So there's an adage in computer science, it comes from Fred Brooks and his very important book The Mythical Man-Month. I think of it as "Brooks's Other Law" ("Brooks's Law" being the more famous dictum, "adding manpower to a late software project makes it later."): regarding building pilot or prototype versions of complex systems, he writes, "plan to throw one away; you will, anyhow." That is, if you dive right in to building what you intend to be your production system, you will make all kinds of mistakes, and instead of noting them for next time and moving on, you will have to cobble together solutions in-place, and by the end either you'll have Ship-of-Theseus'ed the whole thing or it will simply be unfixable and you'll have to start over. If you plan to build one to throw away from the start, you'll end up having done the same or less work, and you'll have an equivalent or better end result.
All of which to say: perhaps it's the case that whatever your conscious intention when you set out on a writing project, one way or another, whether by unconscious obsession, conscious inclusion, or conspicuously deliberate avoidance, you will, anyhow.
