Some people struggle to find ideas for working on side/personal software development projects for learning (and fun!). I’ve kept a running list of ideas in an online notebook and if I’m struggling for an idea for a new project I refer back to this list.
Over the past couple of years I stumbled across a series of inter-related projects that’s been a goldmine of learning opportunities, from frontend to backend to serverless.
This series of projects started with a question “Can you solve a problem without understanding the problem?”. tl;dr? No. The problem for this activity was solving Sudoku puzzles, and you can read more about this here:
This lead to the next step, researching more about exact cover problems, and how they can be solved with well established algorithms:
After building an implementation of Donald Knuth’s Algorithm X, I packaged it up as an AWS Lambda, and then built a React frontend for it:
I have one other project in progress for the frontend, replacing Flux with Redux. I’m still working on that one.
After building my solver and the frontend, I starting thinking about how to generate new puzzles. This is in progress here: https://github.com/kevinhooke/sudoku-generator
If you’re generating new puzzles you also need a way to grade the difficulty of puzzles. The unusual thing about this is there’s no established algorithm for grading the complexity of Sudoku puzzles, they’re typically graded by applying human solving techniques, so this led to this: https://github.com/kevinhooke/sudoku-human-grader
This string of related projects has kept me busy of the past couple of years. Not every idea will lead to a series of related projects like this, but if you can find an idea that does, it will keep you busy with plenty of problems to solve and many learning opportunities.