Here's something to think about. As you may recall, I have an hourly art feed, available both in RSS and on Bluesky. Same content.
Today I gave each the same new name, art.feediverse.org.
If you use that URL on Bluesky, you go to its version of the content stream. And if you give the same URL to a web browser, ie accessing it via HTTP, you get the RSS version of the same data.
BTW -- think of feediverse as a web of feeds. It's open of course, because the web itself is open. And it's pretty social as you will see. It's why I think it's really the first instance of the "open social web."
I've long felt documents should have their own CNAMEs. I even went so far as to design a system where each node in an outline could have a CNAME. I still believe in this, why not have a hierarchic file system that's as editable as an outline? It's the same idea that works in so many other contexts. It's why the idea of a scripting system on a graphic PC like the early Mac made a world of sense. It's why Unix underneath that same UI was the right way to go. And why the web is so nice, you just View Source and you get to lift the hood and see how it works.
They were smart at Bluesky to use DNS this way. Why invent your own identity system when the net itself has a great distributed system that scales? They just made the wrong call (imho) in not using RSS for their timeline data, or at least the outside world's view of their timeline, in both directions -- in and out. That's the basic idea of the Internet itself. No one cares what's behind your TCP interface, it could be a network with a million nodes using some wacky protocol to connect them. As long as I can get to you via TCP, you're on the net. This is the kumbaya philosophy of the net, why not just use what already works.
You'll see that we're able to do some interesting twitter-like stuff with this duality.
A thread for comments and questions.
PS: A piece about DNS as ID in 2012.