How to Kill a Decentralised Network (such as the Fediverse) écrit par Ploum, Lionel Dricot, ingénieur, écrivain de science-fiction, développeur de logiciels libres.
A warning and a perspective from an insider who has been through this before.
I’ve never really understood the EEE argument here. XMPP was an open proptocol, Google embraced it and attracted users, then extended it and took those users away. But according to this article, Google didn’t extinguish XMPP. It’s still around and serving its niche community.
That’s already the situation the fediverse is in. This is a niche community and there are already existing social media companies that the majority of internet users are on. If Facebook joins the fediverse, it brings billions of new users to the fediverse. If they then leave the fediverse, ActivityPub will still be here and all of us on the real fediverse will still be here, in a niche community. Everyone here has already chosen the fediverse despite it being a clunky, unpolished, niche network. How is EEE a relevant fear for the fediverse?
When a large company takes an open protocol, embraces it using adding users to the network through heir platform, then extends it using proprietary means, they have full control over how the protocol runs in the network.
When the open standards are forced to make changes to be functional with the dominant proprietary app that is poorly (and sometimes incorrectly) documented, open source groups are constantly on the backfoot in order to maintain compatibility, and that makes it harder to compete on their own right.
A second example given is LibreOffice, whose documents are made to fit the XML standard by Microsoft, but there are quirks in their documented standard that if you follow it too closely it isn’t formatted quite the same as the document produced in Microsoft Office, so they were pressured to effectively copy MS and deviate from the standards MS claims to follow.
Well, isn’t that sort of mentioned in the article?
If fediverse development slows down e.g. because adoption of inofficial Facebook extensions takes time it will harm the whole platform. Not by directly taking away users but by blocking progress.
I don’t think the Fediverse is small enough for this to be a serious concern. Especially once multiple companies (Tumblr?) are invested in the fediverse I don’t see this happening anymore.
Yes I read that and explained why I don’t think its relevant. Facebook can’t slow down progress on the fediverse because:
progress is already slow. The fediverse has been in development for 15 years and still is a clunky, niche network and likely will always be less polished than large corporate networks.
Every developer on the fediverse is aware of the EEE playbook and next to none of them will try to remain compatible with any corporate extensions.
What do you mean? Progress is already slow so any additional slow down will seriously harm the fediverse precisely because of the limited resources IMHO.
I’m not quite as optimistic as you but yeah, I don’t think it will be easy for Facebook and if they misjudge it they will end up making a competitor stronger by bringing more attention to it.
But #1 is predicated on #2. If developers are aware of the risk of EEE, then they won’t try to remain compatible with Meta extensions, which means development of the open AP ecosystem will continue at the same pace.
I’ve never really understood the EEE argument here. XMPP was an open proptocol, Google embraced it and attracted users, then extended it and took those users away. But according to this article, Google didn’t extinguish XMPP. It’s still around and serving its niche community.
That’s already the situation the fediverse is in. This is a niche community and there are already existing social media companies that the majority of internet users are on. If Facebook joins the fediverse, it brings billions of new users to the fediverse. If they then leave the fediverse, ActivityPub will still be here and all of us on the real fediverse will still be here, in a niche community. Everyone here has already chosen the fediverse despite it being a clunky, unpolished, niche network. How is EEE a relevant fear for the fediverse?
It’s in the article but to paraphrase it:
When a large company takes an open protocol, embraces it using adding users to the network through heir platform, then extends it using proprietary means, they have full control over how the protocol runs in the network.
When the open standards are forced to make changes to be functional with the dominant proprietary app that is poorly (and sometimes incorrectly) documented, open source groups are constantly on the backfoot in order to maintain compatibility, and that makes it harder to compete on their own right.
A second example given is LibreOffice, whose documents are made to fit the XML standard by Microsoft, but there are quirks in their documented standard that if you follow it too closely it isn’t formatted quite the same as the document produced in Microsoft Office, so they were pressured to effectively copy MS and deviate from the standards MS claims to follow.
Ironically XMPP is a counterexample to your argument. They made the switch to mandatory TLS even though GChat didn’t.
That’s a neat fact!
Well, isn’t that sort of mentioned in the article?
If fediverse development slows down e.g. because adoption of inofficial Facebook extensions takes time it will harm the whole platform. Not by directly taking away users but by blocking progress.
I don’t think the Fediverse is small enough for this to be a serious concern. Especially once multiple companies (Tumblr?) are invested in the fediverse I don’t see this happening anymore.
Yes I read that and explained why I don’t think its relevant. Facebook can’t slow down progress on the fediverse because:
But #1 is predicated on #2. If developers are aware of the risk of EEE, then they won’t try to remain compatible with Meta extensions, which means development of the open AP ecosystem will continue at the same pace.
Okay, sure. Still a bit skeptical about point 2 but we will see.