Seven Questions for Nathaniel Borenstein, Who Made Email Attachments Easy
The Internet isn’t known for looking backward at its history all that often, and yet once in a while it’s worth a look back to appreciate why things we do every day work the way they do. March 11 is one of those opportunities. It is the 20th anniversary of MIME, which stands for Multipurpose Internet Mail Extensions.
You never think about it, and yet every time you attach a photo or a Word document, or practically anything else to an email message, you’re using it.
It was created by Nathaniel Borenstein, a computer researcher who, 20 years ago, worked for Bellcore, the research arm of the Baby Bell telephone companies. At the time, no one really gave much thought to the idea that email could or even should comprise any more than basic text messages, and when attachments were involved, incompatible formats caused the kind of headaches that we would consider unacceptable today. Curiously obsessed with the evolution of email, Borenstein teamed up with Ned Freed, a fellow Internet pioneer, to write the MIME standard that is the backbone of email attachments today, supporting more than 1,300 types of files and enabling billions of email users to ignore any worries about compatibility among email programs.
The first message containing a MIME-encoded attachment was sent on March 11, 1992, and today the standard is used something like a trillion times a day. And no, he didn’t get rich (but he did once turn down a job offer from Steve Jobs). He’s now the chief scientist of Mimecast, a cloud-based email outsourcing company that just happens to riff on the name of the standard he helped create. I got a chance to talk to him by phone a few weeks ago. Here are some highlights from our conversation:
AllThingsD: Nathaniel, no one really thinks about MIME, but everyone uses it. Tell me how it happened. The story goes, you thought that one day you’d like to use email to receive photos of your grandkids. Is that true?
Borenstein: I can’t say it was my primary motivation, but it was an easy way to explain what I was thinking of. Email had been around since 1965 on time-sharing systems, and then moved to the fledgling Arpanet. Then more and more people outside the English-speaking world started to come up with a lot of incompatible ways to encode their email. At the same time, people wanted to send around files occasionally. The only way that was really safe to do it was to package up a file with a program called UUencode, which had multiple versions that weren’t always compatible. There were all these ad hoc things that people were doing for these complementary needs. In 1980, I was a grad student at Carnegie Mellon, and I was put in charge of maintaining an email program. It was just a job at first. Then we got some Unix machines. I thought I could do a better job by rewriting the email program. And I was also in charge of running an email system. It became sort of a hobby. Then, later, after I finished my dissertation, my adviser asked me to write what he described as the world’s best email programs. Suddenly my career was my hobby.
So what was it you were asked to work on?
It was Carnegie Mellon’s Andrew Project, which was intended to envision the next generation of computing environments for campuses. What was really interesting then was that it had very advanced — for that time — multimedia capabilities. And so we had a chance to make multimedia work. There were a few programs with multimedia that came before, but we had the chance to get it into people’s hands. And then something interesting happened. Steve Jobs came to visit. This was in the days that he was running NeXT. [Jobs founded NeXT after leaving Apple in 1985, and ran it until he sold it to Apple in 1996. -Ed.] He came to the campus, and a light went off in his mind when he saw the mail system. He had not completely gotten email until he saw what we could do with it, and so he tried to hire our entire team. And he got none of us. None of us wanted to go to work for NeXT.
Why was that? Was it about Steve, or about NeXT?
I’m not sure we all had the same reasons. In my case, it was that I had heard from other people that working for him was difficult. I have enormous respect for him, because he was one of the captains of our industry. But I had a feeling that if you went to work for him and you had a disagreement with him, you lost. It was that simple.
And let me guess: He built email into the NeXT operating system anyway?
