Michael Stonebraker (won the Turing award in 2014) pointed out that databases are like a Tardis. Well,
that's my analogy, not his. The point is: they are alot bigger on the inside. There's a ton of stuff to tackle,
and they hit every area of computer science, even though they seem kind of simple from the outside.
For personal projects, it pays better to be a little selfish write something you yourself want.
Since you are the user, you will give yourself much better feedback. If you don't know the use case,
you can't be the user, and that's a sign that your project has little hope of being useful.
This is the same advice that Paul Graham gives for problem selection; that writers are given
when they hear, "write what you know".
So ask yourself, "what software do I want?"