For Christmas, Batman gave me a Rubix cube. In Dunedin my flatmate Sandy taught me how to do one, and I could remember a few of the steps, but solving it outright took some practice. After memorising the solving algorithm, I've been able to do it pretty quickly since early January. This week, I taught someone else how to do it; he's down to under 4 minutes.

Rubix cubes are one of those beautiful enigma puzzles. The average Joe suspects that the first step is to make one whole side and then loses their way. The cube when mixed up appears horrifyingly complex, yet once the method, the system, are locked away in the memory bank, it becomes a case of applying several steps of an algorithm.

Sometimes, it feels like this is the way medicine could be done. First, orientate the patient cube in the correct manner. Find the white side and ensure it faces up, and the yellow side faces down.

Step One; make the cross. Ensure you know what it is you're moving. Where you're moving it from and to, and in which order. Make a plan. After you think you've done the steps, check your work. Do the next step, the corners. Gather information, make a plan, action it, check the results.

This routine continues whilst all the steps are performed. And, in most cases, it'll work. The challenge comes in picking the cube that's had the stickers pulled off and put on in the wrong place, or the cube that can't turn in certain ways.

So, it's not really how medicine can be done at all. It's a good safe way for students to learn, to apply a system, to check results. It may not always be the quickest, but for most cubes, they'll face up just right.

But down the track, you don't want to be twisting for hours some poor cube whose stickers are just on wrong.