It seems to me that the main problem with email today is the fact that messages are "pushed" to the recipient's mail server. That worked fine in the early days of the Internet when there were few email users and most of them were legitimate. But things have changed, and so should the way email servers work.
The solution to spam seems clear to me, and it requires little more than upgrading mail servers and email user agents. Suppose Alice wants to send Bob an email. Alice composes her message and sends it to her mail server where it sits a while. Her mail server sends a simple message header to Bob's mail server. Bob checks his email and downloads the new message header from his mail server (that's just the header; none of this preview-pane stuff where the message automatically appears). Bob has two choices: (1) he decides he knows Alice and wants to read the message, or (2) he suspects the message is junk. In the case of (1), Bob double-clicks the new message and his user agent sends a request to his mail server to go get the message from Alice's mail server. His mail server happily retrieves the message and forwards it to his user agent. In the case of (2), Bob deletes the junk message's header and continues with his business.
Note that Bob notices little difference in his new user agent compared to his old one, except for the no-automatically-displaying-messages-in-the-previe w-pane thing. But underneath, there are a lot more advantages to the new system. For starters, if Alice is a spammer, her junk message takes up space on *her* mail server instead of Bob's. A mass-spammer immediately has a storage constraint on her side, because if no one retrieves her spam, it continues to sit in her server. Hence, less spam can be sent out in the first place. Another advantage is that there is a drastic reduction in wasted bandwidth if Alice is a spammer, because the entire junk email doesn't automatically get sent through the Internet, only the simple header does (assuming Alice is "close" to her mail server).
This system is kind of like the plain-old telephone system. With a telephone, when you hear the ring and/or check your caller ID unit, you make a decision to answer and retrieve the rest of the message. The person on the other end cannot just start talking to you after they dial your number (kind of like they can with email). This is also how email should work. Given a simple header ("ring, ring"), you decide whether or not to continue. The sender, meanwhile, must wait for your acknowledgment.
Why has something like this not been implemented yet?
The author's analogy between Henry Ford and Shawn Fanning doesn't quite hold. Ford legitimately used the assembly line idea to provide a commodity. He didn't create a skeleton key to the "Big Business"'s warehouses full of automobiles.
For a proper analogy between Ford and Fanning, Fanning would have had to create his own music production company, get artists to sign with him, and sell his CDs for $5. That would be the legitimate way to do it. Instead, he chose to give consumers the skeleton key.
Slashdot Mirror
Use this to get back to classic mode.
It seems to me that the main problem with email today is the fact that messages are "pushed" to the recipient's mail server. That worked fine in the early days of the Internet when there were few email users and most of them were legitimate. But things have changed, and so should the way email servers work.
e w-pane thing. But underneath, there are a lot more advantages to the new system. For starters, if Alice is a spammer, her junk message takes up space on *her* mail server instead of Bob's. A mass-spammer immediately has a storage constraint on her side, because if no one retrieves her spam, it continues to sit in her server. Hence, less spam can be sent out in the first place. Another advantage is that there is a drastic reduction in wasted bandwidth if Alice is a spammer, because the entire junk email doesn't automatically get sent through the Internet, only the simple header does (assuming Alice is "close" to her mail server).
The solution to spam seems clear to me, and it requires little more than upgrading mail servers and email user agents. Suppose Alice wants to send Bob an email. Alice composes her message and sends it to her mail server where it sits a while. Her mail server sends a simple message header to Bob's mail server. Bob checks his email and downloads the new message header from his mail server (that's just the header; none of this preview-pane stuff where the message automatically appears). Bob has two choices: (1) he decides he knows Alice and wants to read the message, or (2) he suspects the message is junk. In the case of (1), Bob double-clicks the new message and his user agent sends a request to his mail server to go get the message from Alice's mail server. His mail server happily retrieves the message and forwards it to his user agent. In the case of (2), Bob deletes the junk message's header and continues with his business.
Note that Bob notices little difference in his new user agent compared to his old one, except for the no-automatically-displaying-messages-in-the-previ
This system is kind of like the plain-old telephone system. With a telephone, when you hear the ring and/or check your caller ID unit, you make a decision to answer and retrieve the rest of the message. The person on the other end cannot just start talking to you after they dial your number (kind of like they can with email). This is also how email should work. Given a simple header ("ring, ring"), you decide whether or not to continue. The sender, meanwhile, must wait for your acknowledgment.
Why has something like this not been implemented yet?
The author's analogy between Henry Ford and Shawn Fanning doesn't quite hold.
Ford legitimately used the assembly line idea to provide a commodity. He didn't
create a skeleton key to the "Big Business"'s warehouses full of automobiles.
For a proper analogy between Ford and Fanning, Fanning would have had to create
his own music production company, get artists to sign with him, and sell his CDs
for $5. That would be the legitimate way to do it. Instead, he chose to give
consumers the skeleton key.