I think the problem is there are two ways of using Mastodon. Some people use it like Twitter. The fact that it’s broken up into individual servers in a federated network is just an implementation detail. Blocking aside, you can talk to anyone, and anyone can talk to you.
Other people use it like SSO for a bunch of PHPBB bulletin boards. Finding out that a bunch of rando Facebook users are going to be able to start posting on your gay motorcycle enthusiasts board is very annoying! You don’t want more people on your board; you just want your friends on your board! The fact that it’s all on the open web and they could always just read it anyway was just an implementation detail.
Mastodon is not, but there are other things. Kbin and Lemmy on Fediverse. Even Lobsters (codebase) itself is self-hostable, isn’t it? Some services already have groups, some are working on a similar feature. So it should be easy enough to convince that one friend, get federation AND forum-like closed group.
But yeah, it’s still not ideal. When someone just offers you a free (or “free”, doesn’t matter) service, that’s gonna be enough.
You touch on the key point: is decentralization a feature, or a detail?
For the builders, it’s a core feature. But for almost all users, it’s certainly a detail. This disconnect is pervasive in the decentralized web, and is the main reason decentralized applications aren’t really gaining traction.
I don’t remember these events playing out quite the way the author does.
I wrote the server-side chat interfaces for the Sidekick and I can say with confidence that MSN was, uh, how to say this nicely… by far the least popular of the big three (AIM, YIM, MSN). Jabber (now XMPP) was a distant fourth – we actually implemented it, but there was no product interest so I don’t think we even shipped it. I would say that XMPP had the worst protocol (they tried to invent a kind of “streaming XML”) but MSN raised the stakes by trying a kind of “binary SOAP” that’s best left unmentioned.
Google back then was not known for doing anything social or popular, so I think their use of XMPP was an attempt to get a handful of free users and good will from techies. I don’t think that initial team had bad intentions, but a company like Google has massive turnover, and their chat system didn’t really get any traction until AIM was starting to die.
I wrote the server-side chat interfaces for the Sidekick and I can say with confidence that MSN was, uh, how to say this nicely… by far the least popular of the big three
This was highly geographically specific. It was by far the most popular in the UK, where AIM had very little usage and no one even knew Yahoo! had a messenger. My geek friends used ICQ then moved to XMPP, everyone else used the MSN Messenger account that they got for free with their Hotmail account (and everyone had a Hotmail account. I remember being grumpy with people saying ‘what’s your hotmail address?’ meaning ‘what’s your personal email address?’).
I only ever had contacts on MSN, so I think it just depended on your social group. Google was flying high off the launch of Gmail so they had a huge amount of Goodwill and I think the kind of company they were at the time genuinely wanted to do the right thing. They just got bored over time since no one else followed suit
This. Facebook used XMPP as well as it was an easy way to get started, but the thing is that evolution of the XMPP was too slow and too chaotic for the “big tech”. They wanted “new shiny features” but whole process was so chaotic and hard to go through, that they stopped giving a damn. WhatsApp is also XMPP (just heavily modified that it is mostly not interoperable with other nodes).
So it is not Google that killed XMPP IMHO, it is XMPP that severed their relation with the “big corp”.
There’s a social dynamic here that’s not new, and has been around for decades as other platforms and protocols have gone through the same thing.
Basically there’s an unstable point where the thing can either go big – explode in popularity and become a general-appeal place where all the normies come to hang out – or go bust and forever remain a collection of niche communities withering away from attrition.
And there are always people who think the “go bust” side is the preferable one, because they’ve been there for a while and have what they think is a comfortable and familiar community that needs to be defended against the onslaught of the barbarians. They hope that by beating back the normies, they can preserve their community, and the special way it feels, unchanged.
(of course, that’s a fool’s errand. Preserving a community unchanged is impossible, and every community sooner or later starts to alienate its old-timers and hard-liners just through normal evolution)
But that’s where “the Fediverse” is right now. There are quite a few people – many of them instance admins – who feel like they have a nice, cozy, comfortable community that needs to be preserved against the ravening hordes who might show up and change things. And they’ve been fighting since at least late last year when people started migrating over from Twitter. All the angst and arguing and threats of defederation and legal action and whatnot over things like search and discoverability features are part of that – all of those things would change the nature of “the Fediverse”.
And again that’s really what the trouble is all about: change. No matter someone’s political views, this dynamic is deeply rooted in conservative feelings and in seeing change as an inherently bad thing, a thing to be fought and resisted and pushed back. Which leads to a linking of “change” with death – if more normies show up and start turning this into a general-purpose social network, then the cozy thing the early adopters had will “die” in the sense that it will either change into something else, or refuse to change and simply disappear as a result.
All of that can happen, and is potentially happening right now, regardless of whether any particular big evil corporation stands up an ActivityPub instance. It doesn’t take an outside threat embracing and extending and extinguishing to kill “the Fediverse”. “The Fediverse” – not the protocols or the technologies, but the people who see themselves as a community of that name, and as guardians and protectors of it – does not need help to be “killed”. It is perfectly capable of dying by its own hand.
“The normies” will kill the fediverse network? Nope. That’s not what this article is about. The article says that Meta will kill the network. The normies won’t even know they’ve killed anyone. The fediverse admins that are “conservative and afraid”, like kev, are not against this because of fear of change - it’s because “Meta’s moral compass and their own” are far from aligned.
It may all end up the same - fediverse half-dead and niche (not that it’s mainstream even now) - but the article disagrees (and I do too, for that matter) on the reasons why. I’m not a fedi admin (I just self-host an instance for myself). And I don’t care if nobody ever reads my statuses. But the “Embrace, extend, extinguish” tactic has a large chance of just overwhelming gotosocial devs (the software I’m using) so that they’ll give up. So I don’t mind the changes that the GTS devs bring. I do mind some changes that mastodon devs are bringing, but I don’t mind enough, I can live with it. What Meta will do is simply kill off the protocol and then there won’t BE a protocol or software to self-host. Or change. Or anything.
“The normies” will kill the fediverse network? Nope. That’s not what this article is about.
I never said it was about that.
I said that “the Fediverse” doesn’t need an external technical factor to kill it, because normal human social dynamics are perfectly capable of doing that, and some of those dynamics are already on display and have been for a while.
I disagree with most of what was written in this post; According to what was written, Google “killing” XMPP was enabled by XMPP’s developers slowing their pace to suit Google. Had those developers not hobbled themselves to accommodate Google, they would have come out having accomplished more than they did. Handled correctly, Google’s involvement in the ecosystem should have been a fluctuation in user count, and nothing more.
If Meta decides to execute the same maneuver and ActivityPub implodes as a result, it will be the fault of developers prioritizing working with Meta. In this case, I find that somewhat unlikely, because most of the people I know that participate in the Fediverse do so out of hatred for Meta/Twitter/etc.. Meta will be regarded as just-another-end-user at best.
The XMPP developers were swooned by a corporate adopting their protocol, that corporate then had the largest user base and any developer should focus on their largest user base. The corporate then ruined the protocol.
