Hacker News new | ask | show | jobs
by nocoolnametom 1479 days ago
In order to talk to my friends, family, and coworkers I need to have the following apps installed and running: Slack, Teams, Telegram, Signal, WhatsApp, Google Chat/Hangouts/Allo/Whatever, FB Messenger, Discord, Twitter, etc.

It'd take a pretty strong argument to convince me that this is so much more productive and allows for more innovation than the old days when the spec for things like Email, HTTP, IRC, XMPP allowed for a plethora of different tools unrelated to the company sponsoring the tech and people figured out how to make money USING the interoperable tech instead of OWNING the tech.

4 comments

I actually love the choice and the separation. And when an innovation good enough appears on a platform, it is quickly copied on the others (reactions…)
I'm willing to bet that your friends and family wouldn't be happy if the European Union would mandate using IRC everywhere. Heck, why not go further and stick to the good old ntalk [1].

[1]: https://en.wikipedia.org/wiki/Talk_(software)

What about the XMPP standard? I use it everyday for messaging family and friends.

WhatsApp is basically an unfederated XMPP provider.

I used XMPP around 2003 when it was still called Jabber. I can't say there's something major wrong with it (only XML verbosity comes to my head), it's just the idea of making it mandatory. By the way how come some EU officials use Zoom? [1] Where are those good open standards?

[1]: https://meeteu.eu/events/

Funny you should mention Zoom, which just like Whatsapp is pretty much a half-baked proprietary XMPP implementation. Now if we had proper interop regulations mandating interoperability between commercial entities (no need to apply that to hobby/research projects), we could talk between all these networks.

Sure it would take a few months of serious dedication for these chat vendors to write specifications for the protocol spaghetti they came up with, but the benefits would be tremendous.

So why is "nobody" using XMPP protocol? The problem is not exactly with the specifications (although there's still a little margin for interpretation here and there, they keep evolving for the better) but rather with the implementations. Since a protocol is not tied to a single implementation, it requires additional resources to develop user-friendly clients. This fact is used by an argument by some people (see also: m0xie's The Ecosystem is moving) to justify centralizing all communications and protocol development. This argument was amply debunked by Daniel Gultsche (who maintains an Android XMPP client) and Drew Devault (who maintains an (unfederated-so-far) forge):

https://gultsch.de/objection.html

https://drewdevault.com/2018/08/08/Signal.html

There's also a lot to say about the Matrix/Element approach, which has some good and bad sides. I'm happy to elaborate if that's of interest to someone.

If you want to have provider choice for customers and interoperability between messaging apps, I don't think there is another way than making standards mandatory.

Why do you think WhatsApp reached a billion dollar evaluation? Not because users have the freedom to move to a different provider and still be able to talk to all their friends...

The only other possibility is that users start rejecting providers who do not comply with internet standards (and I don't see that happening, even here on Hacker News).

Making something mandatory will definitely push things forward because what else are the providers going to do? Though I still have doubts about real progress being made. EU can't can even add VP9 and AV1 to the list of codecs used for Digital Video Broadcasting (DVB).

Instead maybe the government should start eating its own dog food and use an open standard for both internal and external use.

Interoperability of major chat software will be mandated by the EU in the coming years (around 2023 or 2024 hopefully).
It will be very impressive if they get this to actually work, as opposed to an endless flood of unstoppable spam forever.
Seems like a disingenuous argument, why not make a real one?
ntalk is for me the most pleasant non-in-person way to communicate.

Let's please go back to ntalk!

I really hope Matrix bridges will help bring back some sanity on this front.
the performance is too bad
The bridges are not horrible. But they aren’t super reliable. I have seen them go down for a few days once, generally be a bit slow, forward messages out of order, etc.

The free matrix.org server is also overloaded. The paid server is much faster.

Still, bridges do not really solve fragmentation problems the same way compliance with internet standards does.

For example bridges break important features like end-to-end encryption.

Internet standards have consistently failed to innovate. Email and IRC have failed to progress along with proprietary platforms. Features like end to end encryption require user effort and plugins which never took off.

Having 5 IM apps installed is much preferable to me than 1 worse app. Not sure the situation for XMPP but I have been told its highly fragmented with extensions that not all clients support. If I'm talking with someone, I want a high level of assurance that their client is pretty much the same as mine and that all features work and look roughly the same on both sides.

> Internet standards have consistently failed to innovate.

Internet standards do not innovate. People that adopt standards and built interesting things with them do.

For example, Google originally wrote the Jingle XMPP extension. Sadly, it seems like there are no real economic incentives for interoperabilty in the IM space (quite the opposite if you already have a lot of users on your platform), so we don't see a lot of investment going into the adoption of internet standards from big players anymore.

I'm using Element One these days (https://element.io/element-one) which at least gets me Signal, WhatsApp, Telegram, Matrix and IRC all in one place.