A decentralised system doesn’t have a flagship instance. Just sayin’.
We’re not an alternative as long as we adopt their success criteria. Let’s do better.
⇧
A decentralised system doesn’t have a flagship instance. Just sayin’.
We’re not an alternative as long as we adopt their success criteria. Let’s do better.
jack
in reply to Aral Balkan • • •Eugenus Optimus 🇺🇦
in reply to Aral Balkan • • •as long as the instance list is not randomized in the Mastodon app and large instances do not impose a cap on the number of users, there are technically flagship instances
If any of them were to be banned by a government, it would significantly harm the Fediverse
Claus Cramon Houmann
in reply to Aral Balkan • • •CynAq🤘
in reply to Aral Balkan • • •Scan le Gentil
in reply to Aral Balkan • • •Hugs4friends ♾🇺🇦 🇵🇸😷
in reply to Aral Balkan • • •This too, shall pass.
The Sleight Doctor 🃏
in reply to Aral Balkan • • •I think the argument goes that because picking a server is so confusing, newcomers need a flagship server for easier on-boarding. They can then migrate when they're more familiar with the fedi.
The trouble with this argument, is that we can't typically import content archives when we migrate, so folks tend to stay on the flagship server. Which then gets bigger and bigger, undermining decentralization.
neatchee
in reply to Aral Balkan • • •Hard disagree. This is like saying email isn't decentralized because Gmail exists. Decentralized =/= homogeneously distributed.
If .social disappeared overnight my server's federated timeline would hardly change much. And another large server would quickly step in to fill their role as "newbies' first instance".
IMO it's all about any server being replaceable, even the largest "flagship" instance.
Aral Balkan
in reply to neatchee • • •neatchee
in reply to Aral Balkan • • •uniform power distribution is an ideal to be sought, not a litmus test for decentralization.
Social dynamics inherently preclude uniform power distribution in any system that is predicated on human interaction.
The only solution to the "problem" you've defined is programmatic enforcement of user distribution which reduces individual user choice and undermines principles of self-selection that are fundamental to the advantages of a federated platform.
Don't let the perfect be the enemy of the good.
jstogdill 🙊
in reply to Aral Balkan • • •Aral Balkan
in reply to jstogdill 🙊 • • •There’s a difference between designing a system where a single node can support 1 person or 1 million people and one where a single node supports only 1 person.
The former will always tends towards centralisation.
The latter maybe has a chance at remaining decentralised based on how easy it is to own a node.
jstogdill 🙊
in reply to Aral Balkan • • •Aral Balkan
in reply to jstogdill 🙊 • • •jstogdill 🙊
in reply to Aral Balkan • • •jstogdill 🙊
in reply to jstogdill 🙊 • • •I’m not completely clear on what you are working on, but as soon as I saw “dns sub domains” in the description I saw 🚩’s
Who ever owns those top domains holds a position at the top of a hierarchy.
Aral Balkan
in reply to jstogdill 🙊 • • •@jstogdill If you ever have some time, here’s a talk I gave last year (the tools have come along a bit since then) but the first ten minutes or so is probably a good summary.
ar.al/2024/06/24/small-web-com…
Basically, you’re right, they could easily be, which is why Domain is designed not to scale, why I’m spending so much time/effort to ensure anyone run their own, why you can easily point any other domain at it after the initial setup and easily migrate to a different server, and why it’s all free software, etc. Not saying some asshat won’t find some bloody way of centralising/sabotaging it but trying my darnest to ensure it’s not in any way interesting for those folks. 🤞
Small Web: computer science colloquium at University of Groningen
Aral Balkan