The moral of the story is for the fediverse not to federate with corporations. Worldwide ActivityPub access is not for corporates, because they will ruin it, like Google ruined XMPP, and like the corporates have enshittified their own networks and driven users away to what the fediverse offers.
The dream of the fediverse has always been to get the big entities to join. To refuse to federate with them it cutting off your nose to spite your face.
I think you’re mistaken. I get the 100% opposite impression. The fediverse would like lots of participants (humans), but has a pretty strong antibody reaction to corporations, even the one that writes the most popular fedi server.
To get a large userbase, you need lots of people to be able to join. Even on something like lobste.rs, with a very high percentage of geeks as a proportion of the total population, the number that self host anything is (I would guess from prior threads) under 50%. If you need everyone to self host, a system is obviously dead, the question to me is what proportion of people need to run servers. I would guess that 1% is probably too high, perhaps 0.1%. That’s a problem because most people have fewer than 1,000 friends. Hosting a server for your friends to use has obvious rewards (your friends are happy). Hosting a server for 1,000 people, most of whom you’ve never met, is far harder to motivate. You need some mechanism to persuade people to run servers for 1,000-10,000 people for systems like this to scale. A very small number of people will do it because they believe in federated systems as a public good, but that’s well under 0.1% of the population (my guess would be a few thousand people worldwide). Without that, there are two mechanisms that have worked in the past for running large-scale infrastructure:
Public funding
Profit
Public funding may work in some cases. I can imagine having governments fund instances for government employees to use. Some of these may hit the kinds of economies of scale that means that they may as well make them available to the general public as well (though that then has censorship implications).
Without that, we’re left with profit: people running the services to make money. Personally, I’d love to see something like Mastodon include a monetisation model. When WhatsApp launched, it was free for the first year and $1/year after that. At that price, they were making quite a lot of money (WhatsApp’s Erlang-on-FreeBSD infrastructure was impressive - they had hundreds of thousands of active users on a single machine and could scale to over a billion users) but they were also so cheap that it was totally in the noise for most people. If running a Mastodon server for 1,000 users brought in $1,000/year, then that’s more than sufficient to cover hosting costs and so at least you’re volunteering only your time, not time and money.
I can see it having value, because the yearly fee should ensure the continued existence of the server, but you’re free to interact with the rest of the fediverse using your account there.
I think it’s important to distinguish member funded vs public funded. Public funding implies allocation of a fraction of resources of the whole state to a particular purposes. Member funded implies those involved fund their own operations.
There are certainly people who feel this way, but it’s very misleading to frame it as “the dream of the fediverse” when it’s really just a sub-contingent.
The rest of us are just happy to have a place where we can hang out with our friends without getting sucked into corporate social hellscape. Not every garden has to grow to cover the face of the entire earth to be a success.
Join on what terms though? It’s entirely possible for every ActivityPub user to join Facebook already: it’s called creating a Facebook account. Clearly that’s not acceptable terms, because people aren’t doing it.
So, what exactly does the term “join” mean here, if not the community having control over what features are included in the project going forwards? And if there’s ever a disagreement with Facebook, how can ActivityPub users actually push back against Facebook in any way except by defederating?
The dream of the fediverse is, if anything, to replace Facebook entirely. The platform is nice, but the corporate entity behind it is driven by profits, and there is nothing ActivityPub can do to force Facebook to adopt different values.
Thu benefit of federation is you don’t really need terms in the same was as corporate integration deals. To make a Facebook account you need to agree to Facebook ToS etc, but for Facebook to become part of the fediverse they just need to support certain standard protocols on their endpoints and advertise them. Then a Facebook user can follow any ActivityPub user or vice versa, and neither Facebook nor the ActivityPub admin need to even be aware that this is possible for it to happen.
Would it be nice if this gave more people the freedom to leave Facebook because they are no longer locked in to its network effect, since they can still communicate with their Facebook using contacts from elsewhere? Yes. That’s awesome. It’s also the reason no one ever believed we’d see Facebook or Twitter seriously considering federation, but if they can be convinced to see past this short term loss of control I think it would be good for them in the long term as well, and maybe they do see that since they are thinking about it.
To make a Facebook account you need to agree to Facebook ToS etc, but for Facebook to become part of the fediverse they just need to support certain standard protocols on their endpoints and advertise them.
Will Facebook then have some data on me? That I don’t want them to have and hence self-host a fedi node for myself? I think that’s kinda the point of profits and values.
No, I’ve meant kinda in network. When you, say, search for tags or something. I know that the distinction is mostly irrelevant, but the point is, I’d be willing to give my tag search preferences to another fedi admin, but not to a corporation. But if the corporation is networked, they’ll get that data.
And yeah, by “a corporation”, I mean specifically meta. I left Facebook years ago because of that data that I did not want to give them.
If that point is correct, then there should exist at least one XMPP extension that was widely implemented except by Google, and its fate discouraged XEP authors/developers, leading to that slowdown. Right? I don’t recall any such extension. Did I miss any?
I know what Google did and did not implement, and I know the uptime problems. I don’t know that it led to a slowdown in XEPs being written, a slowdown in XEPs being implemented in either servers or clients, or anything else I’d call XMPP’s developers slowing their pace to suit Google, and that absent arguments, blaming google is a just propter hoc to deflect from XMPP’s homemade problems.
FWIW, I liked XMPP, wanted it to succeed, even had an XMPP address on my business card for years. It seems likely that I’d hear about such problems, even if not certain.
According to what was written, Google “killing” XMPP was enabled by XMPP’s developers slowing their pace to suit Google
I think that’s quite an interesting interpretation of what happened. My recollection (I was involved in XMPP before and a bit after it became an IETF standard and wrote a couple of clients) it was dying before Google got involved. The standards process ran ahead with complex designs that were 2-3 years ahead of what anyone could implement. There was a push for pub-sub (with no implementations) then everything had to be built on top of pub-sub. Personal Eventing over Pub-sub (PEP) provided a generic mechanism for doing event notification things. Nothing could be standardised if it could be built on top of PEP but wasn’t, but PEP depended on server support and client support, neither of which existed in the popular implementations. Instead, everyone pushed their own extensions that weren’t interoperable. There were multiple different ways of doing file transfer and even of sharing avatars, most of which worked in 1-2 clients.
Google came along and dumped the Jingle (voice / video chat, primarily) specs on everyone. They were overengineered but at least they were well supported by the client that suddenly had as many users as all of the other XMPP clients combined. Everyone else struggled to catch up with Google Talk’s Jingle implementation. Other things were layered on Jingle (it could establish streams, so let’s use it for file transfer and throw away the other 10 ways of doing file transfer that had XEPs), but Jingle was so big that many things that did X-over-Jingle didn’t actually interoperate because Jingle they each supported disjoint sets of Jingle transports.
