(1/2)

"And even our terms 'Server' Or 'Instance' make sense in an engineering flowchart but why in all that is holy would we foist those onto users to pretend to understand? At the very least we should talk about new folks joining a 'server community' of fellow users."

@tchambers, 2025

timothychambers.net/2025/06/18

So much for "server" making more sense to newbies than "instance". I followed the herd on this usage, even though I questioned the wisdom, given how overloaded the term "server" is.

(2/2)

I'm hesitant to add yet another meaning to "community" for the same reason. Also because not all instances *are* communities, in any meaningful sense of the word.

What about just "service"? Eg example.social is a service connected to the fediverse. Just like it's a service connected to the net, and presuming it has an https portal, it's a service connected to the web.

I'll admit "service" has a lot of existing meanings too, but most of them apply in this usage.

(1/2)

"Worse? There’s no encryption [on DMs]. So it’s not just accidentally public—it’s intentionally insecure. It’s plaintext dressed up as a secret. There is some fine print warning you, But let’s be real: nobody reads fine print mid-conversation."

@tchambers, 2025

https://www.timothychambers.net/2025/06/18/113327.html

Gargron is squarely to blame for this.

(2/2)

Those of us who've been here since GNU social was the dominant software, remember him shoehorning DMs into OStatus so Mastodon would be more like Titter. Without even making sure they wouldn't display publicly in other software. Spoilers: they did.

So the other projects had to scramble to reverse engineer Mastodon's posting scopes to avoid that. As did AP editors. Even though it would have been better to hold off on DMs, and work towards a protocol-level consensus on doing them right.

Just some idle thoughts @tchambers. I love that you've put all this in one place, I love the tongue-in-cheek way you explain it all, and I'm really excited to see part II, with the potential solutions you have to offer.

I'm especially hoping to see some love for the FEP process, given the way you seem to casually dismiss it towards the end of the piece. Some of these UX fixes do need protocol-level support, and FEPs are a community-driven process for filling AP potholes.