Ken: The Moscow adventure (1)
In the early 1980s I spent a week in Moscow with Unix pioneer Ken Thompson. That was quite an adventure.
I have written about this elsewhere. If you know the story just ignore this article. If not, have fun. I remember everything quite vividly.
Ken Thompson is one of the most influential people in my life. Few others have had a similar impact on the way I think and understand things. He even contributed directly to my somewhat devious sense of humour. The adventures I have been through with him and his family in close to forty years and around 30 visits are the subject of a series of stories on my biographical blog.
In the late 1970s Ken, who has given the world the operating system Unix and the computer language C, both still going strong (!), had revolutionized computer chess — by building a hardware-based chess playing program. It was called “Belle” and its development is described in my first article of this series: Ken — an introduction in three syllables.
The Soviet Union, at the time still alive and kicking, had for decades dominated chess. They had all the strongest players (except for Bobby Fischer). They were maintained by the state, which used the latest research technology — teams of grandmasters, libraries of books and a giant “kartotek” I actually saw and have described here. They had also started dabbling in computer chess, and their brute force program Kaissa had won the first World Computer Chess Championship, in 1974. The title was subsequently taken from them, in 1977, by a giant American mainframe program, Chess 4.6, which in turn was defeated in 1980 by Ken’s Belle. I have traced the ascent of Belle here.
The Soviets were troubled by this development, and were looking for alternatives. The great World Champion Mikhail Botvinnik, an electrical engineer, was trying to find a way of managing the Soviet economy using artificial intelligence. With a team of computer scientists Botvinnik designed a highly selective chess program that used general principles of chess to drastically restrict the search. The project is described here.
In any case, in 1982 the Soviet Chess Federation and Botvinnik invited Ken to Moscow to demonstrate his machine. As a friend of both Ken and Botvinnik, whom I had interviewed for a German TV documentary a few years previously, I too was invited.
After this somewhat long-winded introduction we can proceed to today’s story. Before flying to Moscow I coordinated with Ken. We would be staying in the same hotel — but just in case, if they moved him somewhere else, I told him to leave the new address in the original hotel, so I could find him. And: if things went completely astray, he should the next morning at 10 a.m. go to the central Red Square and stand in front of the cathedral that looks like a wedding cake. I would meet him there.
Well, everything did go astray, and on my arrival I could not find Ken in our hotel. The hotel staff did not know (or care about) a Ken Thompson, and so we did not make contact that night.
The next morning I told my Intourist guide that I would be going to the Red Square to meet a friend, to which he just said: “No, you won’t.” I was a little shocked: “I am a guest of the Soviet Chess Federation. I believe I can go anywhere I want. Nobody said there were any restrictions.”
But my guide insisted: “You are not going there! Today is Gitler Kaput (“Hitler kaputt” = the yearly celebration of the end of WWII) — there will be a million people on the Red Square, holding up pictures of fallen soldiers.” Of course, the plan had to be abandoned.
This is a picture Ken took a few days later, when the celebrations were over. The arrow indicates where we were scheduled to meet at 10 a.m. The next picture shows you what Red Square and the streets leading to it looked like on Gitler Kaput. You couldn’t even reach the Square, let alone enter it.
What to do? I decided to call Botvinnik, but only got a lady who spoke no English. So I asked my guide to call for me. “You want to make phone call with great champion Mikhail Botvinnik?” he asked in disbelief. When I convinced him it was okay, he said he would do it, but needed to know Botvinnik’s “otchestvo”. His what? “What is his father name?” he repeated in clarification. “I don’t know that,” I said impatiently, “just call him, please.”
But my Intourist guide was adamant. He actually went down the street to a bookstore and came back triumphantly with the information he needed to be able to call Botvinnik: Moiseevich, son of Moise (Moses). Without that, he simply couldn’t call, it would have been too rude to address him as Mr Botvinnik.
This is how the system works: If you were from the West, you could call him Mr Botvinnik. But if you were Russian you absolutely had to know that his father was Moise, and call him Mikhail Moiseevich. Patronyms, which simply mean “son (daughter) of” and are formed by adding a suffix: -ovich/evich for a boy, -ovna/evna for a girl. In Russia I would be Frederic Aloisovich (here’s why).
