But wait – it gets much, much worse
As I was finishing up the above post, I noticed something a little strange in the code – something I’d glossed over earlier. There are a ton of references to what looks to be functions related to Google’s #Firestore database.
“Popular,” and even “ease of use,” are not relevant for the label of Gold Standard when we’re talking about security. Functionality for purpose is relevant, but if we’re allowing for weaker security in trade for ease of use then I’d say just use SMS; sure, it’s not as secure as Signal, but it’s a lot easier.
Reductio ad absurdum aside, there are by my count about a half-dozen systems which are more secure than Signal. Systems which don’t require you to give up your phone number, or publish it, or leak other personal metadata. You mentioned one, Briar, and there’s SimpleX Chat, Tox, and Jami (the latter two have been around for a few years, and IIRC Jami’s been audited). There are any number of apps (web and mobile) that claim encryption and anonymity such as Confide, Onion Chat, ChatS, Speek!, Peekno, and Threema. Ocelot and retroshare.io are peer-to-peer with no central servers, and are probably (metadata) secure.
I wouldn’t call any of these individually the gold standard, but several are obviously more secure than Signal.
I can’t get over how any system that required such a tracable and abusable piece of PII as a cell phone number could be considered the gold standard for privacy.
First, ease of use is absolutely relevant when it comes to security. If it’s too technical, difficult, or confusing, nobody will use it. Just look at how prevalent PGP is in emails - it’s basically doesn’t exist outside of niche nerd circles. What percentage of Linux admins ever deal with SELinux before getting told to just us AppArmor because it’s easier? So yes, ease of use is a factor.
Second, ‘security’ is too broad a topic. I don’t see a point in debating what is “the best” if a threat model isn’t outlined first.
I originally stated “Signal is the gold standard for encrypted private messaging”, which stands true regardless of other security features because it defaults to end-to-end encryption for everything by default and works out of the box. At the end of the day your messages are guaranteed to be encrypted and private - anonymity is not in the equation.
That said, I did bring up the point about leaking metadata, but looking at SimpleX I see that even they claim [0]:
So, without digging much into it, it seems there’s some limitations to your claims about SimpleX’s superiority to Signal in terms of even anonymity.
I tried it when it was called Ring, tried it again sometime after the name change. It’s a P2P messenger that provides E2EE. The architecture means all metadata leaks to ISPs and the internet. So you should be using it with Tor (or some other layer), and because your contacts also need to do that, and one of them is bound to fuck up, it’s better to use either something that’s metadata-resistant by default (like Briar) or to stick to Signal. Also, because its P2P, it requires both parties to be online to even work - at least last I tried it. This doesn’t work in the modern world.
Without getting into the various security issues over the years (here are two recent ones [3] [4], one which allowed remote code execution!), the Android client is spartan to say the least, and there’s no iOS client [1], making this unusable with half the people I’d like to communicate with in the US. Your regional mileage may vary [2].
Isn’t even open source so completely out of the question - security through obscurity, as the story post about the Converso apps proves, cannot be trusted.
I’ll skip the rest as I’ve already spent too much time on this, but I will say I do believe Threema might be as good if not better than Signal, but it’s a paid app and it’s hard enough to convince friends/family to get onboard with a free app, never mind something that requires payment.
[0] https://github.com/simplex-chat/simplexmq/blob/stable/protocol/overview-tjr.md#trust-in-servers
[1] https://tox.chat/clients.html
[2] https://www.statista.com/statistics/236550/percentage-of-us-population-that-own-a-iphone-smartphone/
[3] https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2021-44847
[4] https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2018-25022
Oh, I didn’t intend to skip the Tox comments. I haven’t used that in a whole, and was unaware of the CVEs. Those, and the fact there’s no iOS app, are good reasons to not use it. I found its use of DHT limited its performance and often had device battery life impacts; it still had a better protocol than Signal. The CVEs and other issues are technical implementation problems that can be fixed, unlike Signal’s design flaws.
Confide was just an example of a new class of fully anonymous, ephemeral chat clients, and maybe not the best choice. There are a half-dozen of these, all using similar mechanisms, some of which are OSS. I need to do a deeper survey of these, because they’re an interesting new approach to full-security chat.
Anyway, just saying I hit "sendx prematurely.