So, Merge Sort. It's probably the most commonly used sort these days. I read that somewhere. People think it's Quicksort, but they're wrong. I read that somewhere, too. There are lots of cool things about Merge Sort, like the fact that it's the fastest known sort for linked lists. And you can even write it to be in-place on linked lists. Anyway, it comes to us from my homie John von Neumann in 1945. (Think about that. 1945!) John von Neumann was awesome. We ought to sing a song about him every morning. Maybe it could be to the tune of Handel's Hallelujah Chorus (Handel's my other homie). Johhhhn von Neumann! Johhhhn von Neumann. John von Neumann! John von Neumann! You get the idea.

So, how does it work. Well, you get a list to sort. First you check if it's small. If you have zero items or just one, then nothing to do. Otherwise, make two parts. Recurse, recurse. Then put the two parts together into one big sorted part. That's merging (so Merge Sort; makes sense, right? Right???). Then you're done.

Anyway, cool stuff about it. It's stable. It's log-linear time. (You can do the merge in linear time, so Master Theorem, figure, figure, figure, yup, log-linear time.) It works on just about anything: arrays, linked lists, whatever. But there's one nasty thing about it, and it's that to do the merge on an array you need to make a big ol' buffer, so linear extra space. Ick.