The whole thing was a mess. It needed a strict IETF-style rule that you can’t propose a standard without one implementation and you can’t advance it without two independent interoperable ones. It also needed a permissively licensed reference implementation of a client library and a blessed server (there was a lot of drama here too: the JSF produced a reference server, then decided to rearchitect it completely in a way that broke plugins, then decided to endorse a different one, then another different one).
There are a lot of lessons to learn from XMPP, but ‘don’t let corporations get involved’ is not the right one.
It’s also a weird thing to argue because XMPP was not hurt by Google leaving except that all those users stopped being able to federate. The rest of the ecosystem kept going just as it always had and still is today
If Meta decides to execute the same maneuver and ActivityPub implodes as a result, it will be the fault of developers prioritizing working with Meta. In this case, I find that somewhat unlikely, because most of the people I know that participate in the Fediverse do so out of hatred for Meta/Twitter/etc.. Meta will be regarded as just-another-end-user at best.
For what it’s worth, the B?DFL of Mastodon (Eugen Rochko/Gargron) seems indifferent to positive about Meta’s involvement - at least, he seems to have agreed to speak to them under NDA instead of laughing them out of the room* - so I wouldn’t be so sure.
Whether or not Facebook will be able to “kill” the Fediverse,
I think it will definitely try to,
and that’s reason enough for me to be concerned.
Besides that, let’s remember: its business model is collecting data and using it to market products and ideas to users.
I don’t want that for myself or my userbase,
so I’ve signed the fedipact
and proactively blocked threads.net.
MTL Rocks will not federate with Facebook,
and I encourage other Fediverse admins to do the same.
For Facebook to maintain control of all eyes and attention which equates to ad revenue, users need to stay on and interact with Facebook and/or look at Facebook ads. Allowing unrelated entities on the internet to have access to their walled garden whose whole intention is to produce a network effect and keep people goes against their financial incentive. It just doesn’t make any sense. And any time something doesn’t make sense, it likely points to missing information. The hypothesis that Facebook wants to kill the Fediverse fills that gap in ideas: Having learned from Twitter’s exodus, by embracing and then killing the Fediverse off, Facebook will be able to better control and maintain its position. Now it all makes sense.
That’s not the only hypothesis that makes sense. I’m sure there are other perspectives. But putting together the timing with Twitter’s exodus, attitude of Meta toward its users, history of exploitation, and financial incentive to exploit some more, it’s hard to think of an alternative viable explanation for Facebook’s decision.
The Fedi user base is a rounding error for FB. Mastodon.Social has 1M active accounts. FB has 2.85B. That’s not counting Instagram and Whatsapp, both of which also have billions of users. The idea that FB would actively[1] crush the Fediverse is ludicrous. Most people who are active there hate FB anyway, but they are mostly preaching to the converted. It’s fun seeing yourself as the rebel alliance fighting the empire, but the empire probably doesn’t even know you exist.
My take - someone inside FB really likes the idea of ActivityPub, and has enough clout to get the company interested in it, at least in an exploratory way. I’m betting they’ll use an in-house solution for Threads in the end.
[1] I’m not saying there can be inadvertent damage though.
Facebook is not an indie passion project run by a couple of people. It’s a multi billion dollar advertisement company whose sole purpose to increase revenue at whatever cost to others. It’s not plausible that some dev somewhere likes ActivityPub and all of a sudden a whole team is engaging in “private meetings” with admins which include “confidential information”. That’s naive at best, and malicious at worst.
Obviously “The Empire” knows you exist and is threatened by you enough to take action and allocate resources to fight, so that point is invalid.
I also don’t buy the “inadvertent” here. The executives at companies that generate billions are way too smart for “inadvertent”.
all of a sudden a whole team is engaging in “private meetings” with admins which include “confidential information”.
This is entirely plausible, if the meetings are “if we build this, your server will get an estimated N million visits a month. Can it handle that?” That estimate is privileged information because it reveals details of how much traffic Meta is estimating .
Obviously “The Empire” knows you exist and is threatened by you enough to take action and allocate resources to fight, so that point is invalid.
LOL no, they’re trying to build a Twitter clone and are investigating whether AP is a good fit. I suspect it’s not, they’ll roll out their own solution, and the angst will begin about how AP is a “dying protocol” because it wasn’t chosen by Meta.
Why would traffic estimate be so confidential that admins are put under NDA and prevented legally from discussing it? What part of that is confidential? Makes no sense.
If they want a Twitter clone, they can unilaterally implement it. No need to get involved with multiple ActivityPub instances. Again, there is no money in that, no incentive. Makes no sense.
The question to ask is: How can federating with these public instances generate advertising revenue for Meta and put it in a more dominant position? I’m not smart enough to answer that. But these arguments just fall apart like meat on a bone that’s been cooking slowly over 24h. Mmmm I’m getting hungry.
I just can’t see Meta being a good webizen. Fool me once, twice, three time, seriously, that’s enough. So unless everyone who works there had some sort of ethical epiphany and are now singing kumbaya by the fire pit thinking about world healing, it’s unlikely this is a benign venture.
Serious question: why do you think FB will try to kill the Fediverse?
Facebook is a for-profit company, why the hell would they want to open up their platform? And if they’re not interested in opening up their platform, then 1) any relationship is doomed from the start anyway, and 2) their obvious motivation is EEE.
That’s not “killing the Fediverse”, it’s at most setting up obstacles to full federation from other AP clients and servers other than FB’s. Basically what the FB haters want to happen anyway.
The late 90s called, it wants its conspiracy theories back.
You can run Linux fully supported within Windows.
The most popular code editor in the world is open source, and backed by Microsoft.
Open source is good business for Microsoft now, there’s no need to “extinguish” it, because it’s making MS lots and lots of money.
Anyway, Microsoft and Meta are 2 different companies, run by different people, with different main revenue sources. Why one should follow the deprecated playbook of another, to counter a threat that’s insignificant, is beyond my limited understanding.
The most popular code editor in the world is open source, and backed by Microsoft. […] Open source is good business for Microsoft now, there’s no need to “extinguish” it, because it’s making MS lots and lots of money.
It’s a neutral observation. The MS in the 90s under Steve Ballmer is a different company than the one today. They lost an entire generation of programmers to Linux and open source languages and had to adapt.
In the context of an argument that “you don’t have to worry about a big company coming in and bringing their values into your existing community” it doesn’t feel neutral.
They wanted control; their original method of getting it failed, and then they found a more effective way to bypass the community’s resistance, but in the end they still got what they wanted.
Having not really been paying attention at the time to the various XMPP servers and whatnot, was there the same sort of issue around blacklisting and petty politics that seems to plague the Fediverse?
I’ve avoided getting involved in that scene in large part due to the outsize amount of “we don’t peer with so-and-so”, “omg you peer with this site, well one person on that site said problematic thing so we don’t peer with them anymore”, etc. That would seem to be a bigger threat than Meta.
