Someone should create a (dynamic) DNS provider (ideally a non-profit) whose only stated mission is to help people take ownership of their Fediverse, e-mail and XMPP identities.
If my identity were herrabre.mastodon.xyz (a subdomain instead of user-@), I could move to another instance by requesting a DNS record change, and my social graph would remain intact.
A trustable social contract would be needed so I kept effective ownership of the subdomain; the non-profit would be responsible for that.
This is noble. I'm curious how history and resolution of identities (from identity @ domain to resolving from a domain alone) would work.
Is it something that the #IndieWeb can play a role in? Having people own a domain, following others and taking ownership + reign that way?
@cwebber Perfect is the enemy of good, right?
DNS is full of warts, but it's what we have today.
Giving people more autonomy within the systems we have today motivates me; I'm tired of waiting for the flawless replacements from the future.
But I'm not sure how to get this particular idea off the ground.
@HerraBRE Sure! DNS and SSL certificate authorities are the centralized systems that plague our decentralized systems and make them not very decentralized after all, but given that it's difficult to move people off them, providing improved spaces within them is still good. SSL CAs are an awful design, but at least Let's Encrypt has reduced the awfulness level a lot, and that's nothing to sneeze at. Doing so for DNS likewise could be very good.
@frankiesaxx @cwebber I loved the ideals behind OpenID! Today all that is left is "sign in with Facebook/Titter/Google".
AFAIK, nobody bothers to support anything else.
We may not like the solution, but the consumer does. Kontalk (XMPP) and down the road also Matrix are trying to also use phone numbers, but the sad news is it's intrinsically not really something you can do in a decentralized way.
https://meta.stackexchange.com/questions/307647/support-for-openid-ends-on-july-1-2018
Bonus meme link:
https://meta.stackexchange.com/questions/207388/why-is-the-handlesuccess-method-such-a-terrible-one
@profoundlynerdy @HerraBRE No need for namecoin... a petnames system is better (but can include a namecoin-like system as an equal participant)
@profoundlynerdy @HerraBRE I think Namecoin is really too much like DNS to be the *root* naming system. However that doesn't mean Namecoin, something Namecoin-like, or even DNS are entities that should not exist... to the contrary, we will always want naming hubs, but in a petnames system dns and namecoin are equal participants among many.
@HerraBRE aren't the encryption keys tied to your domain so if your domain changes then your encryption keys change and then you lose all your account data?
@gme What encryption keys?
@HerraBRE the encryption keys that protect the data while its at rest. i'm pretty sure hubzilla has that not sure about mastodon but wouldn't be surprised. @deaduperhero do you know?
@gme Doesn't matter, if you're moving the provider has to decrypt it and give you a copy of your data that you can work with.
Unless they refuse to give you your data, in which case you GDPR their assets. ;-)
@HerraBRE you're assuming the provider is in europe. i'm in the US and toot.zone is in Canada. I'm not complying with GDPR that I know of. Actually I haven't researched GDPR since I have no European presence.
@gme It's a red herring either way.
If your instance refuses to give you your data you're screwed (you may have legal recourse, maybe not).
If they cooperate, whether they store it encrypted or not is irrelevant, they just decrypt and give you data you can work with.
@HerraBRE We have a phrase here in the States that's very popular: "You can't squeeze blood out of a turnip." 😀
https://www.usingenglish.com/reference/idioms/squeeze+blood+out+of+a+turnip.html
@HerraBRE good tools and docu could make it feasible for many users to host their subdomain themselves, possibly pooling with people they already have a "social relationship" with
@ln4711 The goal here in part is to allow people to make use of shared infrastructure, without losing control over their identity.
So the target audience is largely non-technical.
Geeks can, and do, already do all of this.
@HerraBRE Perhaps allowing #Mastodon to run under custom domains can be good enough to give people control over their identities. You can then either have your private domain point to an arbitrary server or become @me@identity.virtualinstance.tld if virtualinstance.tld wishes to provide custom DNS record for each user.
@KrzysiekJ I have proposed exactly that to the Mastodon team on their account migration issue.
That was about a year ago, I don't think the idea got much love.
It might get more traction if it were easier (less geeky) for users to obtain their own domains for this purpose.
@HerraBRE I was a big proponent of i-names 15 years ago for that very reason https://en.m.wikipedia.org/wiki/I-name
You could tell your i-broker to send Mastodon follow requests to a@example.com one day and xyz@example.net another, completely transparent to anyone using it as long as the right protocols were implemented.
@daveross Yes, there have been a few attempts at this. I'm vaguely of the opinion that we already have DNS, and although it's flawed it's not going away any time soon...
DNS is already the identity layer of the web. E-mail, XMPP, the web, and the Fediverse - all rely on it.
We could improve things a lot without inventing new protocols, it's low hanging fruit we should pick.
@HerraBRE all the dyndns-provider i know rather operate in this way...
but maybe the indieweb people went further with it?
@tethre You're right, the technology largely already exists. The NGO might not do much tech at all, maybe it would partner with existing providers and just explain and advocate.
But trying to get normal humans to understand that a domain (or subdomain) is a way to own their identity, and then getting service providers and software devs to respect that is still a big job that needs doing.
@clacke @cwebber There is a chicken and egg problem here; supporting migration isn't considered important because migration is hard and the because the identity isn't portable anyway.
Making the identity part portable might help break that cycle.
Of course, there are also politics. People will sugar-coat it with talk about community and moderation, but the fact is a lot of people like the fact that user's cannot easily move from one instance to another.
That's a feature, if you like power.
Anyway, anyone adopting your suggestion would have to implement both. I'm just saying there are several puzzle pieces missing.
Export is available for any OStatus account today, you just scrape the whole feed and there are scripts for that. So it's the import that's missing. I have several old accounts backed up that I would love to pour into some instance some day. Maybe I'll get around to writing or improving one some year.
> supporting migration isn't considered important because migration is hard and the because the identity isn't portable anyway.
Who said migration wasn't important? Is anyone claiming that?
Migration tools are some of the most commonly written tools written for the client-server APIs of fediverse servers..
@cwebber @clacke Hmm, here's today's take from a GNU social developer: https://social.umeahackerspace.se/notice/1872997
The Mastodon issue has been open for a year and a half. It's marked as high priority, but... oh well. I'm impatient. https://github.com/tootsuite/mastodon/issues/177
I'm happy to see a comment there that they now consider this the highest priority issue. The issue was open for a full year before they made that choice, but I shouldn't be looking the gift horse in the mouth, should I?
¹ https://xmpp.org/extensions/xep-0283.html
@kaniini @clacke @cwebber Oh. Wow, that's complicated!
How does that handle adversarial admins or instance-death?
I mean, one of the reasons I want to move might be that I just broke up with the admin and we're not on speaking terms. Or you know, any number of other messy human situations. How does this handle witches.town going permanently offline?
Is there a doc somewhere I can read instead of bugging you here?
... I'm busy, but if anyone else is interested in working on this, I'd make time to advise and help out.
I've been sitting on this idea for years and I just have too much else going on to make it happen on my own.
Any ideas on how this could move forward? Who to work with? @cwebber ? @fsfe ?