Git is one of the most useful tools I've found in quite a while. It's just a version control system, developed for the Linux kernel, at a time when they needed a quick replacement for their existing solution. Linus was not happy with anything else available, so he decided to roll his own. What came out is, on the one hand, beautifully simple, and on the other, deeply powerful and complex.

The core idea is simple: keep track of text files as they evolve. Each copy of the data has an associated repository, and they communicate with each other, exchanging revisions. This maps well to human networks, with relationships of trust and no central authority. But when you look at what people want to do with their data, especially their source code, it gets really complicated really fast.

I've been using this wonderful tool for a few months now, and it took me a while to understand how it works, to the point I'm confortable with it. There's a Git repository that keeps track of my daily todo's and projects and contacts and notes (with two working copies, one at home and one at work), there's several repositories for work projects and most of them follow remote SVN trees, there's a repository for articles on this blog and another one for the source code, there's one for project Melody and one for its web front-end, and many more (some of the public ones are hosted on my github account). They are cheap to create, self-contained, and generally handy.

One thing you notice about Git is the user interface. It's all command-line, but it's the most well-designed command-line interface I've ever seen: excellent choices for default values, a good combination of short and long options without pointless synonyms, smart interpretation of values (the same parameter can mean a revision or a file or a remote repository, depending on its format) and good, simple configuration files. Oh, and everything has a "--color" parameter, wich makes the command-line tools a joy to use. There's gitk and gitx to visualize a repository, but I mostly found them useful when learning the basic concepts of git; the command line tools are nearly always better. I get the feeling that some smart people put a lot of thought into designing this thing.

To be honest I don't really understand most of the functionality of Git. I took the 2008 Git survey, and there was a question asking which features do you use: I ticked off a few, but for the vast majority, I have no idea what they do. Perhaps this is evidence that Git makes easy things easy and hard things possible, I'm certainly still learning new things all the time as I need them, and I'm continually impressed.

So if you're not already a fan, and you're willing to try out something new, I heartily recommend Git. And you might find one or more of these useful: github, gitx, msysgit, git-svn.