The ARPANet is for porn
In 1973, Alexander Sawchuk needed a photo with which to test a new digital image compression algorithm he was developing at the University of Southern California. A glossy page with a variety of image properties to test against. In particular, he wanted a human face. A brief search around the lab turned up a photo of Lenna Sjööblom, Playboy's Miss November 1972.
They say that anything you post to the internet can live on forever. For Lenna it turned out to be true, even in that pre-web age. Sawchuck's digitised photo went on to become one of the stock images used in compression research, having now appeared in countless papers on the subject.
Many more women have since followed in Lenna's footsteps, proving the rule: give computer geeks a public network and before you know it they'll fill it with smut.
The web gets out of sync
Reading email on the web sucks; that was Microsoft's judgment in 2000, back when the web was a page-based medium. Each HTTP request meant a round trip to the server that refreshed the entire page, which was no good for high-activity applications such as email clients. So, to make Outlook Web Access 2000 more usable, Microsoft developers created a way for browsers to communicate with web servers, by loading small amounts of data asynchronously.
Surprisingly, the idea stuck. Despite a troubled history with Internet Explorer, the Mozilla Project built similar functionality into Mozilla 1.0 in 2002, calling it XMLHttpRequest. The floodgates were opened, and a new way of coding for the web was born.
It's hard to believe that Facebook, Gmail, Google Maps and countless other Ajax-enabled sites owe their existence to Microsoft's lead. But it's a good thing they did; if they had waited for the W3C to standardise XMLHttpRequest, they would still be waiting today.