Comment by jgg | Hacker News Reader

Comment by jgg | Hacker News Reader

Parent Full thread

I voted you up due to sentiment, and I don't mean to sound like a grumpy member of the Cabal of Crypto Criticizers, but what's the point in Nadim developing the application if he won't respond well to honest peer review and wakes up every day and decides to fuck around with the basic structure of the application, and not only that, but makes fatal flaws when doing so?

I'm all for learning and experimentation, but not when some dude from Syria is literally Tweeting you and saying, "Hey, thanks!", which is the point of the article. I think end-point spying is probably bad enough at this point that we really don't need a broken protocol too. I don't really see the point in constantly changing out crypto primitives like he does, and it has introduced major security vulnerabilities in the past. He should find something, stick with it and then get it audited and reviewed, and then I think less people would bitch.

I support the goal of making crypto more usable.

I think you responded to my comment that I somehow managed to delete, but re-posted here: https://news.ycombinator.com/item?id=6990738

I agree with you. I don't think Cryptocat or how the implementation is handled is perfect by any means, but people have been complaining about that for a long time, and we still don't have good alternatives. What I'm saying is that if somebody makes something that is technically sound and is as attractive for regular users, in Syria and elsewhere, then this whole debate becomes relatively moot.

If Nadim is as unwilling to cooperate as you're implying (which I don't think is totally true, but I will grant you that the underlying constructions for the group chat component were switched out recklessly and ignorantly), then surely complaining about it in blog posts will be very ineffective.

I don't think he's unwilling to cooperate, and I don't mean to pick on the guy, because he's basically become a whipping boy. I think the reason he reacts so adversely to the criticism is the obvious use of him by bloggers and "crypto pundits" to bolster themselves. It's not constructive, and I don't wish to encourage that.

That said, security issues are still security issues, and my tl;dr as someone no one on this site cares about is for Nadim to just stick with a set of crypto primitives and protocol design that work, fix the problems that arise and stay there until he's more confident in what he's doing.

It would be useful to get the opinion of the Crypto Cabal on what is good safe software that is accessible to the general public and deserves to be promoted and well marketed?

Any suggestions?

The best general answer to this question is probably "Use Tails: https://tails.boum.org/ ." Especially for activists in Syria/regular users who have a genuine concern for their lives.

(Tails uses Pidgin with the OTR plugin.)

Things that use OpenPGP or libOTR. At the level of an organization where you can run your own CA, X509 might be easier (it's integrated into outlook IIRC).

I don't know what the accessible frontends are and no doubt there's work to be done there, but the basic primitives are a solved problem, and I'm pretty sure a better frontend on top of either would be very welcome.

I'm not a member of any "Crypto Cabal", and I use the term sardonically, directed at the people people who come out of the woodwork to pimp/gratify themselves and their businesses by posting bullshit about crypto, especially on HN (e.g., the "USE BCRYPT USE BCRYPT USE BCRYPT" guy who claimed that the problem with an RSA exponent of 1 was that "1 is a prime number" (?), or ironic articles of the "I JUST LEARNED ABOUT [crypto topic] SO DON'T EVEN THINK ABOUT ENCRYPTION" variety).

My opinion, as someone who is not important, is that most crypto software is bad, and most software that is fun to use has bad crypto. The Silent Circle stuff looks good, as does the Whisper Systems stuff, and I personally use Pidgin + OTR, which is crap from a UI standpoint.

I totally understand the design/UI motivations behind Cryptocat, but IMO Nadim needs to stick with a protocol design and crypto primitives that work, fix any flaws and then leave it alone until he's more comfortable (perhaps he's done that already).

Reply on news.ycombinator.com