Question 945

Two Copies, One Truth

A comments table stores author_display_name on every row so the feed renders without joining the users table. A user changes their display name, and the new name appears on their profile but old comments still show the previous one. What is the correct reading of this situation?

A

This is the expected cost of denormalization: the duplicated field needs its own update path, so a rename must fan out to every copy

B

This is a caching bug: the feed is serving a stale rendered page, and shortening the cache TTL will make the new name appear on old comments

C

This is a transaction isolation problem: the rename ran outside a serializable transaction, so other sessions still observe the prior value

D

This is a replication lag problem: the feed reads a follower replica that has not yet applied the rename, and it will correct itself shortly