I can't praise highly enough your article about Hans Berliner in the Spring 2012 issue. May I offer a few of my own reminiscences?
The arrival of a world chess champion in Pittsburgh had the same galvanizing effect on Pittsburgh's tournament chess players that the arrival of grandmaster Alex Shabalov had twenty years later. In 1973, when I arrived at CMU, Hans' name was on my friends' lips, and I soon obtained a copy of an article he had written about chess openings--the genesis of the book he wrote decades later called "The System." I read it, and was influenced by it in my own choices of openings.
I had written a chess-playing program as an undergraduate, and that had probably been a factor in getting admitted to the computer science doctoral program, but I didn't continue down that path for my research, and so I never worked with Hans. We met almost weekly, however, in the lounge, to review the New York Times chess column.
He played in one tournament in which I played, in 1974; and in 1975, we played in one match on the CMU team in the Pittsburgh Chess League. That was his last serious chess, however. I now understand why someone might retire from chess; back then, my mind could comprehend it, but not my heart.
Hans wrote a 12-page monograph about the opening variation of one of his famous correspondence games, which he entitled, with characteristic iconoclasm, "From the Deathbed of 4. N-N5 in the Two Knights Defense." I took it to tournaments and sold it. At the 1978 Master Challenge in Chicago, it flew out of my hands, but after that I didn't have as much luck. The variation continued to be topical, and several years later he published a revised version.
It's easy to recall conversations with Hans. He was an iconoclast in person as well as in print. His short stature and his smile only made this trait more appealing.
His current assessment of "the whole AI thesis" from those days is a little harsh. All the research we did back then, not only AI, was deeply colored by the puny computing power that we had available to us. In the 1970s, research into speech recognition and synthesis funded much of the CS department, but reviewing the research from those days after using Siri feels like looking through a telescope in the wrong direction. And so it has gone with chess. The case of Hitech was particularly poignant because its turn as king of the hill was so short. But faster hardware doesn't make domain knowledge irrelevant; today's software still needs domain knowledge.
--Bruce Leverett (CS'81)
Mt. Lebanon, Pa.
Photo ID, please
Can you identify the two other people in the picture on page 18 of the Berliner piece? The guy typing at the keyboard looks like none other than the legendary Ken Thompson.
SCS Technical Staff
We asked Hans Berliner. "The guy on the left is Ken Thompson," he says, "but I am clueless about who is on the right, and I can't even figure out where and when the picture was taken."--Ed.
- Questions? Comments? Criticisms? Suggestions? Email us at TheLink@cs.cmu.edu or write to The Link Magazine, Office of the Dean, School of Computer Science, Carnegie Mellon University, 5000 Forbes Ave., Pittsburgh, PA 15213 USA. We welcome both bouquets and brickbats, though we do wish someone would explain just what the heck a "brickbat" is, anyway.