I dunno, the network has been growing pretty well for 6 or 7 years at this point and while there are occasional issues with admin drama, it hasn’t yet ended up causing a mass exodus, and I don’t think it will. The closest it got was the snouts.online incident, but like everything with OStatus/ActivityPub, that just lead to there being a little island of people who only talk to each other. There similarly is a “sexualized images depicting minors” island, a “Gab/alt-right” island, and a few intentional islands, but that seems like a success of the federated model to me! Those people have their social spaces, and their presence doesn’t cause issues - whether legal or social - for anyone else.
I don’t necessarily think that coupling hosting with moderation in this way is optimal, but it does work, and more so the more instances there are.
The petty politics are pretty specific to a corner of the fediverse (mostly around Mastodon) I think because of the kind of people who came there in the last wave and we they came and some pretty big bugs in Mastodon exacerbate it. I’ve not observed this issue on other federated networks like SMTP, XMPP, IndieWeb, etc to date
Having not really been paying attention at the time to the various XMPP servers and whatnot, was there the same sort of issue around blacklisting and petty politics that seems to plague the Fediverse?
No. There weren’t enough people using XMPP off Google for it to be worth targeting with spam. If you wanted to, it was easy (register a new domain name, deploy a server, and you can spam anyone). Since it was point-to-point, not broadcast, there was no need to block servers for any reason other than spam.
If it had become popular, the lack of spam controls would have been a problem. It’s a bit better than email (you at least need a machine with a DNS record pointing at it to be able to send spam and you can accurately attribute the sender) but not much because the cost of setting up a new server is so low. Some folks were starting to think about building distributed reputation systems. For example, you can build a set of servers that you trust from the ones that people on your server send messages to. You can then share with them the set of servers that they trust and allow messages for the first time from servers on that list, but doing that in a privacy-preserving way is non-trivial.
The problem cited with Google and XMPP seems to be an issue with a single “server” having outsized influence over the XMPP fediverse. In this scenario, Google was effectively a single XMPP server that knew it could wield its users as influence over the existing servers.
If the modern day fediverse decide to play ball with Meta, then make federating with them conditional – Meta has to bring one of their competitors to the federation table.
I’m not saying this would necessarily be a good outcome for the fediverse, but if the fediverse want to disarm the cudgel of unilateral action by an outsized party, then bring more outsized parties to the game.
I’m feeling more hopeful about federated messaging given the EU’s recent Digital Markets Act legislation. As was mentioned in your article, Google was able to implement a federate chat protocol once before. They’ll be able to do it again. We’ll see soon how they’ll implement E2EE (if they’re able to that is).
The same thing happened to one of the ideals of OpenID: being able to use any OpenID provider, including self-hosting, to authenticate. Now it’s basically just a small select list for most instances, with the options available dependent on the slice of the internet you’re using. Heck, even OAuth is a bit of a clusterfuck.
OpenID sort of ate its own tail. Similar to how OStatus blew itself up and we entered the federated social dark ages with the complexity of ActivityPub, OpenID2 got blown up by “well OIDC is coming” and then by the time fully federated OIDC was ready everyone had moved on. You can use federated OIDC today same as you did for OpenID2, but I’ve not yet seen any;he who does. Maybe we’ll have our Mastodon moment for that tech yet and bring us back out of the dark ages.
Well, it didn’t really kill anything. For all intents and purposes XMPP is equally popular as it was before Google Talk was a thing, which TFA actually acknowledges. Yes, for a while it was very popular, but there are plenty of spaces on the internet that run XMPP relays and servers. It’s fine.
As expected, no Google user bated an eye. In fact, none of them realised. At worst, some of their contacts became offline.
Well, ackshwally, we stopped using Google Talk all together, because our clients stopped being compatible with Google’s servers and their own client was a joke.
XMPP became niche.
No, XMPP was always niche. Exactly the same way it’s fine if the fediverse IS and eventually will be niche when Meta gets bored or stops being a good neighbour.
Please piss off ringing the alarmist death knell of technologies. You’re not netcraft.
Well, it didn’t really kill anything. For all intents and purposes XMPP is equally popular as it was before Google Talk was a thing, which TFA actually acknowledges.
Well, ackshwally, XMPP went from lots of interested parties complaining about its complexity in adoption but saying “OK, fine” to immediately saying, when GTalk took it on, “oh cool! My slog will be validated” to a big collective sigh when Google dropped it, followed by a bunch of shitting on XMPP: “I knew it wouldn’t work.” “It’s done.” “Good thing we didn’t actually follow through.” “ejabberd was too hard to run, and a pile, anyway!” “Who writes shit in Erlang?” “Eh, I hate XML. No idea why it had to be XML—they screwed themselves with that decision.”
There was lots of hope in the XMPP space before GTalk dropped it. lots of people considering how to incorporate/ build new. Maybe there aren’t any more/ less contributors to implementations or specs now. Maybe there’s more/less the same number of people connecting over XMPP; but something did change. Potential adopters left, and so did any hope among the commons that XMPP was worth fighting for.
You make some good points, but I would still be wary of confusing the cause for the effect. Perhaps that all the interested parties that stopped investing mind share and development time into XMPP, did so because the technology itself was not worthy. Or that it was worth only as long as a big player like Google was also involved.
I’m not sure if you’d agree that the distinction is meaningful, but for me it makes sense.
did so because the technology itself was not worthy.
I don’t think this is true. I think both Facebook and Google had incentives to move faster than the specification could, and the decentralized foothold wasn’t as strong as say, SMTP / email that came before it. If 99.9999% of your users don’t communicate outside your network, why bother?
–
As for why others “lost interest,” put it this way, I think a lot of developers look to large companies, for guidance. They have the ability to hire the best of the best so it naturally follows that the things they do, the papers they publish “must be solid” or else they wouldn’t do it. If XMPP doesn’t work for Google, how could it possibly work for me when I get to Google scale?
Obviously, you (well maybe you are – but you get my point) and I aren’t “Google scale,” and given we’re actively analyzing it we can think a bit more critically. But most people either can’t, or for whatever reason don’t think critically. It’s exactly why the industry is flooded with articles, conference talks, and products selling you on using micro services running on Kubernetes and stream processing, to publish a blog instead of just using WordPress.
Please piss off ringing the alarmist death knell of technologies. You’re not netcraft.
Why this kind of rudeness and hostility? You have a point and the point was well made but then why this kind of impoliteness in the end? I am afraid but I have to protest. I don’t think this kind of behavior towards a member of our own community is ok in this forum!
Granted, out of context my rudeness seems disproportionate, but as someone that develops ActivityPub software, all my social network inputs have been clogged with similar defeatist hot takes.
The points that OP makes in the article are nothing but inflammatory two bit pieces of information that if you squint at them at juuust the right angle, they might look like they spell T H E E N D.
I’d like to see OP put in the work into making the Meta Embrace/Extend/Extinguish game less of a possibility instead of piling on the negativity like a two bit technological oracle. It takes very little effort to talk down other people’s attempts. I am annoyed.
Exactly. It would be more accurate to say XMPP killed google talk. More accurate, but not accurate.
The worst part of reading all these awful internet history hot takes is that I’m pretty sure they’ll all just become accepted history. I guess that’s what all history you live through looks like.
