Tuesday, March 13, 2007

Acquiring proficiencies by oscilating centralization

I still have to work on the title but it goes something like that. It is about the way I learn, and probably most other too. It applies to getting good at complex tasks. Say driving, relationships, or brushing your teeth. Any task could be seen as complex, I guess. It doesn't work much for achiving perfection, but rather proficency (or the hyperbole of proficency).

Oscilationg centralization means to swing back and forth between centralization and decentralization. How does that apply to learning.. Ummm.

Say the first part to getting good at something is to do it the first time, by any means necessary. If you need to get from A to Z, at this point the important part is getting to Z, not that it took you 56 steps or that you happened to pass by R eight times.

Now consider that you have just done this for the first time and it took you 56 steps. Now you are asked to race against others to get from A to Z. What are you going to do? Probably the same 56 steps, only faster. That is the centralized effort. Centralized in the sense that all of you is centered on this one problem.

If you only attack this problem with a centralized effort the algorithm probably won't get much below 56 steps, which is why you need the decentralized effort.

The decentralized effort is not simply to be dealing with more than one problem, but when different parts of you are dealing with their own problems. Sleep is the ultimate in personal decentralization which is why we get tripped out imagery and thoughts we normally don't. As the different centers do their own thing and communicate casually with each other new elements connect that perhaps never have before. Like, say, microwaves and cooking.

So as part of you is working on this problem you have solved before, though crudely, parts of you are solving other problems and your mind may wander. As this happens interconnections do their thing and you get ideas about how to go about the problem differently. You will get lots of ideas, some bad ones, but some fraction greater than zero will help make the solution easier.

So then perhaps you ween your A to Z task down to 40 steps. Sounds good. Time to test it. How? Centralized effort. You do the same 40 steps faster. The new algorithm may have fewer steps, but some or all of the steps may be harder or cause unanticipated bottle necks and such. Good thing you tested it.

Now you will have new information to use when you return to your decentralized effort. But what if you stay decentralized? Wouldn't you get more and more ideas? Yes, but.

Decentralization is a safehouse free of judgement so that ideas can be generated. Centralization is the reality check so that you know which ideas are worth keeping. It is important to do this periodically so that you know what you need ideas about, ie which parts of the algorithm are tripping you up.

Visualization can play a big role in both identifying problem areas and dismissing bad ideas, thus eliminating a lot of testing, but not all of it. Reality always has surprises for us.

No comments:

Post a Comment