Best practice with git, and why

Dr. Slack, Wed Sept 02 2015, 05:05PM

I'm using git to safeguard an application I'm working on. It's got beyond the toy application stage, several libraries, several tens of classes. Unfortunately, I've not used git, or any source control software before, and beyond doing a commit -a with sensible comments, don't really know what I'm doing, or how it's going to help. I guess I'm trying to throw myself in at the deep end to see what source control can do, having not really understood the advantages or the workflows from just reading about it.

Now I come to add a new function, and realise that it's getting messy because some of my classes are rather badly partitioned. So I need to refactor, and know that I will probably break it several times before the better class design is working.

So do I create two branches, one for refactoring and one for new function experimenting, or use head for one of those? And why? What quirks will I discover about the process that makes a difference to how I should use it? Are there any links that will direct me how to use it to best advantage?