I think the problem is there are two ways of using Mastodon. Some people use it like Twitter. The fact that it’s broken up into individual servers in a federated network is just an implementation detail. Blocking aside, you can talk to anyone, and anyone can talk to you.
Other people use it like SSO for a bunch of PHPBB bulletin boards. Finding out that a bunch of rando Facebook users are going to be able to start posting on your gay motorcycle enthusiasts board is very annoying! You don’t want more people on your board; you just want your friends on your board! The fact that it’s all on the open web and they could always just read it anyway was just an implementation detail.
To be clear I think the second group are looking for OpenID/IndieAuth + Atom/RSS and the whole activitypub thing is a cargo culted accident.
TBF, it’s easier to convince one friend to setup a Mastodon instance than for everyone in a friend group to know enough to self-host.
Sure, but if there were a forum like thing (which Mastodon is not, but lobsters is) with decent SSO and a client API the one friend could run that.
Mastodon is not, but there are other things. Kbin and Lemmy on Fediverse. Even Lobsters (codebase) itself is self-hostable, isn’t it? Some services already have groups, some are working on a similar feature. So it should be easy enough to convince that one friend, get federation AND forum-like closed group.
But yeah, it’s still not ideal. When someone just offers you a free (or “free”, doesn’t matter) service, that’s gonna be enough.
You touch on the key point: is decentralization a feature, or a detail?
For the builders, it’s a core feature. But for almost all users, it’s certainly a detail. This disconnect is pervasive in the decentralized web, and is the main reason decentralized applications aren’t really gaining traction.
I don’t remember these events playing out quite the way the author does.
I wrote the server-side chat interfaces for the Sidekick and I can say with confidence that MSN was, uh, how to say this nicely… by far the least popular of the big three (AIM, YIM, MSN). Jabber (now XMPP) was a distant fourth – we actually implemented it, but there was no product interest so I don’t think we even shipped it. I would say that XMPP had the worst protocol (they tried to invent a kind of “streaming XML”) but MSN raised the stakes by trying a kind of “binary SOAP” that’s best left unmentioned.
Google back then was not known for doing anything social or popular, so I think their use of XMPP was an attempt to get a handful of free users and good will from techies. I don’t think that initial team had bad intentions, but a company like Google has massive turnover, and their chat system didn’t really get any traction until AIM was starting to die.
This was highly geographically specific. It was by far the most popular in the UK, where AIM had very little usage and no one even knew Yahoo! had a messenger. My geek friends used ICQ then moved to XMPP, everyone else used the MSN Messenger account that they got for free with their Hotmail account (and everyone had a Hotmail account. I remember being grumpy with people saying ‘what’s your hotmail address?’ meaning ‘what’s your personal email address?’).
I only ever had contacts on MSN, so I think it just depended on your social group. Google was flying high off the launch of Gmail so they had a huge amount of Goodwill and I think the kind of company they were at the time genuinely wanted to do the right thing. They just got bored over time since no one else followed suit
This. Facebook used XMPP as well as it was an easy way to get started, but the thing is that evolution of the XMPP was too slow and too chaotic for the “big tech”. They wanted “new shiny features” but whole process was so chaotic and hard to go through, that they stopped giving a damn. WhatsApp is also XMPP (just heavily modified that it is mostly not interoperable with other nodes).
So it is not Google that killed XMPP IMHO, it is XMPP that severed their relation with the “big corp”.
There’s a social dynamic here that’s not new, and has been around for decades as other platforms and protocols have gone through the same thing.
Basically there’s an unstable point where the thing can either go big – explode in popularity and become a general-appeal place where all the normies come to hang out – or go bust and forever remain a collection of niche communities withering away from attrition.
And there are always people who think the “go bust” side is the preferable one, because they’ve been there for a while and have what they think is a comfortable and familiar community that needs to be defended against the onslaught of the barbarians. They hope that by beating back the normies, they can preserve their community, and the special way it feels, unchanged.
(of course, that’s a fool’s errand. Preserving a community unchanged is impossible, and every community sooner or later starts to alienate its old-timers and hard-liners just through normal evolution)
But that’s where “the Fediverse” is right now. There are quite a few people – many of them instance admins – who feel like they have a nice, cozy, comfortable community that needs to be preserved against the ravening hordes who might show up and change things. And they’ve been fighting since at least late last year when people started migrating over from Twitter. All the angst and arguing and threats of defederation and legal action and whatnot over things like search and discoverability features are part of that – all of those things would change the nature of “the Fediverse”.
And again that’s really what the trouble is all about: change. No matter someone’s political views, this dynamic is deeply rooted in conservative feelings and in seeing change as an inherently bad thing, a thing to be fought and resisted and pushed back. Which leads to a linking of “change” with death – if more normies show up and start turning this into a general-purpose social network, then the cozy thing the early adopters had will “die” in the sense that it will either change into something else, or refuse to change and simply disappear as a result.
All of that can happen, and is potentially happening right now, regardless of whether any particular big evil corporation stands up an ActivityPub instance. It doesn’t take an outside threat embracing and extending and extinguishing to kill “the Fediverse”. “The Fediverse” – not the protocols or the technologies, but the people who see themselves as a community of that name, and as guardians and protectors of it – does not need help to be “killed”. It is perfectly capable of dying by its own hand.
“The normies” will kill the fediverse network? Nope. That’s not what this article is about. The article says that Meta will kill the network. The normies won’t even know they’ve killed anyone. The fediverse admins that are “conservative and afraid”, like kev, are not against this because of fear of change - it’s because “Meta’s moral compass and their own” are far from aligned.
It may all end up the same - fediverse half-dead and niche (not that it’s mainstream even now) - but the article disagrees (and I do too, for that matter) on the reasons why. I’m not a fedi admin (I just self-host an instance for myself). And I don’t care if nobody ever reads my statuses. But the “Embrace, extend, extinguish” tactic has a large chance of just overwhelming gotosocial devs (the software I’m using) so that they’ll give up. So I don’t mind the changes that the GTS devs bring. I do mind some changes that mastodon devs are bringing, but I don’t mind enough, I can live with it. What Meta will do is simply kill off the protocol and then there won’t BE a protocol or software to self-host. Or change. Or anything.
That’s my problem. Not change.
I never said it was about that.
I said that “the Fediverse” doesn’t need an external technical factor to kill it, because normal human social dynamics are perfectly capable of doing that, and some of those dynamics are already on display and have been for a while.
I disagree with most of what was written in this post; According to what was written, Google “killing” XMPP was enabled by XMPP’s developers slowing their pace to suit Google. Had those developers not hobbled themselves to accommodate Google, they would have come out having accomplished more than they did. Handled correctly, Google’s involvement in the ecosystem should have been a fluctuation in user count, and nothing more.
If Meta decides to execute the same maneuver and ActivityPub implodes as a result, it will be the fault of developers prioritizing working with Meta. In this case, I find that somewhat unlikely, because most of the people I know that participate in the Fediverse do so out of hatred for Meta/Twitter/etc.. Meta will be regarded as just-another-end-user at best.
