What if someone designed a new email protocol/etc from scratch without any reference to the existing one. Could it be made better? Can you even define what better is?
That said, if we could redesign email, my quick wishlist would be:
* Better encryption. No surprises that a HN comment thinks we should encrypt metadata as well as contents, and have something better than the way we have to exchange GPG keys at the moment.
* Better security. Easier to guarantee that the ostensible sender of a message really did send it. Also, help prevent the viral problem of malware, which spreads through inboxes protected by rubbish passwords.
We started out intending to "fix" email with a better protocol, and ended up creating something much greater.
The core problems that we tried to solve (to help define "better") were:
- Security: encrypting and authenticating your messages
- Control: the sender hosts the message, so they can edit or delete them before the receiver sees them
- Flexibility: all messages are, in fact, just key-value stores, so any type of data can be sent on the protocol (whether it's Mail, IMs, Vines, Instagrams, or whatever the flavor of the week social network is)
I think it's a great product, and we are going to start the developer release of the client (Melange [2]) tomorrow.
I haven't seen any IM2000 implementations, so I'd love to check it out if you know of any.
Do you not like IM2000, or have other feelings towards it? I'd love to hear your feedback.
As another poster said, crypto everywhere and a permission system, from DNSSEC on down. Every assertion would carry a cryptographic chain of its authority.
Some kind of multi-user solution would be great. Mailing lists are a hack. To some extent NNTP filled this role though - maybe it's just a case of using the right protocol for the right job.
So yeah, it could definitely be made better - very few protocols are perfect. But it's done pretty well considering.
Like <embed> and <object> tags in HTML?