If you'd like to draw some conclusions about ActivityPub from this, it should not be that a network should disallow supernodes, but that we have to counterbalance them with a wide diversity of other nodes of different sizes with different value propositions.
@evan I want to do this, I want my own server, and contribute to decentralization.
I recently joined infosec's relay and it's filling up my HD and slowing down my server. That is only one relay, a portion of the :fediverse:, if I wanted to federate with Threads, that is bigger than 35x the :fediverse: it's not technically possible.
@anclement@evan "Supernodes" are okay if they are not predators with a track record like meta. We know how meta runs its social networks and we ran away from its playforms for something different.
@dilmandila@evan but Meta brings eyeballs and users to the the fediverse. It also brings brands to the fediverse. I know a lot of people may not like those things.
But rather than block and run away, shouldn't we try and address the issues in the standard?
@anclement@dilmandila nobody should ever have to interact with anyone they don't want to on the social web. EVER. That is fundamental, and I will fight tooth and nail against any structure that doesn't let people shape their own social space.
But that is part of the protocol already. Blocking accounts and defederation already work, as well as blocking words. We need better support for Bayesian filtering, but I think that will come.
No, it is not a violation of "human rights" if, say, my instance or all instances choose not to federate with Threads.
That you bring "human rights" into the issue is a measure of the poverty of your actual argument. And I might add that Facebook has complete disregard for human rights.
Article 19: "Everyone has the right to freedom of opinion and expression; this right includes freedom to hold opinions without interference and to seek, receive and impart information and ideas through any media and regardless of frontiers."
Article 27: "Everyone has the right freely to participate in the cultural life of the community, to enjoy the arts and to share in scientific advancement and its benefits."
@TomSwirly@anclement@dilmandila you have a right to decide who you interact with. I 100% support that and I will fight to make sure you can always exercise it. But all people have a right to be on the fediverse, even if they can't have a conversation with Tom Swirly.
Note for my subtooter: there's not a recipe for this. Sometimes a group of independent Open Source developers and advocates manage to change the world. Other times, they don't. I think one big success factor is having people whose full-time job is expanding the network and making it healthy, unrelated to their own node or implementation. But not always. If I knew how it works and could replicate it consistently, the Internet would be a lot different.
What makes any network “work” or not is a balance of features and humans.
I am amazed that all the new apps or platforms [threads] are not taking the time to be more creative and sociotechnical about the features with a recognition to NOT recreate the past. Whether fediverse or owned.
@evan
Once communication services get to a certain size there ought to be regulations that enforce opening up to standard protocols. Think of the phone systems, email systems and terrestrial TV.
Are you saying that "#supernodes" (like #Threads?) are o.k., but there need to be more "standard" #Fediverse instances to counterbalance their huge number of users?
@evan when Galactica, flagship of the Colonial Fleet but with strict defederation policies, encountered the more powerful Battlestar Pegasus it was important to both collaborate with the Pegasus against a common enemy, but also maintain autonomy as the the interests of the Pegasus were not always aligned with the fleet.
The Galactica, despite being outmatched by the Pegasus, remained crucial to the fleet as a focal point and a strategic asset in a role that smaller ships could not have played.
@evan If this is the conclusion we should draw, is there any case of an open protocol where it has played out this way? Where a flourishing ecosystem of small nodes successfully counter-balanced the big ones and everyone is still free to spin up new nodes and enter the ecosystem? Genuinely curious, because I haven't found a auch a case, but I'm not around that long.
@liaizon@evan@marcelweiss that's a good point. It is a collection of protocols rather than one and a number of them have in turn been somewhat oligopolized, but as long as you play by the rules of the established players, you can participate.
@mcg I didn't say that supernodes are not a problem. I said that a network should not disallow supernodes.
The way to deal with supernodes is to have multiple nodes with similar scale. So, not just Meta, but also LinkedIn, Twitter, Tiktok, WeChat, StackOverflow.
You also have to have a lot of nodes at different scales -- 1 person, 100 person, 1M person, 100M person. Not a bad heuristic to have about same number of people total at each scale.
@evan Yes, we must have "nodes ... with different value propositions." Today, too much attention is focused on the one problem of shaping audiences (i.e. blocking/moderation/censorship) while all the implementations look pretty much the same.
Ideally, we'll see innovation in the user experience. Chemists, economists, gamers, photographers, or journalists could probably all identify domain-specific features that would serve their specific niche better than the general systems we see today...
@evan i don't know enough to weight in on whether 'google extinguished xmpp' is an ahistorical take but some of the reaction to threads does seem like panicking and nothing more. Since i assume this is a post about that
@evan i'm glad you are saying this. That narrative and other stuff are just infesting the community and we are all worse for it.
And for the longest time ive been quite alone in my stance, but it seems to change.
But as one of the protocol people, why don't activetypub have more of a central branding organization like matrix.org. It has its drawbacks and advantages, but I have always wo dered about this.