I feel this is the point of the article.
The XMPP developers were swooned by a corporate adopting their protocol, that corporate then had the largest user base and any developer should focus on their largest user base. The corporate then ruined the protocol.
The moral of the story is for the fediverse not to federate with corporations. Worldwide ActivityPub access is not for corporates, because they will ruin it, like Google ruined XMPP, and like the corporates have enshittified their own networks and driven users away to what the fediverse offers.
The dream of the fediverse has always been to get the big entities to join. To refuse to federate with them it cutting off your nose to spite your face.
I think you’re mistaken. I get the 100% opposite impression. The fediverse would like lots of participants (humans), but has a pretty strong antibody reaction to corporations, even the one that writes the most popular fedi server.
To get a large userbase, you need lots of people to be able to join. Even on something like lobste.rs, with a very high percentage of geeks as a proportion of the total population, the number that self host anything is (I would guess from prior threads) under 50%. If you need everyone to self host, a system is obviously dead, the question to me is what proportion of people need to run servers. I would guess that 1% is probably too high, perhaps 0.1%. That’s a problem because most people have fewer than 1,000 friends. Hosting a server for your friends to use has obvious rewards (your friends are happy). Hosting a server for 1,000 people, most of whom you’ve never met, is far harder to motivate. You need some mechanism to persuade people to run servers for 1,000-10,000 people for systems like this to scale. A very small number of people will do it because they believe in federated systems as a public good, but that’s well under 0.1% of the population (my guess would be a few thousand people worldwide). Without that, there are two mechanisms that have worked in the past for running large-scale infrastructure:
Public funding may work in some cases. I can imagine having governments fund instances for government employees to use. Some of these may hit the kinds of economies of scale that means that they may as well make them available to the general public as well (though that then has censorship implications).
Without that, we’re left with profit: people running the services to make money. Personally, I’d love to see something like Mastodon include a monetisation model. When WhatsApp launched, it was free for the first year and $1/year after that. At that price, they were making quite a lot of money (WhatsApp’s Erlang-on-FreeBSD infrastructure was impressive - they had hundreds of thousands of active users on a single machine and could scale to over a billion users) but they were also so cheap that it was totally in the noise for most people. If running a Mastodon server for 1,000 users brought in $1,000/year, then that’s more than sufficient to cover hosting costs and so at least you’re volunteering only your time, not time and money.
You’re missing the crucial other one: democratic coops supported and governed by their members. This is the best way we know to manage commons today.
Are there any examples of such organizations within the AP space today?
https://cosocial.info/ is one such organisation, run by Evan Prodromou, one of the creators of ActivityPub.
Thanks for the link, this is super interesting.
I can see it having value, because the yearly fee should ensure the continued existence of the server, but you’re free to interact with the rest of the fediverse using your account there.
I use social.coop.
I count that under public funding, though my description conflated public and central government.
I think it’s important to distinguish member funded vs public funded. Public funding implies allocation of a fraction of resources of the whole state to a particular purposes. Member funded implies those involved fund their own operations.
The Mastodon subculture seems to be developing that way, sure.
Source?
I couldn’t point you to anything specific but I’ve seen it before in earnest.
There are certainly people who feel this way, but it’s very misleading to frame it as “the dream of the fediverse” when it’s really just a sub-contingent.
The rest of us are just happy to have a place where we can hang out with our friends without getting sucked into corporate social hellscape. Not every garden has to grow to cover the face of the entire earth to be a success.
Every instance admin that’s been hoarding users in hopes of cashing in on The Next Big Thing.
Join on what terms though? It’s entirely possible for every ActivityPub user to join Facebook already: it’s called creating a Facebook account. Clearly that’s not acceptable terms, because people aren’t doing it.
So, what exactly does the term “join” mean here, if not the community having control over what features are included in the project going forwards? And if there’s ever a disagreement with Facebook, how can ActivityPub users actually push back against Facebook in any way except by defederating?
The dream of the fediverse is, if anything, to replace Facebook entirely. The platform is nice, but the corporate entity behind it is driven by profits, and there is nothing ActivityPub can do to force Facebook to adopt different values.
Thu benefit of federation is you don’t really need terms in the same was as corporate integration deals. To make a Facebook account you need to agree to Facebook ToS etc, but for Facebook to become part of the fediverse they just need to support certain standard protocols on their endpoints and advertise them. Then a Facebook user can follow any ActivityPub user or vice versa, and neither Facebook nor the ActivityPub admin need to even be aware that this is possible for it to happen.
Would it be nice if this gave more people the freedom to leave Facebook because they are no longer locked in to its network effect, since they can still communicate with their Facebook using contacts from elsewhere? Yes. That’s awesome. It’s also the reason no one ever believed we’d see Facebook or Twitter seriously considering federation, but if they can be convinced to see past this short term loss of control I think it would be good for them in the long term as well, and maybe they do see that since they are thinking about it.
Will Facebook then have some data on me? That I don’t want them to have and hence self-host a fedi node for myself? I think that’s kinda the point of profits and values.
When you have a publicly-visible account, information that you’ve chosen to make public is publicly available.
No, I’ve meant kinda in network. When you, say, search for tags or something. I know that the distinction is mostly irrelevant, but the point is, I’d be willing to give my tag search preferences to another fedi admin, but not to a corporation. But if the corporation is networked, they’ll get that data.
And yeah, by “a corporation”, I mean specifically meta. I left Facebook years ago because of that data that I did not want to give them.
They wouldn’t have data on searches on other instances, since searches are entirely local.
If that point is correct, then there should exist at least one XMPP extension that was widely implemented except by Google, and its fate discouraged XEP authors/developers, leading to that slowdown. Right? I don’t recall any such extension. Did I miss any?
It’s well documented that Google implemented partial incompatible XMPP.
In future, it’s considered polite to read the linked article in full before commenting.
I’d love an example. As a veteran of that time all I remember is them contributing to the community, not being gratuitously incompatible
I know what Google did and did not implement, and I know the uptime problems. I don’t know that it led to a slowdown in XEPs being written, a slowdown in XEPs being implemented in either servers or clients, or anything else I’d call XMPP’s developers slowing their pace to suit Google, and that absent arguments, blaming google is a just propter hoc to deflect from XMPP’s homemade problems.
FWIW, I liked XMPP, wanted it to succeed, even had an XMPP address on my business card for years. It seems likely that I’d hear about such problems, even if not certain.
I think that’s quite an interesting interpretation of what happened. My recollection (I was involved in XMPP before and a bit after it became an IETF standard and wrote a couple of clients) it was dying before Google got involved. The standards process ran ahead with complex designs that were 2-3 years ahead of what anyone could implement. There was a push for pub-sub (with no implementations) then everything had to be built on top of pub-sub. Personal Eventing over Pub-sub (PEP) provided a generic mechanism for doing event notification things. Nothing could be standardised if it could be built on top of PEP but wasn’t, but PEP depended on server support and client support, neither of which existed in the popular implementations. Instead, everyone pushed their own extensions that weren’t interoperable. There were multiple different ways of doing file transfer and even of sharing avatars, most of which worked in 1-2 clients.
