Privacy.

If someone sends me spam, I want the ability to easily reject any further mail from either the specific address or the entire domain. It would be nice if this applied universally, on every device. Basically, a user controlled spam filter at the server.

Also, a standardized form of encryption --- to be easily invoked as needed for sensitive data. 98 percent of email doesn't need encryption --- but that last 2 percent does.

Basically, privacy invasion via email needs to be eliminated.

This seems reasonable, and even achievable. It's probably already possible to reject emails that don't have SPF and DKIM (to prevent spoofing), so we just need a way for a sender to determine whether the recipient has whitelisted them. If they haven't been whitelisted, they should be limited to sending a short (100 character?) plaintext introduction, with no hyperlinks.

> to be invoked as needed for sensitive data. 98 percent of email doesn't need encryption --- but that last 2 percent does.

I disagree slightly. The 2 percent that does need encryption also requires that the other 98 percent be encrypted too, so that:

* attackers don't know which messages to target,

* the pattern of encrypted vs. unencrypted emails doesn't reveal information in itself, and

* the user doesn't have the mental overhead (and error potential) of having to decide which emails are worthy of encryption (a task which requires them being able to reliably predict the future).

Much of the logic for encrypting the whole of the web (as opposed to just "important" pages) applies to the need to secure (almost) all email too. The main difference, I think, is that the web is a pull architecture while email is push, meaning that spam and identity works differently, and server-side filtering can't work with E2E encrypted email.

Revealing my email as a potential target is a minor trade off I would accept for ease of use. It should not be a big deal if the encryption is solid. The cost to try and break the encryption is enough to discourage most attackers.

How many email "man-in-the-middle" privacy breeches are you aware of? The ones I am aware of usually happen at a service provider endpoint --- like Google data mining your email. I doubt even Google is motivated enough to absorb the cost of breaking email encryption. There would also be legal ramifications if they (or any other provider) did.

But that's the problem, how does a decision-fatigued user decide what is and isn't sensitive? People have lost jobs over tweets they sent years ago, and we know that users get used to clicking through warnings and find it difficult to tell whether something is secure or not. Dutch Jews filling in the 1930 census form didn't realise that stating their religious affiliation was going to prove a disastrous decision a few years later.

> How many email "man-in-the-middle" privacy breeches are you aware of?

There are all sorts of targeted security attacks and privacy breaches I'm not personally aware of, but protecting all email from potential snoops is the right thing to do regardless of how often they happen.

A totally private (white list only, friends and family) email address would be very attractive to a lot of people.

Lots of people would want 2 email addresses with varying levels of restriction --- one public and one private. I do this now but my private address is lacking any strict control to keep it this way.

If that were true, mail providers would already offer this. In fact, I suspect some unsophisticated ones probably do, with all unrecognised senders having their emails sent to the recipient's Spam folder, with a few exceptions for DKIM-signed mail from frequently seen addresses, like Amazon receipts.

Appeal to Authority?

If people truly had any interest in an email replacement, mail providers would already offer this. So why are you asking your original question?

https://www.thoughtco.com/logical-fallacies-appeal-to-author...

No, appeal to the market. If there is demand for a product in an interoperable market, then one would expect an entrepreneur to create that product. (Admittedly, the unusual economics of email services mean that market forces might differ from theoretical models, but there could still be some pressure for competing providers to add features that users want).

> If people truly had any interest in an email replacement, mail providers would already offer this.

There are dozens of replacements for (some aspects of) email, including WhatsApp and Twitter. To be fair, these services do (I think) have mechanisms that effectively limit your interaction to just people you already know, but the fact that email still exists (and doesn't have this restriction) suggests that there is value in the openness that it offers.

As examples, consider Delta Chat and the Autocrypt specification it's built on.

https://delta.chat/en/

https://autocrypt.org/

Private (figure a way to prevent pixel tracking, etc)

Trusted (figure out how to prevent unsolicited spam - opt-in “friends only” model?)

Overcome file size and type sending/sharing roadblocks

Overcome embedded malware issues

Treat groups the same as individuals at the delivery/trust level

Instant delivery

Easy import/export/transfer/backup/cleanup

First class integration with calendaring

why? If I may ask. Does that not violate end-to-end encryption? Or, I think what you meant is from a delivery perspective, not trust? As in, if you accept messages from a sender, the whole group's messages are also delivered going forward? Or do you mean for authentication?

If I’m trying to remember the name of somebody I talked to years ago, I should have an option to see the first names of everybody who I ever communicated with.

If I’m trying to find an old receipt, I shouldn’t have to search through receipts for half an hour to find out that the word I should have looked for is invoice instead: it should be smart enough to sort through all similar data.

If an email mentions a location, it should automatically be marked in a map. Etc.

I don’t use it myself but hey.com has some interesting and sensible quality of life functions. https://www.hey.com/features/

Depending on how it's implemented, I'm sure there are security issues to consider, but I'm not following how this kind of feature would necessarily cause a breaking encryption problem. If you can add some thoughts on how this breaks encryption, that might be interesting.

> Would you feel ok with having a database on each mail client that caches all text-only content for sorting/searching locally? Let's say a 1GB disk-space allocation for this feature alone ?

Given that modern HDs in most professional machines are in the 1TB and way up range, I'd say that anybody who wants to use the full power of email would probably consider that tradeoff worthwhile.

Maybe nobody wants that though.

To get providers to agree to this system, we could say that all domains registered before 2022 are grandfathered in, and only newly registered domains (with an MX record) have to put money into escrow. Domain owners already have to spend money, and hire technically knowledgeable people, so it's not like millions of ordinary email users will have to suddenly start entering their credit card details in order to send messages.

The only problem is how to decide who is allowed to make a claim against the escrow. If I accuse Gmail of sending me spam, can I take Google's money? To avoid some bad incentives, the money in escrow should go to charity rather than the claimant, but that still allows attackers to make false claims and prevent millions of Gmail users being able to send emails.

A naive voting system would break down because of Sybil attacks, so we probably end up having to trust some cabal of big email providers, who maybe become anchor points in a trust graph, or they employ people to review instances of DKIM-signed emails that claimants report to them as being spam. This cabal could perhaps be made slightly more democratic by allowing UN/ITU members to vote on organisations they trust with this power.