Summary

  • DEC marketer Gary Thuerk’s team used a printed ARPANET directory to assemble recipients for a message promoting demonstrations of the DECSYSTEM-20. The surviving message and later accounts show the practical distance between having addresses and having a purpose-built mailing list.
  • SNDMSG’s field limits turned that distinction into visible evidence: addresses overflowed through message fields, some recipients were missed, and a second send followed. DCA objected under the network’s official-use rule; recipients debated both the burden and the message’s relevance.

The header carries two different clocks

The surviving DEC message is usually introduced as a date in a timeline. Its own record is less tidy. The header says it was written on 1 May 1978 at 12:33 Eastern time. The envelope reproduced in Brad Templeton’s archive says it was received at DEC-MARLBORO on 3 May at 09:55 Pacific time. Templeton’s later account says Carl Gartley sent it on 3 May; a Defense Communications Agency notice quoted in the same archive refers to the mailing on 2 May. The documents do not resolve the discrepancy. They show a prepared message, a later delivery, and an administrative response—not one perfectly agreed timestamp. The archived message and replies are more informative than a date label alone.

The sender was Gary Thuerk, a marketer at Digital Equipment Corporation. DEC wanted to introduce its DECSYSTEM-20 family and its ARPANET protocol support to people on the West Coast, where it had fewer established contacts than on the East Coast. Thuerk later told Templeton that the plan was to invite potential customers to product demonstrations. The message’s purpose was commercial, but its audience consisted largely of people connected to a research network with its own rules of use.

A directory supplied names, not a distribution service

In the late 1970s, ARPANET addresses could be found in a printed directory. Templeton’s retrospective says Thuerk and his technical associate gathered West Coast entries from it, then added some customer and ARPA contacts elsewhere. That directory made the network’s participants legible to one another. It did not, by itself, create a mailing list with a subscription boundary, an owner responsible for maintaining it, or a rule that its entries could be used for sales outreach.

The distinction became tangible in the mail software. Templeton recounts that the TOPS-20 mail program could accept 320 addresses. Einar Stefferud, a recipient whose explanatory note is preserved in the same archive, describes addresses being typed through the Subject, To and Cc fields and then spilling into the body. Templeton says the remaining recipients were handled in a later send after DEC learned that some intended recipients had not received the first message. These are retrospective accounts, not a machine log of every destination.

Together they show that the operation was assembled by hand inside a message composer, not submitted to a mature bulk-mail service.

That matters because the interface exposed a design boundary. The program could carry many destinations, but it did not make that operation invisible, orderly or self-limiting. The recipient list became part of the message itself. A public directory made addresses discoverable; a sender’s manual process converted them into an audience. Neither fact answered whether the network’s capacity and institutional access could be used for a company announcement.

The network operator answered a policy question

The DCA, which managed the ARPANET, contacted Thuerk’s management. Its response, reproduced in the archive, described the product advertisement as a violation of the network’s official-use policy and called on sponsors, hosts and liaisons to explain that policy to users and contractors. This was not a modern anti-spam filter, a court ruling, or a ban on all unsolicited speech. It was an administrative decision about a government-supported network whose access and costs were not open on equal terms to every company.

Jake Feinler’s archived comments make that unequal access argument explicit. If companies already connected to ARPANET could advertise to a technically sophisticated audience while companies without access could not, the network would become a selective commercial advantage. The question was therefore larger than whether recipients liked the pitch. It concerned who could use a publicly funded communications resource to reach a market that other firms could not reach on the same terms.

The complaints did not all make the same case. One recipient objected to the volume and the message’s use of network resources. A user at the University of Utah believed the message had shut down his system; Templeton’s later interview with Thuerk suggested that mailbox storage was already nearly full, so the timing may have been coincidental. The surviving evidence cannot establish that the mailing caused an outage. The Computer History Museum records the claim as part of the reaction, not as a proven impact. Its timeline entry captures the controversy but should not replace the archived technical correspondence.

Richard Stallman’s May 8 response offers a useful counterpoint. He had not received the DEC message and said a relevant product demonstration might not have bothered him. He did object to the enormous header once he saw how it had been constructed. The exchange distinguishes two issues that later shorthand tends to merge: whether the content was useful to some network participants, and whether a sender could turn a shared network directory into a mass-delivery channel.

A 1975 warning described a different failure

Jon Postel had already written about unwanted mail in RFC 706, “On the Junk Mail Problem,” in 1975. Its mechanism was not a marketer addressing hundreds of people. The Host/IMP interface lacked a way for a host to refuse selected messages; a malfunctioning host could send so many messages that a receiving host would have to process them at the expense of normal users. The RFC framed this as a possible denial of service through machine behavior.

That earlier warning is part of the same history of scarce network resources, but it does not establish a policy against commercial announcements and it does not discuss Thuerk. The 1978 episode added a human choice of audience and purpose to the technical question of how much traffic a network could carry. Conflating the two would mistake a capacity failure for a permission dispute.

The phrase “first spam” is a convenient label, not a complete explanation. The term was not used for this message at the time, and no surviving evidence here proves that no earlier unwanted electronic messages existed. The Computer History Museum calls it the earliest known case; a safer description is one of the earliest documented mass commercial mailings on ARPANET. The label does not tell us how the addresses were gathered, which recipients saw the message, or why the network administrator objected.

The event’s durable lesson lies in those boundaries. Connectivity made remote delivery possible. A directory made potential recipients visible. Mail software imposed practical limits. A network operator applied a use rule, while recipients argued about relevance, cost and fairness. Being addressable was not the same as having invited a broadcast, and successful delivery did not settle who had authority to use the network that carried it.

Sources