4 ms·

I've been thinking about building one. I currently have too many other projects though, so this is going to be my next one, when I'm more or less satisfied with

I've been thinking about building one. I currently have too many other projects though, so this is going to be my next one, when I'm more or less satisfied with my ActivityPub server. It's so strange that no one has even tried to fill the niche for a modern desktop multi-protocol IM app. The one I have in mind would be properly-desktop though — multiple windows, 100% native controls, complete ignorance towards touchscreens, etc.

jeroenhd 4y ago
KDE has a built in chat client that seems to do the job most of the time. What's severely lacking there is client support. Because of it libraries for all kinds of services exist already. KDE Telepathy sounds pretty close to what you're describing (native (Linux) controls, multi window, no touch screen design, plenty of different protocols). With the KDE project slowly seeming to move towards Matrix rather than XMPP I wonder how long Telepathy will stick around. It can still be a pretty good start for your own chat application, though, as large portions of the work have already been done for you. An alternative approach I've seen people take is what I'm doing myself: Matrix/XMPP combined with various bridges. That way the chat client only needs to speak one specific protocol. It has its downsides but it does save you from having to develop abstractions over abstractions. You also get multi device support for free without having to scan tons of QR codes if you need to reinstall your client for whatever reason.
- grishka 4y ago
  I looked at the screenshots of KDE Telepathy and I'm not convinced. This kind of multi-protocol IM client exists for every major desktop OS — this and Pidgin for Linux, Adium for macOS, Miranda for Windows, etc. The problem with all these apps is that they're not modern. They do great for text messages, but that's it. Anything beyond text is an afterthought if implemented at all. What I mean by "modern" is first-class support for all the IM features people now take for granted: media attachments, server-side history with search, replies, mentions, reactions, forwarding, bot features, channels, etc. None of these existing apps even try to do any of this. This is what I want to fix with my future project.

jeroenhd 4y ago

KDE has a built in chat client that seems to do the job most of the time. What's severely lacking there is client support. Because of it libraries for all kinds of services exist already. KDE Telepathy sounds pretty close to what you're describing (native (Linux) controls, multi window, no touch screen design, plenty of different protocols). With the KDE project slowly seeming to move towards Matrix rather than XMPP I wonder how long Telepathy will stick around. It can still be a pretty good start for your own chat application, though, as large portions of the work have already been done for you. An alternative approach I've seen people take is what I'm doing myself: Matrix/XMPP combined with various bridges. That way the chat client only needs to speak one specific protocol. It has its downsides but it does save you from having to develop abstractions over abstractions. You also get multi device support for free without having to scan tons of QR codes if you need to reinstall your client for whatever reason.

grishka 4y ago
I looked at the screenshots of KDE Telepathy and I'm not convinced. This kind of multi-protocol IM client exists for every major desktop OS — this and Pidgin for Linux, Adium for macOS, Miranda for Windows, etc. The problem with all these apps is that they're not modern. They do great for text messages, but that's it. Anything beyond text is an afterthought if implemented at all. What I mean by "modern" is first-class support for all the IM features people now take for granted: media attachments, server-side history with search, replies, mentions, reactions, forwarding, bot features, channels, etc. None of these existing apps even try to do any of this. This is what I want to fix with my future project.