[whee]

I think you are analyzing too early. If this project is tackling a problem you haven't solved before, think of your first implementation as a prototype. You likely don't understand the problem until you have spent time solving it.

After you have something that works, you will have a good feeling for what you should've written the first time.

Personally, I don't abstract until I need it. Writing code because you "probably will need it later" can cause unnecessary work and be a distraction.