This blog post by Ploum, who was part of the original XMPP efforts long ago, describes how Google killed one great federated service, which shows why the Fediverse must not give Meta the chance
This blog post by Ploum, who was part of the original XMPP efforts long ago, describes how Google killed one great federated service, which shows why the Fediverse must not give Meta the chance
The case with XMPP is that Google Talk introduced addons and intricacies that were unique to them. So they could federate with you in full with additional bells and whistles while you were stuck in an eternal catch-up. They presented a better alternative regardless of the eventual defederation. Even if we have some viral clauses as in GPL in open-source software that ensures protocol compatible software to be compliant, we can only do that to a certain extent plus enforcement is always an issue. Who are going to spend the vast sum of money in court to defend the “federation”?
This aside, enforcing federation alone does not ensure decentralization. These zero-marginal-cost fixed-cost-intensive businesses of the internet has a tendency to centralize as serving one more seat costs no penny plus one more seat diluates the fixed cost altogether.
Your points are valid, but that doesn’t mean we should do nothing. Enforcing federation and using copyleft licensing are both strong defenses against centralization and network dominance by a well funded third party.
As far as GPL goes, from what I’ve seen, big tech companies tend to take it pretty seriously. There is no reason we shouldn’t be using that, and other license protections if we have the option.
As for natural centralization over time, I think that is a far less urgent problem than the current risks we are facing, those being major network fragmentation due to the use of defederation, and the risk of centralization around a proprietary platform and/or instance.
Removal of defederation and strong copyleft licensing seem to be natural first steps in combatting that risk.