His team built NeXTMail, which looked a lot like Andrew did. In fact, if you use Apple’s Mail.app on the Mac, you’re using something that looked a lot like Andrew did. But he did something interesting. He created a way for people to send files around. And so you had two communities of users on NeXT and on Andrew who couldn’t send files to each other. So after I left Carnegie Mellon, I went to work for Bellcore, which was the research arm of the Baby Bell phone companies. My job was as a researcher, and my mandate was to work on things that would encourage the use of bandwidth. I thought I was done working on mail. Then I started noticing these problems with compatibility, plus I had an idea for something I called active messages. And Bellcore was a very heterogeneous computing environment. There were all these Unix hackers, and each person had their favorite email program — I counted more than 20 in use. And I wanted them all to be able to read these active messages. So what I did was start patching them all. That led to something I called Metamail, which would be triggered by a header in the email that would then call on any one of a number of other programs. So if you received a JPG image in Atomic Mail, which was one of the programs in use at the time, it would display the JPG in Atomic Mail.
So this led to the MIME standard how?
Metamail was already in use when the work that led to MIME started up. I got involved with some efforts at IETF [Internet Engineering Task Force]. I got introduced to Einar Stefferud, and he became a mentor of mine, and introduced me to Ned Freed, who became my co-author on MIME. I was worried about email and multimedia compatibility, and Ned was working on email gateways — the problems of translating messages between different email realms — because we weren’t all running SMTP back then. Stef thought we should work together. Now, Bellcore had allowed me to take Metamail and contribute it to the public domain, or what we would now call open source, and so anyone was allowed to modify it. So every time there was a new draft of the MIME standard, I could update to support the new standard. So when the first public draft of the MIME standard was ready, I was ready with Metamail, and it was just picked up at an incredible rate. I wrote it for Unix, and three days after the release, someone had already adapted it for Microsoft DOS. That’s what told me I had a hit on my hands.
What about MIME made it flexible?
One of the reasons I think we were successful with it was the fact that we had an incomplete vision. Yes, I was thinking about pictures of grandchildren someday — I am a grandfather now, by the way — but I knew that there would be things coming that I couldn’t forsee, and I didn’t want the system we designed to have to be completely redesigned in order to accommodate the new things. That is why the MIME type system is so open. You just go to the Internet Assigned Numbers Authority and register a new MIME type. The original supported 16 MIME types, and when I checked a few years ago, we were up to 1,309 different supported file types. When I tried to explain why I wanted it to be open-ended, I tried to explain things that you could, at the time, just barely imagine. I had two examples — one was smell. I thought that one day you might be able to send files containing smells, and one day you might, but no one is really focused on it yet. The other was one I thought of almost as a joke. I proposed a MIME type for matter transport. I thought we could send matter around as an email attachment, and in a way, it came true. When you think about 3-D printing and the models for that, people are sending around schematics for 3-D objects that can then be printed. So open-ended is very good in a world where science fiction is quickly getting overtaken by reality.
You now work for a company called Mimecast. What do you do there?
The short tag line is that we do unified email services in the cloud. We take all the things that surround and administer your email, and everything except the basic core operation of it, we outsource to the core. We archive your email, we set policies about how and when it can be deleted. We do data-loss prevention. We do continuity and disaster recovery. Our BlackBerry users didn’t notice when the service went down last year. And I’m not the founder. A lot of people think I am, but I’m not. I just work for the company as its chief scientist. Once you get all those things in one place, there’s a lot of potential. You can do things that you couldn’t do before. I’ll give you just one example: Imagine you’re composing an email, and as you type, there’s a sidebar next to it. In the style of Google Instant, it becomes an implicit search query that searches both your email archive, but also, say, news stories. The point is that it might help you shape your message or change what you want to say in your email, who you want to say it to, or whether or not you want to say it at all. Having an email archive solves a very deep problem, which is organizational memory. Everyone wonders from time to time whether someone knows the answer to some question. The point is that the bigger an organization is, the more often it’s necessary to rediscover the same thing over and over. Having an email program that searches for things that might help you would go a long way toward solving this.
Here’s the Infographic on the history of MIME. Click to see it bigger: