The World of Blocks

Imagine a child playing with blocks. Imagine the child’s mind contains a host of smaller minds - “agents”. Imagine an agent called Builder in control. Builder makes towers from blocks:

block-tower-build

Builder is a simple agent. It needs help from several other agents to choose a place to start the tower, add a new block to the tower, and decide when it is high enough:

builder-agent-tree

But doesn’t Add have a big job as well, too big for a single, simple agent? First Add must Find a new block, then the hand must Get that block and Put it on the tower top:

add-agent-tree

Why break things into such small parts? Because minds, like towers, are made that way – except they’re composed of processes instead of blocks. Scientific workflows are also made that way.

Does making stacks of blocks seem insignificant? You probably didn’t always feel that way. You may have spent joyful weeks in early childhood on building stacks of blocks. As grown-ups, we all know how to do such things, but how did we learn to do them?

This forgetfulness, this amnesia of infancy, applies to expert scientific practice as well. All our wonderful methods weren’t always inside our minds – they began and grew, like we begin and grow computer-simulation programs and data analysis pipelines.

This post was adapted from a note sent to my email list on Machine-Centric Science.
I'd love for you to subscribe.