This article appeared in Communications of the ACM, July 2001/Vol.44, No.7 and is reproduced here with their kind permission. Please observe the copyright notice below.
Some people say Internet years are like dog years, with one year on the Net being equivalent to seven years elsewhere. It is perhaps natural, then, that in 2001 we should start to mourn the demise of some of the earliest stars of the Web.
For it is now ten years since the first Web browsers became publicly available, although you would have been unlikely to know about them unless you read news-groups like alt.hypertext, and the earliest ones would hardly be recognized as Web browsers by today's users. A feature added to the original line-mode browser in May 1991, for example, was "An option . . . to return to the top of the present document being viewed." Browsers of that time generally did not have the ability to bookmark your favorite Web sites because there were so few servers in the world you could remember all of their addresses. The more advanced browsers could display headings in a different font size, but we still had some time to go before images, or even colors, would be widely supported.
It's also nearly ten years since some friends and I threw together a few odd bits of hardware and software for a joke, and unwittingly started something that has since been covered extensively in the press and other media around the world. In recent months it even made the front pages of the London Times and the Washington Post, yet it is arguably one of the dullest (or at least silliest) sites on the Web. The full story is at www.cl.cam.ac.uk/coffee/coffee.html, but the plot summary goes roughly as follows:
It's late 1991, and 15 or so researchers in the Systems Group at the University of Cambridge Computer Lab share a coffee machine located in a rather uninspiring area known as the Trojan Room. Not all of the researchers are in the Trojan Room, though; others are two or three flights of stairs away and must travel some distance in search of coffee, often to find those closer at hand have beaten them to it. One pot provides enough coffee to fill just a few mugs, so it's first-come, first-served, and distance is a definite disadvantage. In the interests of fair play, some of the residents of the Trojan Room salvage a video camera, an old 680x0 VME-based computer, and a framegrabber left over from other projects. They grip the camera in a retort stand and point it at the coffeepot. The machine with the framegrabber executes a specially written server program, and an X-Windows client, which can be run by anybody in the group, grabs images at regular intervals and displays a picture of the pot, icon-sized, in the corner of the workstation screen. Those too far away to smell the coffee now have an alternative means of knowing when a new pot is brewed. The Net, once again, helps break down the barriers of distance (even if that distance was only measured in yards), and so streamlines the distribution of a resource so vital to computer science research.
This system used our own software and ran our own RPC protocols over our own ATM network stack on networking hardware that we had also, to a large degree, built ourselves. Few of us had heard of HTTP then, and it would have been of limited use because Web browsers could only display text. But when the Mosaic browser was introduced in 1993, with its ability to display images, Web pages could include diagrams, equations, and pictures of loved ones. We soon realized something that now seems obvious: when your browser requests an image from a server, the server doesn't have to return the same image every time. For us, the most convenient source of constantly changing images was the coffeepot camera, so Daniel Gordon modified the server to respond to HTTP requests, and the first Web cam was born.
This meant users didn't have to run a special application (and, more importantly, a special network stack) to view the images, and as a side effect, it also made the image available to everybody else in the world. Before long, the Trojan Room coffeepot was one of the most popular sites on the early Web. Millions have looked at the small grainy image, and many millions more have seen news reports about it on TV. The story kept running, initially describing a novelty, then, shortly afterward, a historic artifact. Only in Internet time can such a transition take place so quickly.
A number of journalists have contacted us over the years to ask, "What makes it so popular? Why has it had so much attention?" To which the best answer is probably "because journalists keep contacting us to ask that sort of question." It became famous for being famous. But maybe there's more to it than that. The page, quite deliberately, looks much the same as it did when it first went online, to show the limited page formatting available then. This also means it's not the most visually gripping site around. In these days of full color, high-quality streaming video with accompanying stereo sound-tracks, perhaps the coffeepot camera has something of the appeal of early silent movies. There is, after all, no accompanying audio (now why didn't we think of that?), and the image is black and white with a very low frame rate.
The peak of its publicity, however, came recently when news spread that the Trojan Room coffeepot camera was to be shut down. Coffee machines, at least those that get the sort of treatment this one gets, do not last very long, and the current one is, I think, the grandson or great-grandson of the original (which probably proves something about "coffee machine years"). Most of the outdated computer equipment that ran the original system still drives it today; but it is the only segment of the network now running at the University, and it gets ever more difficult to maintain.Today we could go out and buy a $30 USB camera that would plug into the back of a PC and do the same job, but it wouldn't be quite the same.
Even these hurdles might have been overcome, however, but for a more serious problem looming on the horizon. One thing that has been constant about the Trojan Room Coffeepot was that it lived in or near the Trojan Room, but this year the Computer Lab moves to a new building in West Cambridge, and the Trojan Room will be no more. It seemed like the right time for the Web cam to retire, though we had no idea its expected departure would be mourned in newspapers from Canada to South Africa.
The whole experience made me realize that the coffeepot perhaps has one last lesson to teach us, one which could, even now, start another new trend. Putting content on the Web is no longer news, it's expected. No organization can get any column inches by starting a Web server. You want to know the secret of getting attention these days? Switch it off.
COMMUNlCATlONS OF THE ACM July 2001/Vol.44, No.7 pp.25-26
© Copyright 2001 ACM.