edit: Apparently alpha versions of NW.js now support running Chrome Apps. This could be interesting. https://groups.google.com/forum/#!msg/nwjs-general/YuwMHd_uv...
Why a chrome extension and not a desktop app? Because a desktop app is 3 desktop apps if you want to be cross platform, and since it is secure communication software, all of those edge cases that pop up in cross platform desktop application development really matter.
I think it's better that Signal takes its time and gets it right so that eventually we have a good solution. If Signal were to throw caution to the wind and hackathon up some desktop apps, then we may never have a single good option.
Signals competitors don't share it's security standards, and so it's not really reasonable to compare it to its competitors feature for feature. I expect that adding a given feature or other unit of complexity bears a higher cost for Signal than, say Whatsapp or Telegram. In the meantime, we still have the Signal mobile apps for situations where inconvenience isn't an insurmountable barrier.
[1] Signal-Browser doesn't seem to be able to add contacts properly when used with the production server, and the staging server looks like it's down right now.
[2] NW.js currently refuses to recognize Signal-Browser as a Chrome App unless I rename package.json to something else. Remote debugging doesn't seem to work with Chrome Apps running under NW.js at the moment - the inspector just gives me an empty response for each page I try to access. And there will need to be some way of configuring the Chromium engine to use Signal's self-signed SSL cert, though they'll have to solve this for the Chrome App as well.
Right now, Telegram suffers from the same faults (phone number = identity, closed/central server), but excels in usability and client availability. Signal is - for me, right here - worse. And I _should_ be part of Signal's target group.
I don't know your mail address or telephone number. If this message manages to reach you - can you explain your point a bit more?
If the private key would _be_ the identify, that'd be awesome. And maybe I fail to understand ChatSecure/Signal. I'd be glad to be corrected. But as far as I understand, that system ties a user to a mobile number, because 'that is as good a unique identifier as we get' and uses that instead. I think Threema does what you describe - or at least expects you to exchange keys via QR code when you physically meet?
My gripe with telephone numbers is this: I don't want to be tied to an identity I cannot control, to an identify that is public knowledge and unchangeable. I want to contact people via IM without them being able to call me.
Phone numbers are for calls (okay, texts for historical reasons).
On the bright side, there's now an open source fork called SMSSecure. As the name implies, it does encrypted SMS. It works pretty well. I just hope the open source maintainers are keeping up with security updates to the protocol and not introducing any new bugs...
I believe that is only temporary. The latest version of Signal hints at multi-device synchronization, though it appears to not be fully implemented. One example of this is the Chrome browser extension[0].
EDIT: Apparently they're bringing desktop and "web" through a Chrome extension and possibly a desktop browser wrapper. Also multi device support for desktops but apparently not yet for mobile.
Signal is getting way better, but Telegram is just a better messaging client at the moment.
