Andrew Dumont makes an interesting point:
Know that when you start just a side project, you're starting so much more. It'll completely consume you. The worst failure in any side project is to devote time, energy and sanity for any sustained period only to close the doors.
Side projects are a means to an end.
They need to start with an end-state in mind - create a passive income stream, validate my idea. They need to have deadlines and key metrics - six months to profitability, 10 paying users to validate my idea. But most importantly, they need to be a sprint. The longer a project lingers, the harder it becomes to keep morale high and pull the plug if it's not working out.
There is, of course, a valid point there. Then again, I am reminded of those people who assault young musicians with questions like "But why do you want to play the violin? What are you trying to achieve?"
Isn't playing violin and starting side-projects just for the fun of it enough for a start?
I am pretty sure that most people who eventually started side projects with clear end-states in mind first started side-projects just for the heck of it. They probably did that for years before the thought that a side-project could have an end-state right from the beginning entered their mind. That certainly was the case for me: most of my early programming, writing or web activities were done purely for the heck of it.
So, perhaps better advice is to be aware that some day you will want to graduate to definite goals.
But if you want to start a side-project today for the heck of it? Go for it.
If you read this far, you should follow me on twitter here.