It's taken me two years to accept it -- that I'd have to kill my darling.
My darling was a 3000 brush submarine with a non-linear, multi-floor layout and a dynamic life support system, done in a non-photorealistic art style. I kept thinking Design could save it and so I plugged away for months on a bloated concept. Always just one better tutorial, one better puzzle, one better detail prop away from being good. Add more scripting! More complexity, more depth! Always more... Then one month became two and two became twenty-four.
The level was so difficult I couldn't complete it without resorting to cheats -- and I designed the damn thing. So I started stripping things away, digging through the debris and the cancer to rescue the concept. Delete mechanics, close-off rooms, simplify. Get to the bottom in time, quickly.
But it was already dead.
I've never thrown away so much work before. I'm sure it'd feel even worse to work in an industry that regularly de-funds entire studios and projects, or to spend a decade of your life on a space probe that plows straight into the Martian surface from a minor conversion error.
Still, this is the first time (in a long while) that I've made some real progress. This is what Radiator 1-3 looks like right now:
I hope it's better for it.
And to my darling: I'm sorry. I'm so so sorry.