Loved this interview with Goetz Graefe, the DB researcher behind 50%+ of the research papers from the past 25 years that actually make sense to me - I'm a big fan. Most of his "peers" seem to have a kind of practiced indifference to reality (no idea why). sigmodrecord.org/publications/s
Conversation
Replying to
Wholeheartedly agree. That interview is full of very pragmatic insights. Sometimes a bit sad (futility of trying to solve everything in a query planner), but very true.
1
2
Replying to
Right - in a way it's surprising that optimizers don't have far *more* problems. Also liked the related point about robust performance being hard largely because it's hard to measure or quantify.