> In the 90's many of us (yes including me) thought that object databases would solve the problem by eliminating relations on the disk. We all know how that wor

> In the 90's many of us (yes including me) thought that object databases would solve the problem by eliminating relations on the disk. We all know how that worked out. Maybe I'm too young (or too old) because I don't know what he is suggesting here. I've used an object database on a large commercial project, and it was a dream: everything I like about ORMs, and none of the drawbacks. My impression of OODBs, therefore, is that they were a technical success, but a commercial (and open-source) failure. But I don't think that's what Martin is suggesting. Can someone explain what "we all know" about database history here?