News for hackers. Stuff that matters.
Re:MySQL (InnoDB) clustered and non-clustered indexes by misframer on Wednesday June 04, 2014 @09:25AM (#7845378)
A Z-order curve maps from two (or more) dimensions to one while preserving locality: http://en.wikipedia.org/wiki/Z-order_curve
Re:MySQL (InnoDB) clustered and non-clustered indexes by fredophile on Wednesday June 04, 2014 @10:32AM (#7845722)
Depending on your application you might want to pick a different space filling curve (http://en.wikipedia.org/wiki/Space-filling_curve). Different curves lead to different characteristics for nearness. You can also use multiple types of curves or two of the same curve oriented differently and then use the minimum distance of the curves as your distance estimate for some applications.

If you just want to generate keys you should look at Locality-Sensitive Hashing (http://en.wikipedia.org/wiki/Locality-sensitive_hashing).

Re:MySQL (InnoDB) clustered and non-clustered indexes by falcolas on Wednesday June 04, 2014 @10:13AM (#7845616)
That is most likely it, I was guessing they used the interleaving method presented there, but didn't know the formal term.