I agree, but I cut them some slack based on history. They started off using SMS for transport, encrypting everything in place and encrypting sent messages when the recipient also had Signal (Text Secure at the time, I think).
When they switched to using internet for encrypted messages, they kept SMS capabilities. I liked that they did that, because it let me use one app to communicate with everyone: encrypted with other Signal users, SMS with those using competing and incompatible products or just stuck on SMS.
They recently dropped SMS support, so now my contacts are fragmented and I still have to use SMS as a lowest common denominator unless I want to install multiple apps. And since I'm on Android, one of those apps is not possible to install.
Since dropping SMS, there is no reason to hang on to phone numbers, and they seem to be headed that direction.a
I understand all the reasoning. I don't understand why nobody seems to be building nice apps using interoperable protocols and encryption. (Well, I do understand the greed, whether for money, power, or fame, I just don't like it.)
I don't understand why nobody seems to be building nice apps using interoperable protocols and encryption. (Well, I do understand the greed, whether for money, power, or fame, I just don't like it.)
Just join the bunch of us using XMPP then, nothing to miss from Signal.
Please let us know how that goes! And in case you want the "phone number based" onboarding experience of Signal & al. on XMPP, you can recommend https://quicksy.im (I personally find it a terrible and short slighted idea to irrevocably associate one's online identity and presence to a phone number, but at least the option is there!)
I've been the phone number route, so now it time to try the right way. If I can get my son on board, my wife will follow and that's all who's left on Signal anyway.
Thanks. I'm working on a requirements document to help me figure out exactly what I want to do and how to accomplish that.
What I've got so far is leading me in the self-hosting direction. Scale is certainly part of it, but it seems likely that the best way for me to get more people on board is to set up a family server. That way it's not so much about "here's a new way to communicate" but "here's a small, private space to plan family reunions and keep everyone up-to-date on family events."
Also, I have some other projects that I've been putting off because of the need to figure out self-hosting. Running something for the extended family might be the push I need :)
Snikket looks like it checks off a few boxes. It looks pretty simple. It uses Docker (not that I care, exactly, but it seems that Docker is something I will have to come to terms with in general for other things on my self-hosting roadmap). The Android app is available through F-Droid, something that only I care about, but I do care about that.
So, finish fixing the snowblower (today, I hope), figure out how to get my OCI server running properly, then work on snikket. Maybe this time next week? 🤞
but it seems likely that the best way for me to get more people on board is to set up a family server. That way it's not so much about "here's a new way to communicate" but "here's a small, private space to plan family reunions and keep everyone up-to-date on family events."
This was a convincing argument in my case :)
Also, I have some other projects that I've been putting off because of the need to figure out self-hosting. Running something for the extended family might be the push I need :)
Good motivator at least! The downside/risk in your situation might be that you will expose them to quite a bit of downtime and instability initially. If you are new to this, you can probably start small, off of a raspberry pi for instance, and work out a way to always have a backup/working config on a "B" SD card while performing updates and maintenance on the "A" card (or even have two raspberries), it should help in times of doubt :) (you can just slide the other card and deal with it later)
Snikket looks like it checks off a few boxes. It looks pretty simple. It uses Docker (not that I care, exactly, but it seems that Docker is something I will have to come to terms with in general for other things on my self-hosting roadmap).
You don't necessarily have to use docker. I deploy my own instance (of ejabberd, another server software) without docker, and have been at it for about a decade now. My (hot?) take is that it trades a tiny bit of convenience at the start for more clunky tooling and abstraction getting in the way after. An XMPP server isn't that complicated to warrant docker (IMO). And you can get help from here, as I'm pretty sure many other admins deploy without docker and would be keen on helping you with that: https://prosody.im/discuss#chatroom.
https://chat.snikket.org/
The Android app is available through F-Droid, something that only I care about, but I do care about that.
As I wrote earlier, snikket is merely a packaged and preconfigured version of prosody & some popular clients, this is as generic and compatible an XMPP setup as you can think of. Any maintained XMPP client would work (but sticking with the ones branded as "snikket" guarantees that you won't accidentally pick an old/unmaintained one), for instance, in my case, I use Conversations and Siskin which you can find on Google/Apple stores, just so you know those are options as well.