Writing wonkery ahead.
Over the weekend I finished putting together the scene map for Always Outnumbered, Never Outgunned, and slotted into it the "things to do" collected from the notes I'd kept during the first draft process. Some details follow for those not in the know.
To sustain the momentum I need to write and finish a first draft, I tend to postpone any major changes that might need to be made. I keep a document named "Things to do in the next draft" (TTDITND) in the wiki for the project, and if anything comes to mind that seems like it needs fixing or some close attention, I note it down in there and keep going.
When the rewrite process starts, I make a map of every scene in the book as it's been written — rather than as I originally imagined it — and use that to get my bearings. Then I take everything out of the Things To Do document and try to find places in the story where those questions/worries/nitpicks can be addressed, by way of the scene map. If it's something I've already handled, or if it's no longer relevant, it gets filed away. Then, during the writing of the 2nd draft, I open a new Things To Do for that draft, and the process begins again.
Once I reach a point where it's just polishing this or that bit of brass, so to speak, the book is most likely finished. It generally doesn't take more than three or four total iterations to get to that point anyway. One draft to get it down, another draft to get it right, a third draft to get it cleaned up. Everything beyond that is diminishing returns.
Something else about this iterative process: It forces you to notice how details that seemed important earlier on either recede far into the background on their own, or can be elegantly folded into other concerns. Sometimes those concerns turn out to be utterly trivial, and in ways that make me smack my forehead and mumble, "Why the bejesus was I ever getting my scarf in a wad over this?" I must have noted down something like two dozen impossibly trivial details in this vein — not even continuity stuff, but more like implications of implications, things that nobody except me and maybe one other person would ever know, let alone care about, let alone allow that to affect their feelings about the quality of the work.
François Truffaut once made a movie, Day For Night, arguably still the best movie about moviemaking and certainly the most affectionate. The movie being made in the film is probably not very good, but the movie at large is not about that; it's about the surrogate family that forms on a film set, about the madcap circus of the whole thing. It is also about the process of creativity, and how the discipline enforced by the process allows great things to happen.
At one point the Truffaut-surrogate in the film muses, "Making a film is like a stagecoach ride in the Old West. When you start, you are hoping for a pleasant trip. By the halfway point, you just hope to survive."
The same goes for a novel, although my take on it is a little less despairing. At some point you realize the quality of the work is not going to be amplified by the depth of your fretting, so you learn to trust your own high-level decisions, to be attentive to the details that matter and to let go of the ones that don't. You don't pre-emptively censor the fretting; rather, you let the fretting come out in your iterating over the work as a way of turning it loose and letting go of it. Eventually, you turn yourself over to the process, and let that buoy you up. In the end, you are surprised in the best possible way. Then you take what you've learned from the whole thing, roll it up into a few lessons for yourself, and start on the next project.
There is another line from the movie that comes to mind, where an actress is reminiscing about her mother, also an actress: "She hated the way we shoot movies - in bits and pieces! I remember escorting her to the premiere of her first big Hollywood movie. A fantastic evening! When the film ended, she sat there, then turned to me and said: 'I did all that? All I remember is the waiting.'"