of the three problem files.
Work Rate
I found WordEmbed to be a very efficient tool. The book I indexed had some detailed and time-consuming chapters and also chapters that went quickly. Overall, I indexed 18 manuscript pages an hour. If this were a print job, I’d be pleased with that rate, but of course it’s Word manuscript pages. I’ll have to learn whether that’s a good rate.
Editing went quickly, if only because I didn’t do the read-over of page numbers that I usually do. The visual nature of WordEmbed selections made it easy to identify the text associated with locators.
My article Word Embedded Indexing gets into specifics for sorting and handling special characters.