Title: How a Student's Daydream About Stellar Spectra Became one of the World's Most Important Digital Libraries, the Smithsonian/NASA Astrophysics Data System (Video)
Abstract: In 1979 I was daydreaming out the window of the Kresge Library when I had a vision of stellar spectra as occupying points in an n-dimensional vector space, where n was the number of resolution elements in the spectra. This led to my thesis work, where I used, for the first time, n-dimensional vector distances and Eigenvector techniques to classify astronomical spectra.
My thesis became well known enough that I was still talking about it five years later, at a meeting in Bavaria, where someone in the audience came to me and said that the exact same mathematical techniques could be used to classify documents. I view the ADS as being born in that moment.