The projects that stick
My favorite projects are usually the ones where the problem is concrete enough to test, but open enough to make design decisions matter.
I like when a project has a loop:
- Build a small piece.
- Try it with realistic data.
- Notice what feels slow or confusing.
- Improve the next version.
What I pay attention to
I care about whether the tool makes the user faster, whether the code is understandable later, and whether the project teaches me something beyond the assignment.