Right now, a subset of our replication triggers are set to be "verbose" – that is, include the full set of columns in the "before" data for updates and deletes – which was added for the original search server updater code written by murdos, via MBS-5249. This isn't really our approach to that at this point, so it seems like we should probably standardize and remove the unneeded distinction.

Questions:
1.) does anyone use this data that I don't know about? presumably we don't strictly support people using our replication packets outside of proper systems for applying them, but it's possible. Rob, I guess you would be the one who knows?
2.) which way do we want to standardize anyway? It's possible that we'd like to keep this data there and make it this way for everything, instead of the only record for a deletion (for example) being the PK of the deleted row, though it does of course mean replication packets are somewhat larger (perhaps insignificantly, since they're compressed anyway?)