Comment by B1FIDO

13 hours ago

I must protest that this kind of announcement belies the stupidity of proprietary chat protocols.

Remember when IRC was king, and basically, anyone could write an IRC client? Anyone could write a MUD client, or even a Telnet client. Those are open protocols.

When Pidgin came out, it was like a breath of fresh air for me. In the early 90s I had multiple IM accounts (starting with ICQ!) and unifying them, especially under a Linux client, was a dream come true.

But of course, AIM purported to use Oscar at the time, but they really hated F/OSS and 3rd-party clients, and so did the other proprietary guys, so it became cat-and-mouse to keep the client compatible while the servers always tried to break their functionality.

Now this dumb announcement comes out that a 3rd party has (apparently legally) established interop with a Meta property with (I am guessing) a completely proprietary, undocumented, secret protocol underneath.

I am not impressed. I am McKayla Maroney unimpressed.

I want open protocols and I want client devs who are free to produce clients in freeform, as long as they can follow the protocol specs. Now we have email clients who speak SMTP, IMAP, and POP3, including the "secured, encrypted" versions of those protocols. We should ask for nothing less when it comes to other communications.

We had XMPP, and even Google Chat used that in the early days.

It's not like users haven't had choice over the decades to choose software that runs on open standards. It's that the features and UX provided by closed software has been more compelling to them. Open standards and interoperability generally aren't features most people value when it comes to chat. They care mostly about what their friends and family are using.

  • The issue isn't closed vs open but business models. The reason most services don't support third-party clients is that their business model is based on advertising (aka wasting the user's time) and a third-party client would reduce said wasted time.

    A proprietary/for-profit messenger can very well use open protocols and embrace third-party clients if their business model wasn't explicitly based on anti-productivity.

    • Right. Unfortunately, people have overwhelmingly voted with their wallets, and prefer to pay with their time and attention (and ignore the fact that they're being psychologically manipulated into buying random products and services) than with actual cash.

      I expect you could get some people to pay for a messaging platform, but it would be a very small platform, and your business would not grow very much. And most of your users will still have to use other (proprietary, closed) messaging services as well, to talk to their friends and family who don't want to pay for your platform. While that wouldn't be a failure, I wouldn't really call that a significant win, either.

      This is why legislation/regulation is the only way to make this happen. The so-called "free market" (a thing that doesn't really exist) can never succeed at this, to the detriment of us all.

    • The problem is that there's not much of a market for an ecosystem of commercial chat clients that use open standards underneath. It's not like it hasn't been tried. What ultimately ends up happening is the market becomes a race to the bottom, chat clients become a commodity product, and innovation ceases. It's essentially what happened with Web browsers and why we don't have a particularly robust for-profit market in that space.

  • Google Chat used XMPP to build an user base and then cut it off from the Jabber network. That's when I stopped using it. Or was it when it got integrated into Gmail? Then they rebranded it and binned each iteration several times.

    • Similar to Slack and IRC? I guess that's just part of the enshittification process.

> Now this dumb announcement comes out that a 3rd party has (apparently legally) established interop with a Meta property with (I am guessing) a completely proprietary, undocumented, secret protocol underneath.

Resd the article - this isn't a proprietary secret API, it's the official intended interop API the EU now obliges them to provide. Not exactly 100% what you're asking for (I too would prefer common standards) but forcing interop access is a very good start.

Social networks and chat apps are mostly dominated by the network effect.

Since the purpose of these apps is literally putting you in contact with other people, you tend to use the same app/social network most of your friends and family are using.

This is not necessarily true for platforms you use to find new people, but even then, you're going to use the websites/apps people with your interests are using.

  • I don't think his rant is against social networks or instant messaging perse, but about vendor lock in.

    The way I read it is along the lines of Mike Masnick's protocols not platforms.

    https://knightcolumbia.org/content/protocols-not-platforms-a...

    • I understand, but in this specific arena, because of the network effect, interoperability is important so you can hope to make a competitive product.

      More in general,standard protocols are important but they don't necessarily avoid lock-in.

      For example, imagine a Dropbox equivalent with a public API specification.

      At some point you want to leave. You are ready to use Postman or even curl and download everything to upload it somewhere else... but download is capped at 10 files/day per user. And you uploaded 100,000 files over years.

      The API is public but good luck leaving with all your files!

      In other words, standard protocols help avoiding client lock-in, but when the value is on the server side (data,...), they are not enough.

Matrix is getting traction though...

  • Matrix is a lost cause. The protocol is too complex/ambitious and the company behind it doesn't have the resources to actually produce a good server nor client implementation. I was hopeful for it at first but at some point you have to be realistic.

    • While I agree with you, and there should more diverse members than just the people from Element.

      What I do like about them is the zero server trust stand they are taking on their clients which makes migration a pain in the butt, but that is what one would expect from a true e2ee chat app.