@evan another narrative on activetypub and digital politics that I hate is that since "activetypub is so loose, isnt these specific things it is a bad protocol", which is even said by like proper developers, and that makes no sense to me.
Technology and infrastructure don't need to be ideal, being ideal can actually be a drawback. It only needs to be good enough. The actually good stuff are just good enough on what matter.
@evan Spam and harassment also existed in email, at orders of magnitude greater levels, and yet Google did not try to kill open email standards but simply solved the issue.
They rolled out Gmail on April 1, 2004, with a good spam checker that improved over time, and didn't get into XMPP until much later.
You present no evidence of the claim that "spam and harassment outweighed the benefits". No evidence appears to exist except a single press release, bereft of even a single number.
@TomSwirly what you're ignoring is the topology of the network. Google could turn off XMPP for GChat because there weren't enough people on the network to make it a business necessity. Google cannot do that with email, because there are enough people on servers outside Gmail that they can't turn off SMTP without pissing off their customers.
@evan
"And they didn't do anything wrong by disconnecting from the network when the spam and harassment outweighed the benefit to their users."
Thats what some instances are trying to prevent here. disconnect from a major spam, harassment & "you're the product", ad platform.
harassment from meta "bots" to those who block threads, has started....
ads too.
@evan This message requires a certain amount of maturity and seriousness to hear. Puts responsibility on the folks who didn't step up. Which is part of why there's so much resistance to this truthbomb.
@colo_lee@evan Can you explain what "stepping up" would have been, exactly, and who exactly are the people who didn't do it?
What solid evidence do we have that "not stepping up" caused Google and FB to silently and gradually stop supporting XMPP, except a single unsupported and unspecific statement by a Google spokesperson vaguely blaming Spam, with no details or numbers?
If Google had broken email for the same reason, would you have agreed?
@TomSwirly@colo_lee@evan You keep saying that, what evidence do you have of EEE besides people's blogposts? Really go back and ask people about their experiences with it, go back and read about people's thoughts on their experiences, not just your opinions and those that align with you.
What percentage of the traffic was spam or harassing? Was it 10%, 1% or 0.0001%? How much effort did dealing with this take? Why couldn't the spam protection mechanisms of gmail be used to the same purpose?
If they wanted to be transparent at the time or any time since, they could. It has been proven in court repeatedly that Google are deliberate monopolists. There is no reason to give them the benefit of the doubt.
Businesses try to make money. That's what they are for.
The open XMPP connection for GChat wasn't making them any money. Users on GChat weren't clamoring for it. They were complaining about the spam coming through.
We didn't have enough other networks connected to make being on the federation a business necessity.
Compare email: most email addresses aren't on Gmail. If they disconnected from other servers, it would be unacceptable for their users.
@TomSwirly@damon@colo_lee in the case of XMPP, stepping up meant getting more servers installed and implemented. At schools, churches, libraries, businesses, households. Having a diverse, resilient, heterogeneous network.
@evan late to this, but as I’ve argued elsewhere on this site, it is also worth noting now poorly XMPP worked on mobile and now long that took and how much denial there was (and to some degree STILL is) about why that was a fundamental flaw of the protocol. Doing mobile is a different set of constraints & it wasn’t prioritized & the attempts to modernize it were largely rebuffed. It’s understandable Google moved from a standard that didn’t work well with mobile when it focus shifted to mobile
@film_girl@evan Do you have any pointers on where to read about the history? It seems even if building a good mobile experience required a new client <-> server protocol XMPP would be fine as a common basis for server-to-server communication.
@film_girl@evan Thanks, interesting. XMPP on the user's device was probably doomed by that point, the "custom proxy server" approach where the "client" state lives server-side seems tractable but maybe by the time you're putting that much work into your client it's not worth keeping track of any remaining requirements to stay compatible with XMPP.
@film_girl Agreed. #XMPP was useful for Google given where they were, it didn't make needed progress, so of course they moved away.
Your point about how badly XMPP worked on mobile, and @evan's perspective about the responsibility he and others in the XMPP community had in not addressing the spam and harassment problems, are both important and don't usually come up when people tell the story.
@film_girl@evan@boris the big issue I had with XMPP was how it would tightly bind a conversation to the specific endpoint clients instead of letting conversations migrate. If a conversation started on my work computer it would stay there, rather than following me home in the evening or weekend, and friends would wonder why I wasn’t replying to their messages I wasn’t receiving.
@film_girl@evan I agree. The mobile / push support is solved, at least on the server side, but it was slow to take off. It still not well known or widely adopted by mobile developers.
@evan I feel like we need actual historians to figure out what really happened, because it seems to me everyone seems to have their own version of what happened.
@evan Fair enough, but I never heard spam and harassment cited as their reasons for dropping #XMPP.
In fact, I don't remember Google really giving any reasons at all; it was a typically opaque Google decision: $THING is out, $NEWTHING is in, we will be turning $THING off permanently in 90 days. Good luck everybody!
