Programming just feels like a large game of managing expectations. It often feels like I never have time to do anything "right", so I have to make trade-offs before some impending deadline. This is a serious source of discomfort in this profession that never seems to want to go away.
To answer your question, I'd write a fantasy economic sim game like a sequel to Patrician III or something, except with wizards and weird races and stuff. A big bit of the game would be negotiating with customers for custom goods and services.
Exactly how much money would it cost to for a casting of Bull's Strength before a mercenary engagement? What are the obligations a wizard has in the event their town gets attacked? When is the Teleport spell economically viable vs just taking a boat? How do you negotiate with beings very alien to you?
Classic pet project that I've wanted to do for years, just never had the time/concentration to do it.
Sure, it may be possible to get things working smoothly today, but that doesn't mean that all the areas I mentioned (documentation, etc.) have caught up. The primary reason they don't catch up is the straggler libraries keep many teams in a past era.
Right now, the state-of-the-art is exciting but falls short in some world-changing areas like question-answering. As much as I want to, though, I have no idea how to push the field forward, however.
For the other things I want to do, I am probably sufficiently skilled to do, but I just need the time and motivation.
If I could be bothered, I would write a really detailed political simulator a la Democracy but with much more detail and elections.
When I see the same thing, I think a lot of better stories can emerge - a lifetime bond between comrades, a romance between a prisoner and the nurse who comes to heal him, rivalries between factions, a 500 year old cyborg who questions her humanity. So many stories can emerge.
disregarding transaction fees of course, otherwise it was horrible :)
But then I grew tired of waiting and just did it already :)
A content-first web browser (think Reader-Mode for every page), depends on: a) Learning the differences between the HTTP|HTML|CSS|JS specs, and what browsers actually implement, and b) heuristically lifting out the content.
b) is a programming problem, not easy to solve well, but there's been a lot of work in that area to lean on.
a) is a documentation/people problem. Not my area of expertise, but definitely the harder problem of the two.