      And now they have two stable servers in rust. The French and German government including military are using the protocol to make their own apps. Maybe it should be something the EU should put some more resource into it?

      3 replies →

    • It was the invite floods of what was probably CP and cat torture that made me uninstall it and never look back.

      No thanks on that. I don't have time or energy for these things.

  • Is it? My experience with it has been middling at best, and I communicate with exactly zero people through Matrix outside of the context of open source projects.

    The UX is still pretty bad, with many rough edges around sign-in and device verification. The message/encryption story has gotten better (it's been a long time since I've gotten spurious errors about being unable to decrypt messages), but it's still not particularly easy to use. Performance-wise I've found it to still be fairly bad; loading messages after I've been offline takes a noticeable amount of pause, something I rarely see with other messaging platforms.

    On the plus side, Matrix does have many chat features that many people like (or even require) in a chat platform, like formatting, emojis, message reactions, threads, etc.

WhatsApp uses the open Signal Protocol.

  • That's a bit misleading. WhatsApp uses Signal's end-to-end encryption scheme, but not Signal's networking protocol. It's still proprietary. Otherwise, we could have cross-messaging between Signal and Whatsapp.

    • WhatsApp just implemented cross messaging using the open Signal Protocol forced by the EU. We will see if the Signal messenger enables interop with WhatsApp, they are not forced to do this.

> I must protest that this kind of announcement belies the stupidity of proprietary chat protocols. [...] In the early 90s I had multiple IM accounts (starting with ICQ!) and unifying them, especially under a Linux client, was a dream come true.

ICQ was also a proprietary chat protocol. The Pidgin (then "Gaim") developers had to reverse-engineer it. Fortunately the folks at ICQ were less hostile toward third-party clients than AOL was toward Gaim's reverse-engineer of AIM's protocol, as you note. (Not to mention sending legal threats to the Gaim/Pidgin team to get them to change the name of the app.)

IRC was indeed king, when the internet was populated mostly by technically-savvy folks who could deal with its rough edges. (For example, you probably forget how annoying it was to get file transfer working over IRC; sometimes it was just impossible to do, depending on clients and NAT conditions and so forth. Things like ChanServ and NickServ were creative, but inelegant, hacks, functions that the protocol should handle directly.) And consider that IRC has more or less not changed at all in decades. I am a technically-savvy user, and I gave up on IRC, switching to Matrix for those types of chats, which has its own rough edges, but at least has modern features to sorta kinda make up for it. (Otherwise I generally use Signal, or, if I can't get people to switch, Whatsapp.) I want to be able to do simple formatting, react to messages, edit messages, etc. And most people in the world seem to want those things too. IRC has stagnated and doesn't meet most people's needs.

But I absolutely agree in that I want open protocols too. It's just hard to fight against big corporations with endless development, design, and marketing budgets. And those big corporations are not incentivized to build or support open protocols; in fact they are incentivized to do the opposite. As much as the EU does get some things wrong, I think we need strong governments to force companies to open up their protocols and systems for interoperability, and to stamp down hard on them when they comply maliciously, as Apple and Meta does. The EU is pretty much the only entity that comes close to doing that. I really wish the US was more forward-thinking, but our government is full of oligarchs and oligarch-wannabes these days, thanks to the lack of any meaningful campaign finance limits. At least California (where I live) has some GDPR-inspired privacy legislation, but I think something like the EU's DMA is still too "out there" for us here, unfortunately.

  • ICQ was not only proprietary, but it was centralized and server-based, even though the messaging part was peer-to-peer.

    Even in those heady early days of the mid-90s, it was recognized that many end-users were behind NAT and firewalls and otherwise-inaccessible endpoints of the Internet. Many of us were also on dialup lines that were intermittently connected, so they needed to establish some sort of persistent presence.

    So the ICQ client was designed to check-in with a central server to indicate the online/away/DND/offline status of the client. I do not know how much of ICQ's messaging went through that server, but I believe that a lot of clients in those days were designed to, eventually, connect peer-to-peer for delivering files and stuff. Mainly, because the operators of servers didn't want to be overwhelmed with transferring lots of data!

    Interestingly, ICQ and Livejournal as well were completely invaded and taken over by Russians. Or perhaps it was not an invasion, but a planned psy-op all along. My original UIN was 279866, and my girlfriend's was slightly below that: she had signed up first and got me on-board.

    But eventually, Russians broke into my account, changed the profile, and commandeered it for their own purposes. And Livejournal got sold to Russian interests too.

    I believe it was them watching us over here all along. It must have been a personal-data goldmine to know when teens and young adults were online and who they were connected to, on the social graph, whether it was IM'ing or blogging the old-fashioned way on Livejournal.

    So beware with your modern "disruptive" apps, particularly ones like those fun e-Scooters you can share and rent. They are probably psy-ops from foreign-based actors who enjoy watching and recording our movements.