5 ms·

Doesn't this sort of abandon the trust model of PGP? If you make default encryption a part of the system and an elevator pitch bullet point, I don't think it's

Doesn't this sort of abandon the trust model of PGP? If you make default encryption a part of the system and an elevator pitch bullet point, I don't think it's really fair to call it simple because a UI change would be needed to ensure that though email may be encrypted, users can't necessarily trust the identity of the sender.

digi_owl 11y ago
And how will software ever help with that? Put a signed image of the device user at time of sending? You are pretty much walking into the same issue that big media has had with DRM. You can't trust something user facing to be doing the "right" thing unless the whole chain down to the individual chips are certified by entities you control. Right now email is more like sending post cards than letters, yet people treat them like the latter. encryption by default would at least make them behave like letters while crossing the networks. Figuring out if the sender is who he or she claims to be is a very different issue, and one that has plagued humanity since the dawn of correspondence.
- stinkytaco 11y ago
  My issue is not default encryption, which I think is a "good thing"™. It's that if you say "Thunderbird: Encrypted by Default"™, you give users a sense of security that is at least partially an illusion. This is something we would roundly criticize Google for if they were to say, for example, "your email and searching is secure because you are using SSL". I feel it's incumbent on an organization like Mozilla, one that I support at least partially because they are not building a platform to advance their own interests, to be very clear that email is encrypted en route, but not necessarily secure. I'm guess I'm viewing it from a communications angle.
  - digi_owl 11y ago
    Here we are back at the varying definition of "security". The encryption will defend against NSA style "grab everything, sort it out later" attacks. but it will not defend against targeted attacks where someone had compromised the remote sender somehow. The former is a boon for most people vs the present state, the latter is a hairy problem that has been with us since we first started sorting ourselves into ingroups and outgroups (how do you know that the email is not sent by someone with a gun pointed to their head?). Or as i was reminded by a different comment: perfect becomes the enemy of good.
pekk 11y ago
In the status quo nothing is verified or encrypted anyway. When we make the perfect the enemy of the good, we reinforce that status quo. As long as the trust model of PGP depends on everyone to get the same religion, we are stuck waiting for a messiah which will never come. If we have any hope of improving the status quo, we need some kind of bootstrap, and that bootstrap won't look exactly like the final result we want. Ubiquitous email encryption would improve on the status quo, and prepare the ground for further improvements. People who don't care would still be generating more verifiable information than the status quo. People who want to use encryption wouldn't be forced to drop it in order to talk to family and friends. People who want to carefully manage their identities and use signing to build a web of trust could build that on top of ubiquitous encryption. This would get us closer to the PGP trust model than anyone has been able to do purely by evangelizing the PGP trust model.
- digi_owl 11y ago
  > When we make the perfect the enemy of the good, we reinforce that status quo. Story of present day FOSS right there.
- anexprogrammer 11y ago
  After so many years of email that's entirely in the free and clear, I'd consider any steps nearer encryption as beneficial. Even if fundamentally weak improvements with annoucements littered with caveats for those in the know. I'd like for encrypted email to become a standard thing in my lifetime. We are no nearer than when I trialled PGP in the 90s. Then we can think about email that doesn't leak all your meta-data.

digi_owl 11y ago

And how will software ever help with that? Put a signed image of the device user at time of sending? You are pretty much walking into the same issue that big media has had with DRM. You can't trust something user facing to be doing the "right" thing unless the whole chain down to the individual chips are certified by entities you control. Right now email is more like sending post cards than letters, yet people treat them like the latter. encryption by default would at least make them behave like letters while crossing the networks. Figuring out if the sender is who he or she claims to be is a very different issue, and one that has plagued humanity since the dawn of correspondence.

stinkytaco 11y ago
My issue is not default encryption, which I think is a "good thing"™. It's that if you say "Thunderbird: Encrypted by Default"™, you give users a sense of security that is at least partially an illusion. This is something we would roundly criticize Google for if they were to say, for example, "your email and searching is secure because you are using SSL". I feel it's incumbent on an organization like Mozilla, one that I support at least partially because they are not building a platform to advance their own interests, to be very clear that email is encrypted en route, but not necessarily secure. I'm guess I'm viewing it from a communications angle.
- digi_owl 11y ago
  Here we are back at the varying definition of "security". The encryption will defend against NSA style "grab everything, sort it out later" attacks. but it will not defend against targeted attacks where someone had compromised the remote sender somehow. The former is a boon for most people vs the present state, the latter is a hairy problem that has been with us since we first started sorting ourselves into ingroups and outgroups (how do you know that the email is not sent by someone with a gun pointed to their head?). Or as i was reminded by a different comment: perfect becomes the enemy of good.

pekk 11y ago

In the status quo nothing is verified or encrypted anyway. When we make the perfect the enemy of the good, we reinforce that status quo. As long as the trust model of PGP depends on everyone to get the same religion, we are stuck waiting for a messiah which will never come. If we have any hope of improving the status quo, we need some kind of bootstrap, and that bootstrap won't look exactly like the final result we want. Ubiquitous email encryption would improve on the status quo, and prepare the ground for further improvements. People who don't care would still be generating more verifiable information than the status quo. People who want to use encryption wouldn't be forced to drop it in order to talk to family and friends. People who want to carefully manage their identities and use signing to build a web of trust could build that on top of ubiquitous encryption. This would get us closer to the PGP trust model than anyone has been able to do purely by evangelizing the PGP trust model.

digi_owl 11y ago
> When we make the perfect the enemy of the good, we reinforce that status quo. Story of present day FOSS right there.
anexprogrammer 11y ago
After so many years of email that's entirely in the free and clear, I'd consider any steps nearer encryption as beneficial. Even if fundamentally weak improvements with annoucements littered with caveats for those in the know. I'd like for encrypted email to become a standard thing in my lifetime. We are no nearer than when I trialled PGP in the 90s. Then we can think about email that doesn't leak all your meta-data.