Proofreading vs. Copyediting
Folks seemed to get enough out of my take on copyediting that I thought I'd talk a little bit about proofreading.

First, let's define the terms, because what is called "proofreading" by folks outside the publishing industry is usually copyediting instead. The copyeditor works on the book at the manuscript stage, marking grammatical and spelling mistakes, querying inconsistencies and awkward phrasing, verifying facts, preparing a style sheet, and keymarking the manuscript for design. The proofreader works on the book at the galley proof or page proof stage, comparing every word of the manuscript with every word of the proof, verifying correct word breaks, making sure that all editorial changes were input, and (with some publishers) verifying that elements were set according to design specifications.

Proofreaders are not supposed to make substantive changes to a manuscript (a copyeditor shouldn't even make substantive changes without querying) or go against the style sheet provided by the copyeditor, but they do serve as an additional reader. No one—no author, editor, copyeditor, or compositor—is perfect, so the proofreader is expected to pick up any mistakes the others missed. While copyeditors are expected to complete about ten manuscript pages an hour, a proofreader (in fiction) is expected to complete ten to twelve book pages an hour—and book pages are usually much more dense than manuscript.

Proofreading is a tough job. Even for someone with a good memory like me, it's difficult to hold more than a few words in your head at a time without leaving out a comma or mixing any of the words up—confusing an east for a west, for example (a very common compositor error when books are reset)—so you can only read four or five words from the manuscript before comparing it to the proof. It's intensely time-consuming and tedious.

How do proofreaders manage to do so much in such a short time? Well, perhaps someone out there can truly manage it, but I never could. I haven't accepted proofreading for years because I ended up spending far more hours than I could charge on every book I did. From years spent conscientiously supervising copyeditors and proofreaders, though, I can verify that many proofreaders make do by simply not proofreading: they read the proofs but trust that they were set from disk and don't compare the manuscript and proof word for word. This approach is very problematic, though, and can lead to errors because a) it often happens that the disk doesn't contain quite the same version as the manuscript, for various reasons; b) it's easy to miss changes that should have been input, because you don't have a finger each on the manuscript and the proof at all times (the only way I could ever proofread accurately); and c) the book sometimes hasn't been set from disk at all. I suspect that the proofreaders who don't take such shortcuts simply resign themselves to losing their butts on wages, the way I did.

So be appreciative of your proofreaders. They are your last line of defense against mistakes in your novel.