This propaganda about good guys backdoors being impossible again. This is cryptographic bullshit of the highest degree. We've had the DUAL_EC scandal, for once, as an example of NSA backdoor which as far as was proved, only the NSA could crack. And with the fact that the NSA had so many bad leaks, yet still everyone except the NSA can't crack it, proved that backdoors are not only possible, but were going on behind our backs.
Don't get me wrong - I'm completely against backdoors. But when you shift the argument into "we won't do it because it's impossible", you're already agreeing that it should be done, while your argument won't hold because it is in fact possible.
There's three kinds of people:
1. Non-technical people (theresea may) that want backdoors are don't care about whether it's possible or not.
2. Technical users, with a vague general knowledge of cryptography, and the imprinted thumb rule of "backdoors are bad"
3. People with actual knowledge in cryptography which had already been doing research about why it is possible for years. Just a teaser: https://en.wikipedia.org/wiki/Kleptography
Of course, the real issue would be the scale and the distribution of access to the backdoor to various agencies.
>And with the fact that the NSA had so many bad leaks, yet still everyone except the NSA can't crack it, proved that backdoors are not only possible, but were going on behind our backs.
How do you know that no other intelligence services have broken it? You don't show someone you've broken their ciphers unless you have to.
This is not disputed. What is disputed is that there is some safe way of doing it
> but were going on behind our backs...
A number of people outside of the NSA had identified the possibility of a backdoor. It seems some decided to overlook it or felt powerless to do anything. Indeed some others patented it, thereby making it public. Maybe the $10M grant affected some peoples judgement
> only the NSA could crack
If I understand correctly the flaw weakened the selection of key parameters in a way that made it easier to brute force, by dramatically reducing the possible start points. If another nation state had brought their resources to this, maybe they could have exploited it and, most likely, they wouldn't have made that public. Moreover the Juniper scandal (mentioned at the end of the same article which was rather convenient) suggests that a third party did indeed make use of this backdoor.
> the real issue would be the scale and the distribution of access to the backdoor to various agencies.
Which was a point discussed in the original article at some length
Also, when you don't fight the real issue, which is that we shouldn't want backdoors in the first place, all those agencies can offer their backdoor to the decision makers (maybe it even happened with DUAL_EC - who knows? maybe already happening as we speak), and when all decision makers hear is "impossible" and "impossible", but never "don't do that", they will silently agree and you wouldn't know a damn thing about it.
I can also imagine a front-door approach where they legalize encryption only if you let the the government cosign the message with their own private key, say, through a web service. Not saying that's practical but it is possible.
I just posted a similar, if much less comprehensive comment. I'm troubled you've been downvoted for this.
Your second point in particular is highly significant. In general, "it's impossible" should be avoided as an argument when one's true objections are ethical. You don't want to be left standing naked because of some clever engineering.
It depends what you mean by "it's impossible" when you are discussing 'good-guy backdoors'. I don't think the article was saying its impossible just because it is technologically hard. The article discussed at length that what was impossible is that the keys would be guaranteed to stay in the hands of good guys, and that the good-guys would always be good.
I think it is perfectly reasonable to say that that is impossible.
I couldn't agree more. Saying that it's impossible to backdoor encryption in a "secure" (so long as you trust the government holding the 3rd key) is a very dangerous stance to take. It just sends this message to people who might campaign against backdoors that we don't need to bother - after all it's impossible.
The NSA was smart. They weren't advocating banning encryption, but what they did was weaken cryptography by tricking everyone into using ellptic curve cryptography with their EC.
Don't get me wrong - I'm completely against backdoors. But when you shift the argument into "we won't do it because it's impossible", you're already agreeing that it should be done, while your argument won't hold because it is in fact possible.
There's three kinds of people: 1. Non-technical people (theresea may) that want backdoors are don't care about whether it's possible or not. 2. Technical users, with a vague general knowledge of cryptography, and the imprinted thumb rule of "backdoors are bad" 3. People with actual knowledge in cryptography which had already been doing research about why it is possible for years. Just a teaser: https://en.wikipedia.org/wiki/Kleptography
Of course, the real issue would be the scale and the distribution of access to the backdoor to various agencies.