cryptolect

In case nobody has pointed it out yet.

By default, with Ubuntu's dnscrypt-proxy package - the resolver is "cisco".

That's right, you're encrypting your DNS traffic just so Cisco can read it...

There are alternate resolvers but most people wouldn't change it out of the box.

I was really hopeful for a ChromeOS competitor. Guess I'll have to wait. Maybe Ubuntu's mobile OS?

Interesting to see that "end-to-end crypto" is becoming a frontpage feature for messaging apps.

Maybe an NFC implementation would be more acceptable.

Amusing to note that a site hosting a detailed article on the "cypherpunk revolution" doesn't support https.

Ever since I realized there will be a persistent log of my browsing history maintained by one or more government agencies, I have mentally-marked http as risky. I cannot wait for something like this to happen, and encourage the wider less-informed community that TLS is critical for both trust (properly implemented, TLS will provide protection against tampering and data leakage) and to provide a minimum level of privacy (I can see you visited site x, but not what sections of it). Bring it on.

Considering the functionality that wasn't reviewed, I don't think it should be advertised as "A".

It might be "A" for the expected permissions, but if there's extra permissions which haven't been investigated, these should be marked down until explained. This would encourage app owners to work with whoever maintains the database to get accurate ratings. If Whatsapp is an A today, there's no incentive for the developers to justify / assist with anything unexplained permissions remaining.

In it's current state, if it's awarding A scores while significant permissions are unexplained, it doesn't help the end user.

This is a wonderful example of NSA meddling. With one hand, Microsoft gives everyone out-of-box encryption, which it can use to demonstrate how well it's protecting consumers. With the other hand, by virtue of a 'feature' to assist consumers, it's providing access to the NSA via SkyDrive copies of encryption keys. Everyone's happy!

Best of all, enterprise customers don't have a reason to complain, because the SkyDrive backup 'feature' shouldn't apply to their deployment scenarios. The only people with a complain are those that use the default option.

We should keep vigilant for these security 'features' that are undermined by implementation. The NSA has years of practice at this, and we're playing catchup.

I'm happy to see that some new 'modern lisp on x' projects are taking design cues from Clojure (See Rhine and now Pixie).

There's a disconnect between the people who develop policy, and those who enact it. One group is elected, the other is not.

So subtlety and obfuscation to make good ideas sound bad (serpent), and bad ideas attractive (in the interest of performance and practicality).

So when any standards body recommends a particular cryptographic approach, we should ask why, and then why again.

I've been looking into ocaml, and was wondering what to do once my programs grow beyond a single file. Unfortunately it doesn't have the equivalent of 'lein new' (Clojure tool for creating a new, empty project skeleton), so I'd welcome a repository of best practice examples like skeleton.io.

To me, this explains why the five eyes nations are pushing for (meta)data retention legislation. It's a condition of participating in the scheme.