datalevin 2026-06-30

Hi all, thanks again for volunteering to review the Datalevin book. If possible, I’d prefer to receive your reviews by July 3rd so I have time to incorporate feedback. Much appreciated!

šŸ‘ 3
šŸ‘€ 1

My one note that I've been able to glean from the sections that I've gone through, is that I'd appreciate an elaboration on how tupleAttrs compare to a unique-composite; in the text they both sounded functionally pretty similar, and I didn't know if there's any reason to use one over the other at a given time (a la performance trade-offs)

what do you mean by 'a unique-composite'?

There's only one kind of composite tuples. Are you talking about its comparison with stored tuple types? These are quite different.

Halfway in the book, i'll try to shift gears šŸ˜… šŸŽļø. So far the most important item that wasn't very clear for me has been answered in the channel: the conn vs (db conn) question, and how this contrasts to xtdb/datomic where you have db snapshots. Overall it's been a delight to read.

šŸ‘ 3

oh wow, apologies on two fronts, I never saw the notification for this response-- didn't mean to vanish, and I also seem to have crossed some wires while reading- don't mind my illiteracy

šŸ˜€ 1