got a little example project going exploring lexicon dev tools for Go, this one on the "using existing schemas with codegen" side, with @atwork.place@bsky.brid.gy as an example
bnewbold.net/atwork-cli
(tbc I think the deadnaming concern is legit, and I don't think there is anything wrong or suspicious with users wanting to be able to move on from historical handles)
there might be a better balance than the current system, but i'm not sure simply dropping handle history is it.
my bigger privacy/history concern is if folks start using DID docs for more social/personal info: for example account linkage in alsoKnownAs array, or many other service and key entries.
handle history being public is pretty helpful as a trust signal both for moderators and individuals. bot nets and engagement farms really frequently flip accounts around.
in the big picture there are a bunch of needs and perspectives to balance.
removing handle history from PLC just adds friction. bad actors could have kept a verifiable log, and there will be signals in records (eg, mentions) than can be looked up. friction can be powerful, but depends on situation
tangentially, the Wayback Machine indexes URLs in "CDX" format, which flips around domain name order so everything goes left-to-right
good DNS lore here for "deeply hierarchical" handle and NSID lovers, from @jbcrawford@hachyderm.io
memories of .us
it has been a public known issue for a long time, though I don't have a link
a simplistic take is that atproto retains individual authorship and allows "backsies". that works for a lot of group/collective stuff, but maybe not all.
use the right tool for the job, AT is just one of many. tangled is a great example of hybrid with other systems (git)
atproto lexicon for reviewing and staring debian packages