Who needs an rdb, the trilogy

I’ve followed up original “Who needs an rdb?” post with two others over on the DI blog:

• TorQ: No stinking rdbs - Adding a TorQ pack to let others easily copy my approach
• The last good reason for an rdb: the return of the attribute - And an approach to have appendable attributes, and close the remaining performance gap with rdbs

I wanted to call it my rdbs are shit trilogy, but I was advised against it by people who are less fun. Hopefully people find the ideas useful!