My knowledge organization is partly inspired by this. I summarize some key points below.
Everything is organized into areas and categories. It is a two-level organization system where it is advised to not have too many of them.
IDs are assigned as serial decimal numbers from 00-99. There's no pre-defined order attached to these numbers and semantics are defined on the go as needed. While nothing stops us from going beyond 99, at this point, one may have over-organized and should split up an area.
Using this organization on a file system, nothing is ever more than two levels deep.
It allows exceptions to all the rules, which aren't many in the first place. It's pragmatic and I like it!