NIP-05 is weird... It's supposed to be a mapping to a DNS identifier, so it sounds logical to use DNS for that - but no, you have to use HTTP/S and a well-known file. I guess I understand that it's more universal through well-known, but it still seems weird to me

Or am I missing something?

Reply to this note

Please Login to reply.

Discussion

The design allows for multiple user names on the same domain. You wouldn’t want to have to update DNS every time you add or remove a user.

Well, the fact that you can have multiple users on the same domain is also strange. We have subdomains for subnames. If each user had their own subdomain, it could be used as an identifier in reverse - now you need some extra username for that.... guess I'm too influenced by AT Protocol, where domain is unique identifier that and you can use both DNS and HTTP methods as you like.

It seems to me that this is moving towards Activity Pub and creating gatekeepers rather than if the identifier were a domain or subdomain, which would perhaps push more towards self-ownership. I'm not saying it's outright wrong, it just feels weird.

Is there a way to validate DNS txt records from the browser?

On the other hand, its much lower barrier of entry if you have your own domain to upload a file instead of mocking with the dns entry.

In browser, using DoH (DNS over HTTPS) is pretty common nowadays. Yes its gateway, but at least you do not need to reveal your IP with direct requests to every domain

yeah smells like a bad decision

You aren't missing anything, web developers are just not familiar with DNS so unless your API is JSON bullshit, you are facing headwinds with web devs. So you won't get as much cool hackathon demos. It might be smart to use the worse tech to acquire Devs... It is depressing though when you get neither.