I commented on a Mastodon issue regarding supporting multiple domains.

I think this can be super important for making users' identities independent of the instances they connect to, and can be part of a seamless user migration strategy.

https://github.com/tootsuite/mastodon/issues/1211

Hat-tip to @danyork for the idea.

I followed up to this with a comment on the Mastodon user migration issue.

Folks there have been thinking about migration for quite some time, but after quickly skimming the thread it looks this idea hadn't really come up.

https://github.com/tootsuite/mastodon/issues/177#issuecomment-294047505

Have now realized what is probably the main roadblock for this: TLS certs.

At small scale (thanks to letsencrypt) that's no biggie, but at large scale having a cert per-user may be unworkable. User-configurable TLS certs also requires integration that may generally be missing.

I still think this is worth doing, but first draft implementations of this would probably be HTTP only.

@HerraBRE probably not ready yet but might be worth consider some of the emerging #blockchain sovereign identity solutions emerging for the future as well. Not that this will speed up Mastodon adoption by any means.
@crscheid For the most part, I'd say we're stuck using whatever tech the standard browsers support. Someday...

@HerraBRE

why does it need to be a cname record? couldn't any standardized DNS txt record work?

just like email, where the dns has an MX record telling you where the federated email server is, DNS can have a record telling you where the federated social server for the domain lives.

@krypteia Because if it's a CNAME or an A record, then all the existing mechanisms just work. AFAICT, this could work tomorrow across the entire #fediverse if someone patched Mastodon to support multiple domains.

If you create a new record type, all the software everywhere needs to be updated to understand it.