Google came along and dumped the Jingle (voice / video chat, primarily) specs on everyone. They were overengineered but at least they were well supported by the client that suddenly had as many users as all of the other XMPP clients combined. Everyone else struggled to catch up with Google Talk’s Jingle implementation. Other things were layered on Jingle (it could establish streams, so let’s use it for file transfer and throw away the other 10 ways of doing file transfer that had XEPs), but Jingle was so big that many things that did X-over-Jingle didn’t actually interoperate because Jingle they each supported disjoint sets of Jingle transports.
The whole thing was a mess. It needed a strict IETF-style rule that you can’t propose a standard without one implementation and you can’t advance it without two independent interoperable ones. It also needed a permissively licensed reference implementation of a client library and a blessed server (there was a lot of drama here too: the JSF produced a reference server, then decided to rearchitect it completely in a way that broke plugins, then decided to endorse a different one, then another different one).
There are a lot of lessons to learn from XMPP, but ‘don’t let corporations get involved’ is not the right one.
It’s also a weird thing to argue because XMPP was not hurt by Google leaving except that all those users stopped being able to federate. The rest of the ecosystem kept going just as it always had and still is today
For what it’s worth, the B?DFL of Mastodon (Eugen Rochko/Gargron) seems indifferent to positive about Meta’s involvement - at least, he seems to have agreed to speak to them under NDA instead of laughing them out of the room* - so I wouldn’t be so sure.
Whether or not Facebook will be able to “kill” the Fediverse, I think it will definitely try to, and that’s reason enough for me to be concerned.
Besides that, let’s remember: its business model is collecting data and using it to market products and ideas to users. I don’t want that for myself or my userbase, so I’ve signed the fedipact and proactively blocked threads.net.
MTL Rocks will not federate with Facebook, and I encourage other Fediverse admins to do the same.
Serious question: why do you think FB will try to kill the Fediverse?
For Facebook to maintain control of all eyes and attention which equates to ad revenue, users need to stay on and interact with Facebook and/or look at Facebook ads. Allowing unrelated entities on the internet to have access to their walled garden whose whole intention is to produce a network effect and keep people goes against their financial incentive. It just doesn’t make any sense. And any time something doesn’t make sense, it likely points to missing information. The hypothesis that Facebook wants to kill the Fediverse fills that gap in ideas: Having learned from Twitter’s exodus, by embracing and then killing the Fediverse off, Facebook will be able to better control and maintain its position. Now it all makes sense.
That’s not the only hypothesis that makes sense. I’m sure there are other perspectives. But putting together the timing with Twitter’s exodus, attitude of Meta toward its users, history of exploitation, and financial incentive to exploit some more, it’s hard to think of an alternative viable explanation for Facebook’s decision.
Also, things like Facebook trying to hold secret conversations with server admins “off the record” doesn’t sound very much like they want to play nice.
The Fedi user base is a rounding error for FB. Mastodon.Social has 1M active accounts. FB has 2.85B. That’s not counting Instagram and Whatsapp, both of which also have billions of users. The idea that FB would actively[1] crush the Fediverse is ludicrous. Most people who are active there hate FB anyway, but they are mostly preaching to the converted. It’s fun seeing yourself as the rebel alliance fighting the empire, but the empire probably doesn’t even know you exist.
My take - someone inside FB really likes the idea of ActivityPub, and has enough clout to get the company interested in it, at least in an exploratory way. I’m betting they’ll use an in-house solution for Threads in the end.
[1] I’m not saying there can be inadvertent damage though.
Facebook is not an indie passion project run by a couple of people. It’s a multi billion dollar advertisement company whose sole purpose to increase revenue at whatever cost to others. It’s not plausible that some dev somewhere likes ActivityPub and all of a sudden a whole team is engaging in “private meetings” with admins which include “confidential information”. That’s naive at best, and malicious at worst.
Obviously “The Empire” knows you exist and is threatened by you enough to take action and allocate resources to fight, so that point is invalid.
I also don’t buy the “inadvertent” here. The executives at companies that generate billions are way too smart for “inadvertent”.
This is entirely plausible, if the meetings are “if we build this, your server will get an estimated N million visits a month. Can it handle that?” That estimate is privileged information because it reveals details of how much traffic Meta is estimating .
LOL no, they’re trying to build a Twitter clone and are investigating whether AP is a good fit. I suspect it’s not, they’ll roll out their own solution, and the angst will begin about how AP is a “dying protocol” because it wasn’t chosen by Meta.
Why would traffic estimate be so confidential that admins are put under NDA and prevented legally from discussing it? What part of that is confidential? Makes no sense.
If they want a Twitter clone, they can unilaterally implement it. No need to get involved with multiple ActivityPub instances. Again, there is no money in that, no incentive. Makes no sense.
The question to ask is: How can federating with these public instances generate advertising revenue for Meta and put it in a more dominant position? I’m not smart enough to answer that. But these arguments just fall apart like meat on a bone that’s been cooking slowly over 24h. Mmmm I’m getting hungry.
I just can’t see Meta being a good webizen. Fool me once, twice, three time, seriously, that’s enough. So unless everyone who works there had some sort of ethical epiphany and are now singing kumbaya by the fire pit thinking about world healing, it’s unlikely this is a benign venture.
Facebook is a for-profit company, why the hell would they want to open up their platform? And if they’re not interested in opening up their platform, then 1) any relationship is doomed from the start anyway, and 2) their obvious motivation is EEE.
That’s not “killing the Fediverse”, it’s at most setting up obstacles to full federation from other AP clients and servers other than FB’s. Basically what the FB haters want to happen anyway.
Microsoft did it a lot.
It has a name in 3 words.
It’s on the tip of my tongue.
The late 90s called, it wants its conspiracy theories back.
You can run Linux fully supported within Windows.
The most popular code editor in the world is open source, and backed by Microsoft.
Open source is good business for Microsoft now, there’s no need to “extinguish” it, because it’s making MS lots and lots of money.
Anyway, Microsoft and Meta are 2 different companies, run by different people, with different main revenue sources. Why one should follow the deprecated playbook of another, to counter a threat that’s insignificant, is beyond my limited understanding.
You say that like it’s a good thing.
It’s a neutral observation. The MS in the 90s under Steve Ballmer is a different company than the one today. They lost an entire generation of programmers to Linux and open source languages and had to adapt.
In the context of an argument that “you don’t have to worry about a big company coming in and bringing their values into your existing community” it doesn’t feel neutral.
They wanted control; their original method of getting it failed, and then they found a more effective way to bypass the community’s resistance, but in the end they still got what they wanted.
Having not really been paying attention at the time to the various XMPP servers and whatnot, was there the same sort of issue around blacklisting and petty politics that seems to plague the Fediverse?
