A long time ago, I was keen on a CODASYL database called IDMS. It had records related by pointers. Along came IBM, who didn't have a Database, but did have indexed sequential files. Data in one could be related to another using the index instead of a pointer. Not as good, but invent a new 'paradym' called relational database, and Bob's your uncle!
At last, many years later, I'm delighted to find that Neo4j represents a return to the best principles I aspired to back then.
