It's 1998. One late night a young engineer at Microsoft is working on the next build of Internet Explorer, version 5. He has this great idea in his head that he wants to implement. Why don't bookmarks (favorites) have an image next to them? Representing bookmarks as images would make it easier for the user to organize and find their bookmark. So he dives into the old Win32 COM Framework and sets up a scheme in which a server adds a "favicon.ico" file to their configuration's root directory, and any time a webpage is bookmarked, the browser will send a get request for the asset in question and display it next to the bookmark's name. Problem was, he wasn't supposed to be working on that. It wasn't a part of the project outlines. So he calls over a naive assistant project manager and asks if he can commit his build to the trunk. The assistant thinks it's a cool idea and OK's the commit, promptly getting yelled at the next day for insubordination.
With a team of over a 100 engineers, it's a little too late to revert the commit, and the feature stays in. Almost 20 years later the favicon is now the de facto "thumbnail" of a web page. Browsers now display favicons on each tab of a website. Shortcuts use it as an icon. It goes without saying that bookmarks makes use of the icon. Billions upon billions of failed HTTP requests looking for a "favicon.ico" on amateur websites that never added one can be blamed at the foot of one young engineer who had a neat idea went through with it, despite the bureaucracy.
It's not exactly the invention of the C programming language, but it's the small innovations like these that count. Currently Microsoft is trying to play catch-up to Apple with Windows 8. Perhaps if they had allowed innovation instead of stifling it, Windows as a platform and an OS would be extremely different today.
