Sending Email only takes an Email client, or even a manual SMTP session typed in by hand with telnet(1). That's because many MTAs do not require any authentication to send. Your "From" and optional "Reply to" addresses are not required to be real.
Receiving Email adds only the need for an Email account on an MTA somewhere, and some Email client software.
Quote:
I'm starting to get this.
|
Excellent. Here's some deep background:
Unix-based Email, which became Internet Email, was in use many years before there was an Internet; the connections were point-to-point, using UUCP.
You had to know your own routing and apply it in the Email address. The "bang path" listed the adjacent computers and the user account.
(example: machine1!machine2!machine3!username).
Email was store-and-forward. Once
machine2 acknowledged receipt of a message from
machine1, responsibility for the message shifted, and
machine1 would delete the message from its storage.
In modern times, Email doesn't take a circuitous path, and you don't have to look at anyone's "bang path" Email address to find a machine mentioned along it that you've heard of and know how to route to.
This means to have an MTA to receive mail, you need to publish MX records so that the Internet can reach it, and to send, you'll likely need to send from a static IP address. I do this with dyn.com (used to be dyndns.org) for DNS and their mailhop.org MTA services; it's cheaper than hosting a dedicated MTA via a service provider -- even on a virtual machine in the cloud.