A small correction from yesterday's post. I said 'not everything has to be optimised'. Olivia rightly yelled 'GEORGE ARE YOU OKAY'. I apologise for the error. What I meant to say was 'things should be optimised to their optimal point of optimisation and no further'
One thing I have to hold myself to when writing is: forward momentum. When I'm at the writing stage, no matter my reservations, it's better to keep trying to turn out a first draft and fix the problems later than tinker with the materials to try to 'get it right' first time. Usually, it's better for me to keep on truckin' and get through a rough version of everything so I can see how it all fits together and finesse the wrong bits. This works best when there's sufficient iteration time for big rewrites as well as just revision. (But also: one of the ways to make more time for rewriting is to write the first bit as fast and scrappily as you need.)
I find, though, that I have a tendency to 'figure out what's wrong' with the fundamentals while drafting, leading me to want to go back and work those out those kinks midway through. That's generally a false economy. Those things can almost always be fixed between drafts. In fact, they're often more easily fixed between drafts than in-flight. Sometimes, they might not turn out to be an issue at all -- or else my understanding of the issue will shift. So: better to press on, you can't polish a blank page, etc. etc.
Except, that's not always the case. Sometimes, there is something major wrong in the fundamentals that needs fixing to enable you to write better and more fluidly, and the correct thing is to pause and backstep through the process.
Which is immensely frustrating, because that means my brain is constantly trying to trick me into believing that this is one of those times when that's necessary (when it almost always isn't). And I can't shut that off by saying 'that's never the right thing to do', because... sometimes it is. It's the variable reinforcement problem.
Still, this almost always holds true for me, and there doesn't need to be a perfect answer. It's gratifying to recognise a problem before turning out a bunch of words that bakes in whatever assumptions might be shaky, but that sort of draft is rarely entirely wasted effort, either.
I'm better at writing first fast and scrappy in games stuff, I think because the narrative units tend to be more tightly defined at the point of drafting. Writing for (or in) a little box in a CMS makes it much easier for me to define function and effect, and just do that bad, then better, then decent. Prose shares some general narrative units (paragraphs, scenes, sentences, etc.), but for me the connective tissue of those tends to be much less crisp when drafting. That's partly a positive affordance of the form, one of the things that makes prose prose and lets you do all those lovely prosey things, and partly something that I find makes it harder for me to be as clear and intentional early on. Perhaps that's something to poke at more in future.
(There's a further distinction between 'prose as artistic form' and 'prose as medium'. 'Prose' can be a tool deployed in other context, like games, or the end in and of itself. Maybe this is where my thinking about narrative units needs to happen. The fundamental atomic units of prose (words, sentences, paragraphs) plus the general narrative units that apply to basically every form (like scenes) are not actually the narrative units of a prose-first form like a short story or novel. Pages don't quite apply there in the same way as a more formalist medium like comics, so... what is the shape of those prosey units?)