And my recollection is that XMPP-based chat disappeared at about the same time G+ was the Next Big Thing, and had its own chat system which used Google's in-house tech, not XMPP.
@kadin What's awesome is we don't have to work entirely from your faulty memory, since we have a World Wide Web where many articles and documents can be found.
Here's a contemporaneous article with a statement by a Google spokesperson:
@evan It’s worth pointing out that moderating and keeping bad actors out IS part of the product, as much or even more important than the protocol and features. Fediverse seems to be succeeding by spreading out the moderating work to many admins. That works in the current model but I predict Masto admins will have to cooperate on blocking lists to protect themselves and their users.
I was a huge champion of XMPP for a long time, but one of my original accounts (dyork@jabber.org ) was rendered completely unusable by the sheer volume of spam!
@danyork I am a big fan of the system, and this is a small corner of it, but I think for a lot of people "Google killed XMPP" is gospel, and I think it's oversimplified at best.
@evan@danyork it is even the reverse. In 2005, XMPP had a huge boost in interest, thanks to Google chat being implemented in XMPP.
The Google team working on XMPP helped the protocol at that time.
Later, Google even used XMPP as a basis for Google wave protocol. They were not hostile.
@mremond@evan I agree that it’s an oversimplification.
My own strong advocacy for #XMPP faded away as I watched all the independent (non-Cisco) clients not be able to keep up with the changing user expectations around messaging.
In particular, as #Slack (and later Teams, Matrix, etc) emerged people started using that as the chat experience they wanted to have.. and it wasn’t there in the XMPP clients I could find.
@danyork@evan I agree. Servers are developed by experts, clients can be developed by people less experts in the protocol. That's by design, when you want a protocol to be used.
As a community we should focus on making #XMPP easier to leverage on the client side by developing extensions that make their life easier and move a lot of the complexity to the server side.
@mremond@evan I think if the #XMPP community wants XMPP to be seen as a messaging alternative, there needs to be a solid client with a fantastic user experience that aligns with expectations people now have from Slack, Discord, etc. (and consumer messaging like WhatsApp, FB Messenger, etc)
I think the Matrix team has done this well with Element. It’s not the ONLY client, but it’s a default client that delivers a great UX. That helps!
@danyork@evan Yes, and for that, we need conceptually more features of the #XMPP client logic to move on the server-side.
It makes the client easier to implement, and the use of multiple clients more seamless.
@evan@EricCarroll Also, getting some kind of generally-agreed-on standard for store-and-forward and MUC through the XEP process by 2008 would have been... helpful. 😬
Seriously though it’s need frankly disconcerting to read the nonstop stream of what I can only describe as alternate-universe histories of XMPP. This wasn’t that long ago and it all played out very much in public!
@evan Furthermore, the XMPP/Mastodon comparison doesn't even make sense, because there's a lot more federated Mastodon nodes out there now than there were XMPP nodes back then.
Like, nothing's stopping us from just mass-adopting XMPP right now, the technology never stopped working just because Google and Facebook abandoned it.
@evan Also XMPP just flat-out lacked any features for moderation, even after I connected to the main dev channel and trolled them mercilessly from my own instance for a week.
How did the failure of Google Wave (which as you know uses XMPP) affect the failure of GChat? I don't want to post anything ahistorical, but I always wondered whether the difficulties of policing GWave bots was too much. Also, with my tinfoil hat, it did seem quite close to military blackboard systems in power, which was perhaps not meant for public use. Apologies if I've gone off topic (I tried to stay on!), and thanks for your inspirational work over the decades.
@pete did it fail? All the Google Workplace Apps like Docs, Sheets, and Slides have realtime collaboration built in. I assumed at least some of that came from or was inspired by Wave.
Yes, I think a lot of the team and functionality went to Docs, etc. So Wave succeeded in proving the feasibility and value of collaboration using web tech.
I just felt Wave was outward looking whereas Google Workplace is inward looking (from Google's perspective).
@evan@pete GWave was an independent effort that didn't work well with others in the company. Google has made this mistake a few times. They attempt to reduce the complexity of dependency management by having some group chartered to build up an entirely new system on a green field. That works in theory, but not in practice.
I don't think it is correct to attribute the Workplace Apps' collaboration features to GWave.
@evan Why stop at XMPP? The same thing happened to SIP. The prospect of an open interconnection was obliterated by spam and fraud. The lack of a path to innovation was why I ultimately got out of telecom.
@pettter@evan If I read this thread correctly I'd call that excuses for not having enough nodes diversity. (I'm not making any claims over the funding XMPP projects had at this moment because I wasn't there.)
@evan Google never seemed really that invested in XMPP as a standard. they never implemented s2s TLS, for example, and that feels like it was tablestakes
@est My "I am not open to your ahistorical take on Google Chat and XMPP" t-shirt has people offering a lot of ahistorical takes already rejected by my t-shirt
Add comment