Actually, one of the benefits of Git over, say, Subversion, is that it's much easier to hack away on multiple things, and then arrange a series of sensible, self-contained commits from your messy working directory.
It's not necessarily a great way to work at all times, but it's a handy facility to have. Personally I use it most in the early, somewhat experimental stages of a new feature or project, when I find it useful to mentally separate experimental hacking with organised committing.
For most things, though, I try to focus on one thing at a time, and here Git's quick-'n'-easy branching and merging helps out. Whenever I start something new, even if it's just fixing a minor bug, I'll create a new branch just for that work, often naming it after the ticket number in our issue tracking system. Then your being in that branch is a constant reminder of what you're supposed to be working on.
It's not necessarily a great way to work at all times, but it's a handy facility to have. Personally I use it most in the early, somewhat experimental stages of a new feature or project, when I find it useful to mentally separate experimental hacking with organised committing.
For most things, though, I try to focus on one thing at a time, and here Git's quick-'n'-easy branching and merging helps out. Whenever I start something new, even if it's just fixing a minor bug, I'll create a new branch just for that work, often naming it after the ticket number in our issue tracking system. Then your being in that branch is a constant reminder of what you're supposed to be working on.
Ryan Tomayko has written a good article on these aspects of Git here: http://tomayko.com/writings/the-thing-about-git