Facebook Kills PGP-Encrypted Emails | Hacker News Reader

Facebook Kills PGP-Encrypted Emails

techcrunch.com

40 pointsjbegley24 comments

PGP/GPG needs to die. It's too clunky, and that means it does not work the way people need it to work.

Then how do you encrypt your emails?

I don't think I know a single person or organisation who does... (and likely, that's a problem in itself)

Stat of a single person. I found a personal blog with a typo in the commands for one of the tutorials on the site. I found the blog owner's email and it had a pgp key on the keyservers. We chatted, securely, without ever having met before. It was quite a pleasant experience.

Actually in our organisation we use a custom open source solution built on top of pgp to transmit sensitive data across directorates. It works great.

It's this one https://github.com/spapas/etsd for anybody interested

> organisation

Everything related to megacorps is totally encrypted from you, starting from Intel ME and ending with going to a great length to take off a root rights from you on your "smart" phone. Only megacorp's useds do not appreciate privacy.

we have large healthcare networks as customers, some send provider data to our company in a pgp envelope.

That's been discussed: https://news.ycombinator.com/item?id=19173326

S/MIME.

Do you mean PGP/GPG the program(s)? PGP/GPG the file format (RFC 4880)? Or PGP/GPG's "web of trust" key distribution scheme (which may or may not include public keyservers)?

Regarding the three complaints:

1. If you hate GnuPG the program, there are several other implementations which are less "clunky." Have you tried them? Do they also need to die?

2. RFC 4880 was published in 2007. The state of the art of cryptography has advanced since then, much as the state of the art of email has advanced since RFC 822 was published in 1982. And just as RFC 822 has been updated over the years, there's an in-progress effort to update RFC 4880: https://datatracker.ietf.org/wg/openpgp/about/ Are you also calling for that working group to "die"?

3. Some poor implementation choices for the public keyservers made them unusable for a time. There's definitely room for something better—Keybase was great but understandably wasn't sustainable as a VC-funded company. I hardly think we need to throw GnuPG and RFC 4880 away because of particular keyserver implementation details. Nor do I think the keyserver failures mean the entire idea of "web of trust" is doomed.

IMO, most of the calls I've heard for PGP to "die" have conflated these issues, and proposed partial solutions which solve e.g. confusing gpg command-line flags, but don't address e.g. key distribution at all. It feels like when people say, "Gmail sucks, therefore email needs to die."

Why we are not all using email encryption and digital signatures is beyond me.

Worked on projects dealing with digitally signed/encrypted email. The main technical hurdles I always found were around key management:

1. How do you discover someone's encrypting key? There is no simple way and it involves understanding how it all works. Solutions using the DNS were proposed but never really workable or saw wide deployment (I tried). Gets more complex when someone updates keys.

2. Multi-device - not only do I have to mange keys, but now I have to make sure I have the same keys on my laptop mail agent and phone. That sucks. Using a web client also complicates this.

and a non-technical issue:

3. Large email provides and specialized service found things that sorta work enough. Every bank/medical office has a "secure messaging portal" that you can use that aren't as simple as email, but works well enough that there is no pressure to move to more standard email. Large providers like Microsoft have internal solutions to send secure email that works in their service and if it is free to get an account, it works (even if it's annoying).

     1. How do you discover someone's encrypting key?

They give it (the public key) to you or publish it where everyone can find it. You would keep and verify all your friends, colleagues, businesses, etc email..

These are all solvable problems. Much like voting. Someone does not want these problems solved.

> Someone does not want these problems solved.

I don't think it's a conspiracy; I think the people who don't want these problems solved are typical email users, for whom the increased security would only increase complexity without solving any actual problems they have with email.

All email users deal with the tyranny of email. They are just used to it. They don’t know how much better it could be.

I don't think that adding end-to-end encryption would make things meaningfully better for most of the people who use email.

Digital signatures would. Makes it real easy to sort your email.

If proton mail can do it so can everyone else. The ad-tech companies clearly don't want to.

How do you solve the problem of needing to pass an email address via phone? Happened to me dozens of times over the last year. Also, how do you solve the problem of shared inboxes where employees need to access emails and they may get replaced over time? How do you solve the problem of people just straight up losing or breaking their keys/laptops/phones? I knew people who had a new phone every few months because something happened to it.

The average user is very, very far away from being able to use a nearly irreplacible item for communication. And yes, it's irreplacible since there is no rollover mechanism in just publishing the public key.

Because it simply isn't ergonomic and the people who could do something about it don't care.

For instance, people were told to use sub-keys instead of their master key. When Thunderbird integrated PGP (effectively deprecating enigmail), sub-keys were not supported for a long time (are they now?). Thunderbird devs didn't really care that they just broke multiple people's workflow. That's how you butcher adoption.

I was pretty upset by this. It looks like it's still broken [1]. I stayed with the older version of Thunderbird for a while and recently upgraded by following their smartcard guide and it worked. I ran into some troubles that were resolved by installing a 32-bit version of Thunderbird. The issue was with GPGME, not clear to me whether it's a 64-bit library.

I got decryption and signing working with my Yubikey. Enigmail is still missed.

"Can I use an OpenPGP smartcard or a hardware token with Thunderbird 78?

Yes, we offer an optional mechanism. It requires that you install GnuPG and all required software yourself. Please refer to this document for more details: https://wiki.mozilla.org/Thunderbird:OpenPGP:Smartcards"

It just blows my mind they gimped a feature so bad and after so many years still havn't reached parity.

[1] https://support.mozilla.org/en-US/kb/openpgp-thunderbird-how...

I wager because of Google. They have have positioned themselves to see the majority of e-mails and can't allow E2EE.

Email encryption is badly broken and cannot be fixed. It shouldn't be encouraged, users should migrate to a ratchet encrypted messaging service that doesn't have a million different gotchas and potential side-channels for exfiltrating messages.

'twas only a matter of time

Reply on news.ycombinator.com