I would be happy with it but my contact list quite empty.
I'm not outright rejecting XMPP. I would like it very much to just use XMPP, but something needs to happen. It needs to be easy to use, friction-less, have all the fancy features and probably more to sell than Signal, Telegram, WhatsApp, Facebook Messenger and Matrix at the same time, which is no small feat. Core XMPP does not cut it. People expect more from a chat solution and won't adopt it as-is.
XMPP has had decentralization for a very long time, but almost no users (if speaking about actual, decentralized, chat; it's used in many other things). It had both decentralization for a short time with Google chat, and user with no decentralization with Google chat and Facebook chat for a longer time, but both things are dead now.
The network effect is the biggest issue to overcome for a chat system, not the actual tech, and XMPP has not solved it. Unfortunately, if you ask me.
I can't claim it has exact feature parity with Signal/Telegram/WhatsApp, but it has all the important stuff and is good enough for my family to use daily. It also has some advantages, such as not requiring a phone number (which e.g. means that my children can use it on the old "living room tablet").
Snikket counters the "empty contact list" problem and the network effects by focusing on small groups like this, automatically including everyone else in the same group(s) in your contact list. It helps a lot.
The ultimate goal is that we get enough such small groups operating on open networks, eventually the members of those groups will find they can communicate with each other. It's better approach than isolated individuals using XMPP/whatever with empty contact lists just because they believe in it.
Is there a desktop app for Snikket?
Having been down the venture capital route in the past and not particularly enjoyed it, I'm trying to work on Snikket as close to full-time as possible while self-limiting to funding that's compatible with the open nature of the project and the ecosystem it's a part of. Currently grants, sponsorships and donations are helping immensely with that.
There is no desktop app yet, but it's firmly on the roadmap. For now I tend to point people towards Dino or Gajim for Windows and Linux, or Beagle for MacOS. None are a perfect match for the project right now, but power users can generally deal with it, and all of them are under active development and constantly improving.
Without a internet standard for IM we cannot hope for advanced features like usable federation, encrypted messaging or A/V calls between platforms. We probably should not be promoting platforms which are not standards compliant.
Don't you need XEPs to do any messaging in the first place?
Naysayers keep arguing that the jungle of XEPs make it too hard to implement and use and that's what halting the adoption of XMPP, but that's like saying no one is doing any web innovation because the specs are too big. The reality is that only a big name with continuous marketing work and advocacy is missing from XMPP, and is the one reason Matrix got where it is today.
No, this is not a question, this is an answer. At this point, XMPP is not a solution, but a basis for an actual coherent solution to be built on, but it does not exist yet (or anymore).
Matrix is also only a partial answer, because in practice you will direct users to matrix.org, and that makes matrix.org not a single point of failure, but still a huge point of failure and matrix will be down for most people if matrix.org goes down.