At some point, I think I forgot I was making a game. It became more like this expansive simulation of every possible reality, folding into itself; I caught myself wanting to make everything interchangable with everything else, to let every piece talk to another piece and act like yet another. Everything was a network, a graph, and every node was another network. Recursion upon recursion.
There's a tendency in game development to think that you will always be coding "the" framework you can re-use and re-use for everything. Everything should be modular, endlessly recyclable, endlessly useful. It's hard to let that go and accept that you should focus on making a game, first -- a game that works -- and not an engine or a library or something that'll change everything as we know it. I ended up forgetting the most basic lessons of making.
Convo felt too big, and it felt too big by 2/3. I had to cut 2/3. I don't know how I can quantify a fraction like that, but I thought about my ability to build and engineer and develop, within the timeframe I wanted, and decided it was overextended by at least 2/3. I've cut a bunch of stuff, but I'm not sure if the scraps on the floor -- do they add up? Is it enough? How do you know if you've cut too much? Is this what Peter Molyneux feels like?
More details soon...