Where the @ Comes From, In Modern e-mail Addresses

A long, long time ago, individual programmers connected to time-sharing computers using terminals, and each so-called account on the time-sharing computer already had a username. And so a need arose, for people to send messages to each other, on the same computer, because those users were not physically located where the computer was located. The computer would typically be located at a remote, secure location, while the user would be located in a terminal-room, without the ability to speak to the other user face-to-face, but somehow requiring the cooperation of the other user.

If this was taking place on a UNIX system, then we considered ourselves to be privileged, because when UNIX was first developed, it was considered advanced, and we could actually give a command such as:


mail dirk


To send a message to the user named ‘dirk’.

Eventually, multiple computers got into the act – i.e. computers existed on small networks. And when I open a terminal-session on any of my Linux computers, I get a command-prompt something like this:




This command-prompt means that I’m user ‘dirk’ but on the computer named ‘Phoenix’. If each Linux computer on my LAN still had legacy packages installed such as ‘sendmail’ or ‘postfix’, then I could type in the command instead, that would read:


mail dirk@Klystron


Which would tell the computer, I wanted to send a message to another computer on the same LAN.

The fact that the characters which follow the ‘@’ form a Fully-Qualified Domain Name – an FQDN – only really started to exist, once the Internet had started to spread.

Now, the way it is on modern Linux systems, we no longer have the ‘sendmail’ utility installed by default, so that we can no longer send each other emails from the command-line. Like the rest of the world, we will need to open a full email-client, to do so instead. And for users who don’t like the fancier GUIs, there is also an email client named ‘mutt’, which allows for emails to be sent from an ASCII-representation, even with no X-server running.

Users who do this are considered to be something of an oddity, much like users who use ‘Lynx’ as their non-graphical Web-browser. But because some legacy software still exists, which would like to be installed on legacy UNIX, we have a package named ‘lsb-invalid-mta’, which we can install, and which provides a superficial appearance of the old ‘mail’ utility being available. But if we ever try to send an actual message or email with this, we will just get an error-message every time.

OTOH, If we wanted to expand our configuration to such a degree, that we could send an actual email using ‘mail’, effectively, we need to install ‘postfix’ as an exclusive alternative either to ‘lsb-invalid-mta’ or to ‘sendmail’. But I suspect that many users would consider this to be a security risk, because then, any application on our local machine could start sending emails, even if those just had user-status – i.e. without root – because it was the user in question, who used the ‘mail’ command to begin with. In the case of email, we’d give ‘postfix’ the (secured) login credentials to an actual, external SMTP server, over which all our locally-generated emails would go out. I think that most Linux users are slightly more-at-ease, knowing that their regular, user-applications, do not have the privileges to do this.