Programming Internet Email by David Wood

There are many types of electronic mail systems in the world today. Most of them never learned to share.

Most major computer vendors offer a proprietary email system. IBM, Digital Equipment Corporation, Microsoft, Hewlett-Packard, and others have their own ways of performing this simple task. Many of them are based on early versions of the X.400 protocol, an international standard for email. Unfortunately for those of us who live in the real world, X.400 defined the nature of compromise.^[1] In practice, some large corporations and governments mandated the use of X.400, but lack of vendor implementation has hampered its adoption.

While the International Standards Organization (ISO) was haggling over X.400, various vendors implemented their own email systems based on its formats and protocols. It is supposed that they believed it would be simple to migrate their systems to standards compliance, if the need arose. It didn’t. As the TCP/IP protocol and the Internet matured, so did its open mail system. When the consumer revolution hit the Internet in the early 1990s, the Internet way of messaging became a de facto standard literally overnight.

Technically, the Internet mail system is less integrated and complete than X.400. However, it is much more widely used and continuously evolving. It is also an open set of specifications, independent of a particular vendor. Any vendor may create and sell an implementation, of course, and many do just that. Reflecting the Internet’s early days in academia, however, most of the core Internet mail system runs on free software.

There really is no “Internet mail system” in the sense that one could map it, or even refer to a common set of standards. The Internet mail system is like the Internet itself, a collection of loosely connected cooperating agencies that are each running the software that they choose on the hardware that they choose. It only works because it is to everyone’s benefit to cooperate. This aspect, too, is continuously evolving. There are, of course, Internet standards for the core systems operating on the Internet itself. Some governments are currently proposing laws that would regulate which actions are appropriate to be undertaken on the Internet in general and the mail system in particular. These proposals reflect a view of the Internet as an important piece of public infrastructure and no longer simply an experiment.

This loose structure implies that any particular machine initiating an email message may not be compliant with Internet standards, nor may the recipient. Much of the documentation that comprises Internet mail standards concerns itself with translating messages to a canonical format for transfer over the Internet and translation again to the receiving format.

Similarly, sending and receiving machines may have different means of connecting to the Internet itself. The Internet mail system is equipped with the means of handling intermittent connections. In practice, individuals receive mail in many ways: from their local machine, a local area network (LAN) server, a server elsewhere on a wide area network (WAN), or virtual private network (VPN), or a server across the Internet. Even the program used to read and send mail (a “mail user agent” or MUA) is ill defined. An MUA may be a dedicated program with or without a graphical user interface (GUI) or even a Web browser! Custom scripts often act as MUAs to send, receive, and distribute mail.

There is order in the chaos. Internet standards define both the canonical forms and the means for creating gateways to other mail systems. This book describes canonical forms. Those interested in creating gateways will find helpful references in the appendixes and will still need the information in this book for the Internet-exposed portion of the gateway.