On 29 October 1969, two scientists established a connection between computers some 350 miles away and started typing a message. Halfway through, it crashed. They sat down with the BBC 55 years later.
At the height of the Cold War, Charley Kline and Bill Duvall were two bright-eyed engineers on the front lines of one of technology’s most ambitious experiments. Kline, a 21-year-old graduate student at the University of California, Los Angeles (UCLA), and Duvall, a 29-year-old systems programmer at Stanford Research Institute (SRI), were working on a system called Arpanet, short for the Advanced Research Projects Agency Network. Funded by the US Department of Defense, the project aimed to create a network that could directly share data without relying on telephone lines. Instead, this system used a method of data delivery called “packet switching” which would later form the basis for the modern internet.
It was the first test of a technology that would change almost every facet of human life. But before it could work, you had to log in.
Kline sat at his keyboard between the lime-green walls of UCLA’s Boelter Hall Room 3420, prepared to connect with Duvall, who was working a computer halfway across the state of California. But Kline didn’t even make it all the way through the word “L-O-G-I-N” before Duvall told him over the phone that his system crashed. Thanks to that error, the first “message” that Kline sent Duvall on that autumn day in 1969 was simply the letters “L-O”.
They got their connection up and running about an hour later after some tweaks, and that initial crash was just a blip in an otherwise monumental achievement. But neither man realised the significance of the moment. “I certainly didn’t at that time,” Kline says. “We were just trying to get it to work.”
The BBC spoke to Kline and Duvall for the 55th anniversary of the occasion. Half a century later, the internet has shrunk the whole world down to a small black box that fits in your pocket, one that dominates our attention and touches the furthest reaches of lived experience. But it all started with two men, experiencing just how frustrating it is when you can’t get online for the very first time.
This interview has been edited for clarity and length.
Can you describe the computers that enabled Arpanet? Were these massive, noisy machines?
Kline: They were small computers – by standards of that time – about the size of a refrigerator. They were somewhat noisy from the cooling fans, but quiet compared with the sounds from all the fans in our Sigma 7 computer. There were lights on the front that would blink, switches that could control the IMP [Interface Message Processor], and a paper tape reader that could be used to load the software.
Duvall: They were in a rack big enough to hold a complete set of sound equipment for a large show today. And they were thousands if not millions or billions of times less powerful than the processor in an Apple Watch. These were the old days!
Take us inside that moment when you started typing L-O.
Kline: Unlike websites and other systems today, when you connect a terminal to the SRI system nothing happens until you type something. If you wanted to run a programme, you first needed to login – by typing the word “login” – and the system would ask for your username and password.
As I typed a character on my terminal – a Teletype model 33 – it would get sent from my terminal to the program I wrote for the SDS Sigma 7 computer. That program would take the character, format it into a message and send it to the Interface Message Processor. When it was received by SRI’s system, [it] would treat [the message] as if it came from a local terminal and would process it. It would “echo” the character [replicate it on the terminal]. In this case, Bill’s code would take that character and format it into a message and send it to the IMP to go back to UCLA. When I receive it, I will print it on my terminal.
I was on the phone with Bill when we tried this. I told him I typed the letter L. He told me he had received the letter L and echoed it back. I told him that it was printed. Then I typed the letter O. Again, it worked fine. I typed the letter G. Bill told me his system had crashed, and he would call me back.
Duvall: The UCLA system did not anticipate that it would receive G-I-N after Charlie had typed L-O, so it sent an error message to the SRI computer. I don’t recall exactly what the message was, but what happened next was due to the fact that the network connection was much faster than anything seen before.
The normal connection speed was 10 characters per second whereas the Arpanet could transmit characters at up to 5,000 characters per second. The result of this message being sent from UCLA to the SRI computer flooded the input buffer which only expected 10 characters per second. It was like filling a glass with a fire hose. I quickly discovered what had happened, changed the buffer size and rebuilt the system, which took about an hour.
Kline: No, I certainly didn’t at that time.
Duvall: Not really. It was another step forward in the larger context of the work we were doing at SRI which we did believe would have a large impact.
When Samuel Morse sent the first telegraph message in 1844, he had an eye for drama, tapping out “What hath God wrought” on a line from Washington, DC to Baltimore, Maryland, US. If you could go back, would you have typed something more memorable?
Kline: Of course, if I had realised its importance. But we were just trying to get it to work.
Duvall: No. This was the first test of a very complicated system with a lot of moving parts. Having something like this complex work in the very first test was dramatic in and of itself.
BBC
