A while back, Kin asked “What reason (other than simple laziness) would necessitate a mere patch, plug, or ignore of plot holes in a story?”
“Necessitate” is the key word here, because in writing, necessity is in the eye of the author and/or reader. There are never hard and fast rules you can just memorize and apply as needed. What there are, are consequences.
Everything a writer does in a story has consequences for the story, tilting the plot, the characters, the pacing, the theme, the readability, or some other facet in a different direction. It’s up to the writer to decide whether the consequences of any change (or lack of change) are acceptable or not.
Let’s say that I’m halfway through my current WIP and my most reliable beta reader points out that my protagonist, George, has no reason to follow Janet into the warehouse. My plot absolutely requires him to do this, but my beta is right – the action as written makes no sense.
I now have several options. I can rewrite George so that he’s a nosy dude who would follow Janet without a reason. I can decide that George doesn’t follow Janet. I can give George a plausible reason for following Janet. Or I can ignore the problem and just have George follow Janet regardless.
The simplest solution, obviously, is to ignore the problem. Fixing a minor detail that few readers will notice or care about and that is really incidental to what I’m doing may just not be worth my time. If I know that beta is super-picky, if six other beta readers didn’t even notice a problem, or if George’s reasons are really immaterial to what I’m writing, I might choose to do nothing. The consequences of leaving it alone are minimal.
On the other hand, how much work would it be to give George a reason for following Janet? Maybe a message came in for her just as she left, and George tears after her to tell her. But that changes the dynamic of the scene I had in mind; George won’t be sneaking after her to find out what she’s up to, he’ll be blundering in like a golden retriever dragging a dead squirrel into the house and expecting its owner to be pleased. It’ll take George a few minutes to realize that he’s interrupting Janet’s meeting with the head villain. That will affect the pacing and tone of the incident, as well as the ongoing characterization of George, and very likely alter the plot rolling forward.
There are a bunch of reasons why I might decide to go for this sort of minimal patch. I may like bumbling, clueless George better than sneaky George. I may realize that introducing some comic relief and/or slowing the pace a bit at this point in the story is an excellent idea. The new scene dynamics may spin off cool new ideas for the direction of the future plot. The consequences of changing turn out to be desirable, so I do it.
Or those same consequences may be highly undesirable. Slowing down this scene may throw the overall pace totally off. The scene may need more angst, not comedy. Or possibly I was intending this incident to be the point at which bumbling-George starts to transition into sneaky-George (character development!)…which is clearly not working for my beta-reader, but which equally clearly is not going to be fixed by making George’s reason for following Janet one that’s consistent with his bumbling personality so far.
It’s also possible that I know perfectly well why George went sneaking after Janet – it’s due to a complicated piece of his backstory that will take two or three pages to explain adequately. In this case, making up some other reason will vitiate the real, complicated-backstory reason. But dropping a three-page infodump about why George sneaks off after Janet will a) slow the story to a crawl and b) make the whole sneaking-after-Janet incident and the backstory seem a lot more important to George’s development than I want them to be. Since those are unacceptable consequences, I’d have to either go back and find places to drop in bits of the backstory, so that by the time George sneaks off, the reader already has some idea why he would, or else, again, ignore the problem and hope the readers will trust me until two scenes later when George’s reasons come out naturally.
Rewriting George to be a sneaky, nosy dude right from the start, and ditching the rest of my planned plot in order to have George behave in-character with his presentation so far are, to my mind, different aspects of the rebuild-the-boat-in-drydock fix. If I’ve been unhappy with bumbling-George and/or the overall direction of the book, I’ll go back and rewrite; if I like bumbling-George and can see fascinating new plot-jungles ahead arising from his decision not to follow Janet, I’ll pitch my plan and revamp the rest of the story. It depends on whether the story is better, or more fun, or more interesting to write if I go in a completely new direction. If, however, I really wanted to do bumbling-George evolving into sneaky-George, rewriting George to be sneaky from the get-go won’t work for me, even if it fixes the problem.
For some writers, it is not acceptable to leave even the tiniest plot hole, even if the consequences of fixing it are that the story becomes a turgid mass of explanations for why various alternatives didn’t happen or wouldn’t work or were never tried, at the expense of the actual plot and pacing. For other writers, gaping holes are tolerable as long as the plot moves according to their initial vision. It depends on the writer.