[pyjarrett]

I keep a lot of notebooks of general problems I run into and ways I think problems can be solved. I sketch out a bunch of ideas in a notebook and let it sit. I might do some research to see if it's been done before, and if so how, and if that solves the problem. If I keep coming back to the idea to add things, or see where it can be used, it goes in my queue of "things to do".

I prototype with basic materials (post-it notes, notecards, paper, etc.) to see if the idea is fleshed enough and has legs. I do this because a lot of my "failed projects" were just really not well-defined enough to proceed. It also helps build out the idea and helps illustrate deep-ends you should avoid in a prototype, and what an MVP might look like. Simple material prototypes help demonstrate the sorts of capabilities your program will need.

Once a project receives enough attention, I consider actually coding it up. I boil down the idea to the core of what's to be done, and try that. If it fails, maybe it wasn't a good idea, so try alternative approaches if I have any sketched out or any lessons learned from the initial attempt. I try to build up knowledge around the area of the project in-case it deserves a return in a few months or years.

If a project does actually work out, then I justify further development based on effort vs. what I need from the program.