You have to have a killer feature for something to get adopted.

The killer feature of a new email protocol will be that it's not SMTP.

It should have sender authentication. E2E encryption that's easy and works by default. Those would be more than enough killer features, SMTP is just too broken and all the workarounds we have in place like DKIM, SPF, Spam ratings etc etc don't make up for it. We still have spam, important mails still end up in our junk boxes, and nobody trusts it enough anymore to put important content in emails. The war has long been lost.

"Nobody"

"War"

Few, outside small, likely familiar tech circles, use these terms when discussing email.

Not being SMTP usually means being another messaging island or other, which has only strengthened email.

Lists have huge value today, still!

Now, there is one exception: Marketing

The reason? Everyone else is busy getting work done.

Whatever may transcend email needs that quality, or it will, in fact be, yet another messaging island.

Exactly. There's no replacing email with something that's closed membership / invite-only etc. The whole point of e-mail is its your address where someone can initiate a contact, like phone number. You can choose to filter it etc. And you can choose to get attested emails which you can trust more etc. But fundamental nature is it is open for anyone to contact you. If you lose that, and build a custom new protocol, you are just building another closed wall messaging application.

Gmail has done this. I often follow up e-mails from my personal domain with a copy from a gmail account since it gets guessed as spam randomly.

Agree with everything you've said. Additionally, consistent rendering. If you've ever tried to send an email campaign and had to support all the old (and sometimes surprisingly new) Outlook clients and Lotus notes, you'll know just how painful that is.

Email must be: * Secure and encrypted; * Have proven identity; * Have easy to fabricate and predictable rendering;

I feel like the ability to have forms and charts is very nice, but adds a lot of complexity, especially from a security point of view. I'd be looking at this kind of "application level" functionally being a layer added optionally on top, not being in the core protocol.

As far as I know, and I could be wrong, ProtonMail wraps around pgp inbox encryption. Many providers have exactly this service already.

The difference seems to be that ProtonMail don’t allow you to use normal imap and your own client with pgp, but force you to use their client. This is probably a trade-off made to protect users against them selves in one way (disabling encryption for grandpa) and hiding the non tech-savvy from all the technical details.

Again, many assumptions here.

So, to me, mentioning proton here doesn’t make sense as the underlying tech: smtp & pgp have existed and been industry standard for a long time. So there is not issue of adoption.

I'm with thayne above on this one.. Protonmail just offers secure storage on their system but they still see the external incoming emails unless they're PGP encrypted. They make that easy, but still. PGP didn't take off in 30 years, it's just too complex. Tools can help that but the weak point is its key management.

Not the parent, but protonmail is still just more hacking to work around a broken system. It is a valient effort, but to really fix email the whole thing needs to be replaced.