The following blog post will probably bore the stuffing out of most people, but I wanted to vent.

According to an item in today’s New York Times, the publication of Andrew Sullivan’s new book has been delayed.

The entire print run of “The Conservative Soul,” a new book by Andrew Sullivan, the blogger and political commentator, is being discarded after Mr. Sullivan spotted a significant production error in the text: half of the fifth chapter had been inserted into the middle of the sixth chapter. Writing on his blog (time.blogs.com/daily_dish), Mr. Sullivan called the mistake “every writer’s nightmare, especially as I discovered the error myself while rereading the book late one night last week and couldn’t believe my eyes.” The finished books were already en route to retailers, so HarperCollins recalled the print run of 26,500, ordered a reprinting and delayed the publication date by one week, to Oct. 10 from Oct. 3. Kate Pruss, a spokeswoman for HarperCollins, said the printer would cover the additional costs. (The total cost of the error has not been determined.) No editorial content will be changed in the reprinted edition, Ms. Pruss said. JULIE BOSMAN

Prediction: Sometime today some rightie bloggers will claim the book was sabotaged by liberals. But I’m going to guess it was sabotaged either by outsourcing or downsizing, or both.

Without looking at the books I can’t say for sure, but it sounds as if the books were trimmed and bound with a signature out of order. This is an error that would have been made by the printer/binder, which I’m sure is an outside vendor. It’s been generations since publishers printed and bound their own books.

In brief, the process works this way: several pages (usually anywhere from 8 to 32) are printed on one big sheet of paper, and then that piece of paper is folded and the folded paper trimmed to separate the pages. Pages printed together this way make a signature. The signatures are gathered together (hopefully in order) and bound into the book. These days this process is all done by big printer-binder machines.

Back in the day it was standard procedure for the printer/binder to pull some gathered signatures (called F and Gs, for “folded and gathered”) off the printer-binder machine before binding and send them to the publisher, so that the publisher could make sure the signatures were in order. If in fact someone at HarperCollins saw F&Gs and the signatures of the F&Gs were out of order, but HarperCollins staff signed off on them, then the fault is HarperCollins’s and not the printer/binder’s. (It is possible, but extremely unlikely, that the F&Gs were OK even though the books were mis-bound.)

Back in the days when books were set by photo-offset process there was another step before F&Gs, called “blues.” Blues are blueprints of the photographic plates. The publisher received sets of blues that had been folded and gathered as the printed pages would be. The publishing production staff could check page and signature order (and other stuff) at that stage, before the actual printing. But now that photo-offset is being phased out in favor of an all-digital process, blues are going the way of raised metal type.

Over the years book publishers have been cutting more and more corners to save time and cost, and it is entirely possible that all of the checkpoints were eliminated for Andy’s book. If so, no one at HarperCollins cast eyes upon the product from the time production transmitted PDF files of the pages to the printer/binder and the time a box of the first printed and bound copies showed up in the publisher’s offices, by which time the rest of the 26,500 print run was bound and boxed and being loaded onto trucks.

On the other hand, if someone at HarperCollins did sign off on the F&Gs, likely it was a junior staffer who was swamped with a workload that would have been handled by two or three people 20 years ago.

The first books I was responsible for producing, back in the early 1970s, were printed by linotype, meaning raised metal type. When F&Gs hit my desk I had a couple of days to check them before I called the printer and gave an OK to go ahead and bind. I was expected to be painstaking and go through the book front to back several times looking for the several things that can go wrong at that stage before signing off. Today, if a staffer sees F&Gs at all, he or she is usually under a gun to give the approval that same day, or possibly within two or three hours. If the same staffer is already swamped with other books on critical deadline — well, mistakes are made.

When I got my first publishing job, ca. 1973, all of the manuscript development, editing, copyediting, and proofreading typically were done in house. By the time a book was published, several people on the publisher’s staff had examined every page, at several points in the manuscript-to-bound-books process, looking for errors. New staffers were trained in editing, copyediting, and proofreading procedures by senior staffers. Over the years more and more of the editorial functions have been outsourced or subcontracted, however. Today most copyediting and proofreader are freelanced, and good luck finding competent people who have received real training. Increasingly even manuscript development and substantive editing are freelanced, or subcontracted to a book packager. It is not unusual for a book to be published without anyone on the publisher’s regular payroll actually reading it.

Changes in technology introduce new ways for books to be screwed up. The old linotype guys who did the typesetting and created the metal plates for printing were artisans who caught many errors themselves; also, proofs would be read independently by both printers’ and publishers’ staffs. Then we switched to offset process. The linotype operators, who were Union workers, were laid off, and compositors/printers hired nonunion people (mostly young women fresh out of high school) with typing skills to keyboard the manuscript. The keyboarders rarely caught old errors but were champs at introducing new ones. But at least we could still see as many passes of proofs as we needed to ensure the books were thoroughly checked.

Once photo-offset became the standard procedure, page composition and printing/binding were done by two separate vendors.
Vendor #1, the compositor, would keyboard the manuscript and output galley proof. After proofreading and correcting, the vendor output clean “repro” proof, which would be shipped to the publisher. Usually someone at the publisher’s staff would paste the repro proof on boards to make up pages. The boards were shipped to vendor #2, the printer, to be photographed, and the images on the film were transferred to the printer plates or rollers by a chemical process. Again, the publisher re-checked the book at every step.

Desktop publishing changed procedures again, because pasteup was eliminated. It became standard to require authors to submit word processing files, so keyboarding was also eliminated. Some publishers had in-house desktop departments; others continued to use vendors. Most of the time clean, corrected pages were sent to the printer/binder by whoever did the composition, and the printer/binder photographed the pages, but in the case of complicated four-color books sometimes the compositors output film.

Now it’s all digital. The author submits Microsoft Word files of his book, which are edited, and the edited files are given to somebody (often a freelancer) who uses desktop publishing software to compose the pages. These pages are proofread once, maybe by another freelancer. With luck, somebody at the publisher gets a look at a second set of proof to make sure corrections were made, but that individual won’t have time to do a second proofreading. Instead, the staffer only skims the old “foul” proof looking for proofreader marks, and then checks new proof to be sure the error was corrected. Then the desktop compositor outputs PDF files, which are sent to a printer, and the printer imposes the files into signatures (some magazine publishers are using software that creates imposed PDF files; I’m not sure if book publishers are doing that yet) and transfers the digital impositions onto printing plates.

Sometimes in the case of mass market books two books are printed at the same time from the same plates, so that if the signatures get out of order chunks of Bodice Ripper #1 might end up in Bodice Ripper #2. It’s also possible this happens a lot and no one notices. But if you ever read a paperback romance novel in which the heroine’s name inexplicably changes from Mary to Jane, blame the printer.

It’s a wonder to me that more oopsies don’t happen. Yet the urge to downsize and outsource continue. We production editors used to joke that at least the Suits couldn’t ship our jobs overseas. We were wrong. We’re approaching the stage at which no one with hands-on responsibility for producing books for American readers speaks English. I’m serious.

So, Andy, I’m sorry about your book. At least the problem was caught before the copies were put on display at Barnes and Noble.