They're singularly focused on the user experience and what users want.
Uninstall / reinstall / multiple-devices works flawlessly.
So you can use regular texting for everything you don't care much about, like sending youtube videos and memes to your friends, and use the secret chat to things that re more sensitive. That's great for most people that currently use Whatsapp
* Whatsapp is closed source, so you can't tell if the encryption is true anyway (what if they are transmitting your private key?). Telegram has open source implementation and API.
* Facebook will not have your metadata. The subject of these events is basically FB having too much of data about you.
I would expect a Signal implementation of this to allow recovering chat history if you restore a backup onto the new device and if you re-enter your PIN.
(Also, your Signal PIN can be an alphanumeric passphrase.)
I don't either, but a chat app is useless if I can't use it to chat with people and almost nobody I know will consider Signal. It's either WhatsApp, Telegram, or SMS (horrible).
But you're right, Signal is designed for secure communication between parties. Allowing messages to be exported or archived should at worst leave only your own messages (and remove all others including media) and at best not exist at all.
Sure, let me just single handedly backup our conversation and send it to the HK police...
I doubt my mom will send my 'good morning', 'good night' and 'love you' to the HK police.
If you are communicating with people you don't know/trust, then yeah, you want expiring messages that are not exportable
This is the core problem as why many projects don't get mainstream. They have 2 options: they can focus on what they think is a priority, or on what the public thinks is a priority.
I'm not saying Signal is wrong on doing what they are doing, as they are being successful among some niches (i.e: tech). But to grab Whatsapp users, they need feature and UX parity, at least to some level