Why codegraph is the way it is: why facts and inferences never share an edge, why no build is required and what that costs, why identity is a key and never a parsed string, why the city is a model before it is a picture, and what codegraph took from Moose, CodeCity, Structure101 and Gource. Read these when a tutorial or a how-to left you asking “but why”.

Several pages are the project’s design records, kept with their numbered sections (AN-1, CM-2, NV-4…) because the code and the other pages cite them.