I’ve avoided getting involved in that scene in large part due to the outsize amount of “we don’t peer with so-and-so”, “omg you peer with this site, well one person on that site said problematic thing so we don’t peer with them anymore”, etc. That would seem to be a bigger threat than Meta.
I dunno, the network has been growing pretty well for 6 or 7 years at this point and while there are occasional issues with admin drama, it hasn’t yet ended up causing a mass exodus, and I don’t think it will. The closest it got was the snouts.online incident, but like everything with OStatus/ActivityPub, that just lead to there being a little island of people who only talk to each other. There similarly is a “sexualized images depicting minors” island, a “Gab/alt-right” island, and a few intentional islands, but that seems like a success of the federated model to me! Those people have their social spaces, and their presence doesn’t cause issues - whether legal or social - for anyone else.
I don’t necessarily think that coupling hosting with moderation in this way is optimal, but it does work, and more so the more instances there are.
The petty politics are pretty specific to a corner of the fediverse (mostly around Mastodon) I think because of the kind of people who came there in the last wave and we they came and some pretty big bugs in Mastodon exacerbate it. I’ve not observed this issue on other federated networks like SMTP, XMPP, IndieWeb, etc to date
No. There weren’t enough people using XMPP off Google for it to be worth targeting with spam. If you wanted to, it was easy (register a new domain name, deploy a server, and you can spam anyone). Since it was point-to-point, not broadcast, there was no need to block servers for any reason other than spam.
If it had become popular, the lack of spam controls would have been a problem. It’s a bit better than email (you at least need a machine with a DNS record pointing at it to be able to send spam and you can accurately attribute the sender) but not much because the cost of setting up a new server is so low. Some folks were starting to think about building distributed reputation systems. For example, you can build a set of servers that you trust from the ones that people on your server send messages to. You can then share with them the set of servers that they trust and allow messages for the first time from servers on that list, but doing that in a privacy-preserving way is non-trivial.
The problem cited with Google and XMPP seems to be an issue with a single “server” having outsized influence over the XMPP fediverse. In this scenario, Google was effectively a single XMPP server that knew it could wield its users as influence over the existing servers.
If the modern day fediverse decide to play ball with Meta, then make federating with them conditional – Meta has to bring one of their competitors to the federation table.
I’m not saying this would necessarily be a good outcome for the fediverse, but if the fediverse want to disarm the cudgel of unilateral action by an outsized party, then bring more outsized parties to the game.
I’m feeling more hopeful about federated messaging given the EU’s recent Digital Markets Act legislation. As was mentioned in your article, Google was able to implement a federate chat protocol once before. They’ll be able to do it again. We’ll see soon how they’ll implement E2EE (if they’re able to that is).
The same thing happened to one of the ideals of OpenID: being able to use any OpenID provider, including self-hosting, to authenticate. Now it’s basically just a small select list for most instances, with the options available dependent on the slice of the internet you’re using. Heck, even OAuth is a bit of a clusterfuck.
OpenID sort of ate its own tail. Similar to how OStatus blew itself up and we entered the federated social dark ages with the complexity of ActivityPub, OpenID2 got blown up by “well OIDC is coming” and then by the time fully federated OIDC was ready everyone had moved on. You can use federated OIDC today same as you did for OpenID2, but I’ve not yet seen any;he who does. Maybe we’ll have our Mastodon moment for that tech yet and bring us back out of the dark ages.
Well, it didn’t really kill anything. For all intents and purposes XMPP is equally popular as it was before Google Talk was a thing, which TFA actually acknowledges. Yes, for a while it was very popular, but there are plenty of spaces on the internet that run XMPP relays and servers. It’s fine.
Well, ackshwally, we stopped using Google Talk all together, because our clients stopped being compatible with Google’s servers and their own client was a joke.
No, XMPP was always niche. Exactly the same way it’s fine if the fediverse IS and eventually will be niche when Meta gets bored or stops being a good neighbour.
Please piss off ringing the alarmist death knell of technologies. You’re not netcraft.
Well, ackshwally, XMPP went from lots of interested parties complaining about its complexity in adoption but saying “OK, fine” to immediately saying, when GTalk took it on, “oh cool! My slog will be validated” to a big collective sigh when Google dropped it, followed by a bunch of shitting on XMPP: “I knew it wouldn’t work.” “It’s done.” “Good thing we didn’t actually follow through.” “ejabberd was too hard to run, and a pile, anyway!” “Who writes shit in Erlang?” “Eh, I hate XML. No idea why it had to be XML—they screwed themselves with that decision.”
There was lots of hope in the XMPP space before GTalk dropped it. lots of people considering how to incorporate/ build new. Maybe there aren’t any more/ less contributors to implementations or specs now. Maybe there’s more/less the same number of people connecting over XMPP; but something did change. Potential adopters left, and so did any hope among the commons that XMPP was worth fighting for.
You make some good points, but I would still be wary of confusing the cause for the effect. Perhaps that all the interested parties that stopped investing mind share and development time into XMPP, did so because the technology itself was not worthy. Or that it was worth only as long as a big player like Google was also involved.
I’m not sure if you’d agree that the distinction is meaningful, but for me it makes sense.
I don’t think this is true. I think both Facebook and Google had incentives to move faster than the specification could, and the decentralized foothold wasn’t as strong as say, SMTP / email that came before it. If 99.9999% of your users don’t communicate outside your network, why bother?
–
As for why others “lost interest,” put it this way, I think a lot of developers look to large companies, for guidance. They have the ability to hire the best of the best so it naturally follows that the things they do, the papers they publish “must be solid” or else they wouldn’t do it. If XMPP doesn’t work for Google, how could it possibly work for me when I get to Google scale?
Obviously, you (well maybe you are – but you get my point) and I aren’t “Google scale,” and given we’re actively analyzing it we can think a bit more critically. But most people either can’t, or for whatever reason don’t think critically. It’s exactly why the industry is flooded with articles, conference talks, and products selling you on using micro services running on Kubernetes and stream processing, to publish a blog instead of just using WordPress.
Why this kind of rudeness and hostility? You have a point and the point was well made but then why this kind of impoliteness in the end? I am afraid but I have to protest. I don’t think this kind of behavior towards a member of our own community is ok in this forum!
Granted, out of context my rudeness seems disproportionate, but as someone that develops ActivityPub software, all my social network inputs have been clogged with similar defeatist hot takes.
The points that OP makes in the article are nothing but inflammatory two bit pieces of information that if you squint at them at juuust the right angle, they might look like they spell T H E E N D.
I’d like to see OP put in the work into making the Meta Embrace/Extend/Extinguish game less of a possibility instead of piling on the negativity like a two bit technological oracle. It takes very little effort to talk down other people’s attempts. I am annoyed.
Exactly. It would be more accurate to say XMPP killed google talk. More accurate, but not accurate.
The worst part of reading all these awful internet history hot takes is that I’m pretty sure they’ll all just become accepted history. I guess that’s what all history you live through looks like.