I've been working on Notes From Province for 4 years, and during that time, I've gotten much better at making Notes From Province.

Which sounds awesome. Except I've been making the game chronologically. So middle of the game is much higher quality than the start, and the end is higher quality than the middle. And you can't have the start be the worst part of the game. That's when you hook the players. So the solution is obvious. I go back and remake the beginning. But you can't have the middle be the worst part of the game, or else player's won't stick through it until the end. So the solution is obvious. I go back and remake the middle.

If anyone knows how to not get stuck in an infinite development cycle while maintaining a consistent standard of quality let me know. For now, I have three ideas.

Prefer adding rather then remaking
If we add new, better content in between the old, then the average quality rises. I'm lucky with this as the first area has noticeably fewer dungeons than the later zones, creating some gaping content holes.

Define what "good" content is
Some earlier blog posts described a list of rules for making good puzzles and encounters. With something concrete to point to, I can objectively tell if a part of the game is below standard and needs reworking. If it meets the rules, leave it.

Develop with the intention of remaking
Not sure yet if this is a good idea or a horrible one, but I'm developing the rest of the game as if it were a draft. Placeholder treasure and monsters between unfinished cutscenes. Leave out the things that were likely to be remade anyway and once the game exists in it's entirety, come back and fill them in.