Regression Testing for the Novel
I edit a great deal. I wish I didn’t have to, but in general, after a single draft, my work is compelling but shaky. Getting from that stage to the point where I can happily shelve a project takes draft after draft. For novels, the process generally looks something like this:
- Read through the book.
- Notice lots of problems.
- Fix these problems.
- Return to 1.
“Ah, but how do you know you’re finished?” you may ask, reading these steps. You never know you’re finished, because you’re never finished. Someone once said that works of art are only ever abandoned, which makes sense to me. Michaelangelo would still be painting the Sistine Chapel today if they hadn’t kicked him out.
One of the dangers of repeated editing is that you risk undoing a previous draft’s work. A given sentence may seem awkward at first glance. Ah, you think, why didn’t you write it this other way, which rolls better off the tongue? The answer, if you’ve edited your work enough, may well be: “because the more natural way sounds barbarous when read in context with the rest of the paragraph.”
This doesn’t mean you should abandon the awkward sentence-the goal is to rewrite it in a way that sounds natural, yet avoids your earlier problems. To do this, though, you need to remember what your earlier problems were.
In software, when you fix a problem with your code you write a test to make certain your fix works. Those tests stick around after the problem’s gone; whenever you adjust your code, you run your tests to ensure you haven’t broken any prior work. I wonder if the same kind of thing could work for editing. Maybe a better versioning system would help (Track Changes, superhero edition), or perhaps a good commenting framework. Food for thought. For now, I’ll live (and edit) in uncertainty.