process

1. Dream up an idea

2. Figure out which part of I think will be the most fun

3. Prototype that one thing with whatever tools are available

4. If it's not fun enough (it usually isn't), nothing else matters.  Move on and start over*

5. If it is fun, you've found your core (your major) - add production value, features and polish in support of that core.

*The prototype may tell you that something different about your fundamental idea is fun - so you might be able to start back at (2) instead of (1) - but doing that more than a couple times points to a fundamental flaw the core idea in (1)