Electronic Messaging

In 1961, the Compatible Time Sharing System (CTSS) was a popular multiuser system, but lacked any features for allowing direct communication. Users of the system identified this need, however, and developed an ad-hoc system for sending messages to each other by writing files in a common directory with the recipient’s name. This was not secure - any user could read any mail messages, irrespective of their intended recipient.

The first documented proposal for a formal system for sending messages was Programming Staff Note 49 by Louis Pouzin, Glenda Schroeder, and Pat Crisman in 1964. This proposed a MAIL command for operators to notify users of events.

In 1965 Noel Morris and Tom Van Vleck implemented the MAIL command. Although the original proposal intended for it to be used by administrators to send messages to users, their implementation allowed any users to mail any other user. The MAIL program needed to run in a privileged mode to support this (equivalent of UNIX setuid). The functionality was similar to the original informal mail protocol. Each user had a mailbox file which contained their incoming mail. In the new system, this was a private file which only they could read. The MAIL command allowed arbitrary users to append messages to this file.

Another system used for communicating between users was the .SAVED command. This allowed users to write messages directly onto other users’ terminals. This was modified to provide notifications of new mail. When the MAIL command was used, it would first add the message to the recipient’s mailbox file and then write a message to their terminal (if they were logged in) to let them know that they had new mail. This was an important development over the older ad-hoc system, since it gave electronic mail delivery an immediacy not previously present. A user no longer had to periodically check their mail file to see if there were new messages.

The computers of this era were large machines with multiple users connected via terminals. They were very rarely connected to other machines. Mail was used in a similar way to internal mail in an organisation. It wasn’t until 1971 that these systems were expended to support networks. Ray Tomlinson modified the mail application to send messages to other machines via ARPANET, the system that would later evolve to become the Internet. This was the first use of the @ sign, used to signify user @ (at) computer.

The 1970s also saw the first instant messaging application. Emails were viewed as asynchronous messaging systems - replacements for postal mail where an immediate reply was not expected. The talk program, written for the PDP-11 and later ported to UNIX, split the screen of two users into two parts and allowed text typed by one user to appear at the bottom of their screen and the top of the other’s. Terminals of this era had very simple control mechanisms and talk was quite unreliable - two users tying at the same time could cause display corruption. Later terminals, with more advanced control mechanisms, made this kind of problem much more rare.

Electronic mail was originally a one-to-one communication system. In 1978, Ward Christensen introduced the Computerized Bulletin Board System (CBBS) in Chicago. This provided an online equivalent of the community notice board. Users would connect with their MODEMs, upload new files and collect those already there. The original system ran on a single MODEM, requiring users to connect one at a time. This kind of system gradually grew in popularity and eventually evolved into the kind of forums popular on the web today.

Bulletin board systems were very simple networks, typically consisting of a single server which clients would directly connect to. At the same time, ARPANET was evolving a peer-to-peer architecture with large university and government systems being connecting to a network. In 1979 Tom Truscott and Jim Ellis at Duke University implemented a form of message board using the UNIX to UNIX Copy Protocol (UUCP). This allowed users to post messages on a local machine which would then be relayed to other computers on the network. This became known as USENET, a contraction of user network. In 1985, UUCP would gradually be retired in favour of the Network News Transfer Protocol (NNTP), designed from scratch to efficiently propagate messages in this network.

In 1995, Deja News began maintaining an archive of the contents of USENET. Their index included posts from as early as 1981. This was controversial at the time, since many people had posted to USENET over the preceding decade with the expectation that their posts were transient.

Instant communication between computer users at this time was limited to the two conversing users supported by the talk program. In 1980, the Compuserve network launched an application they called a CB Simulator. In this period, increasing deregulation of the Citizens’ Band and the decreasing costs of radio equipment lead to increased popularity of CB radios and Compuserve aimed to exploit this. The CB Simulator introduced a simple many-to-many chat interface. The system was divided into channels, much like a real CB system (although they were identified by names rather than numbers). Text entered into one would be relayed to all other users in the same channel. This interface is still popular in more modern systems such as Internet Relay Chat (IRC) and Secure Internet Live Conferencing (SILC). In 1991, the CB Simulator would host the first online wedding.

Compuserve were responsible for two other innovations in electronic mail. In 1982, they introduced a bridge to Internet email, allowing Compuserve users to exchange email with Internet users. Prior to this, online service providers had maintained private email systems, where only subscribers to the system could exchange mails. This lead to people being forced to maintain several different email accounts in order to be able to communicate with everyone. The second significant development by Compuserve was the introduction of rich text email in 1992, with a WYSIWYG editor.

Instant messaging followed a similar pattern to email. The first system to allow instant messages to be exchanged was Quantum Link (which later became AOL), which allowed dial-up Commodore-64 users to exchange messages in 1985. In 1996 the Israeli company Mirabilis introduced ICQ (‘I seek you’) which brought instant messaging to the Internet. The proliferation of incompatible instant messaging systems lead to the development of the Jabber protocol in 2000. This was standardised by the Internet Engineering Task Force (IETF) as the Extensible Messaging and Presence Protocol (XMPP) in 2004. In 2005, Google introduced Google Talk, a commercial implementation of this standard.

David Chisnall

Further Reading
Tom Van Vleck’s History of Electronic Mail
Ray Tomlinson’s story of the first networked email