Birth of the killer app

Apple Computer is celebrating the fifth anniversary of its “killer app,” the Apple iPod MP3 player. But Eric Allman, Chief Science Officer at Sendmail, is among those Internet pioneers celebrating the anniversary of an even older and more seminal killer app: Internet e-mail, which celebrates an anniversary of sorts this October: It has been 25 years since work started on SMTP (Simple Mail Transfer Protocol), which gave birth to modern e-mail communications.

InfoWorld Senior Editor Paul F. Roberts spoke with Allman, who recalls that even in the early 1980s, e-mail and its free, effortless communications were the magnets that drew many early users online.

InfoWorld: Tell about your role in the early days of Internet e-mail. What were you doing at the time?

Eric Allman: At the time I was a full-time staff member on a research project at Berkeley and a part-time grad student. The research project I was on was a (Relational Database Management System). Today, all databases are relational, but back then they didn’t exist. We were building one of the few original database management systems. So with this we got a DARPA grant and with that came an ARPANET [the Advanced Research Projects Agency Network] connection for our little research machine, which probably had the power of my Palm Pilot and definitely had less memory.

IW: It was the size of a refrigerator, right?

A large refrigerator, yes. Pretty much everyone in our department wanted an account on the machine so that they could access ARPANET, and our machine was not powerful enough to do that. But a couple things occurred to me. First, people were saying that they wanted an ARPANET account, but what they really wanted was ARPANET Mail. Second, that we had our computers connected on a network at Berkeley, and that I could just write some software that would connect two different networks together. So I did, and that was Delivermail, which was a simple little program. It was just a hack really. That was around 1980.

IW: Those early days of the modern Internet seem to be a simpler time when you could get a bunch of people together in a room, cook up a flexible, open standard, and just implement it. Was it really like that or was it more tortured?

EA: Will it really ruin your day if I say “Yeah, it really was like that?”

IW: No it will just confirm that my suspicions were correct.

EA: The Net back then was a relatively small group of people who were technically savvy and all of whom had as their primary goal making the Internet work. Today, most users of the Internet do not have as their primary goal “making the internet work.” At best, they want it as a tool and at worst they are actively trying to harm it.

IW: Could you have envisioned the problems with spam e-mail when you were laying the groundwork for SMTP?

EA: That’s a hard question. Could I have? Sure. Did I? No. Once again, it was a very collegial atmosphere.

IW: Why would you send e-mail to someone you don’t know, right?

EA: Exactly. It would be stupid. Why would anyone do that? Well, now we know. In retrospect, looking at the economics of it, why would somebody not do it? But it’s not the way we were thinking back then.

IW: Recently there’s been a fight over ICANN (Internet Corporation for Assigned Names and Numbers) and U.S. control of ICANN. What are your thoughts on that?

EA: I don’t think it’s reasonable to say the United States runs the Net. I think our policy will be shortsighted. ICANN is just an administrative organization. If enough of the right people say we’re going to change this, they can. For example, China already has systems out there that use Chinese characters in domain names. ICANN says, “No, you can’t do that,” but they’re already doing it. They’re a country. They run their own DNS servers. They can do whatever they want. People seem to think that ICANN has the power to reach out and cut off a switch, and it isn’t true.

IW: As we look ahead 10 or 20 years, what will become of e-mail?

EA: I don’t think e-mail will go away, but I do think e-mail protocols will have to evolve. One of the projects I’m working on is DKIM (Domain Keys Identified Mail). It could be 10 years or 20, but we’ll have an authenticated mail system up.