Well, in the end I reached Botvinnik and through him Ken. They were both in crisis mode: the computer “Belle”, which was going to be part of the lectures and discussions, was not on the plane from the US. Mikhail himself had climbed into the hold to make absolute sure. Ken had checked the computer in the freight department of Kennedy airport, but it had not made it to Russia. He tells you why in this interview recorded in May 2019:
We placed a call to Bell Labs and Joe Condon, the co-builder of Belle. And this is how that worked: the hotel told us the call would take up to three hours to go through (it took four), and we should wait in Ken’s room. And then they called every fifteen minutes to make sure we were waiting. On the other hand, when Joe woke up, he went into his kitchen in pjs, pushed a dozen numbers and immediately spoke to us. It turned out that Belle had been confiscated by the US authorities at the airport, because the export of advanced technology to Russia was strictly forbidden. Ken could face arrest on his return.
So the current problem was that we had no chess playing computer in Moscow. Ken and I were summoned to the head office of the Soviet Chess Federation, where the president, Nikolai Krogius, received us with a very stern face. Nikolai Vladimirovich was a famous grandmaster (he had assisted Spassky in his 1972 match against Bobby Fischer), and he told us that the missing computer was a disaster: “Can you imagine what the Soviet press will make of this?” Ken shrugged. “Ruhollah Khomein has banned chess in Iran,” Krogius said. “Is Ronald Regan trying to do the same in the America?”
Krogius also wanted to know if Belle had any military use. On his return, the Washington Post asked the same question. Ken’s reply: “Well, you might be able to kill a few people if you threw it out of a plane.”
So the lectures were cancelled. Instead Botvinnik invited us to his dacha outside Moscow, where we could chat with his AI chess programmers. Actually, we wanted to meet them in their workplace and fool around with the computer. But Mikhail insisted we all come to his summer residence. There we were served a sumptuous buffet meal, with smoked meats and fish, noodle salads, breads and general delicacies. We ate and ate until we no longer could follow their instructions: “Go on, take a little more.” After we said we simply couldn’t, some helpers cleared the table — and brought in the main meal! Big roasts, steaming potatoes, vegetables and salads. We were doomed!
The lunch included Botvinnik’s programming team, and in the afternoon we at last got to discuss their efforts to implement a highly selective search, using criteria defined by the great world champion, to drastically reduce the number of positions a chess program had to generate and evaluate before it made a very strong move. They could not show us the computer or their program, but spent a fair bit of time explaining how it could find the famous bishop sacrifice that their mentor had played against the former World Champion José Raúl Capablanca in the Avro Tournament in 1938. Chess players note: you can replay the game — it is the second example on this page, and the bishop sacrifice occurs on White’s 30th move. The replayer has a fan icon below the board which you can use to start a traditional brute force chess program. You will see that this JavaScript engine, even on an Android phone, plays the bishop sacrifice in less than a second.
Botvinnik’s Intelligent Programming team explained very eloquently how their algorithms found the bishop sac for purely positional reasons in a couple of minutes. It sounded very convincing — it was clear that they had provided this explanation to experts many times before.
At this stage Ken asked them about a position he had given Belle. It is from a 1913 game in Paris, when the future World Champion Alexander Alekhine sacrificed his queen, and announced a mate in ten. He played 22.Qh5+!! and won after 22…Nxh5 23.fxe6+ etc., leading to mate. “Belle can almost find this move in reasonable time,” Ken said.
The intelligent-method team immediately started to apply their algorithms on the position, using paper and pencil to do so. Botvinnik looked on with pride. But his approving smile froze when he noticed that I had removed the white pawn on h2. As one of the strongest chess players of all time, he realized that without that pawn it was not a mate, and 22.Qh5+ was a terrible mistake. In fact it lost the game. Still, it did not disturb his computer team: after considerable calculation and enthusiastic discussion they came to the conclusion: “Our program will find 22.Qh5+!!, definitely, in under two minutes.” It was the “dynamics” in the position that allowed it to do so. Our conclusion (not shared with them): if these guys had a computer and a program at all, they were mainly trying to tune it to find the famous Botvinnik-Capablanca bishop sacrifice.
In the above May 2019 interview with Brian Kernighan you can listen to Ken describe his involvement in chess, building chess machines, and the trip to Moscow. It gives you an impression of what Ken is like, even today — the things he has done and the way he narrates episodes from his life. It will also tell you why I went to stay in his home, first in New Jersey and then California, around twenty times during the last forty years.
Before I break off and announce a somewhat harrowing part two on the Moscow visit, let me share one little incident that is typical when you are hanging out with Ken. The Falkland War was under way and our hosts started talking about it. “The British are doomed,” they said. “The Argentineans have sunk a British warship with a French-built Exocet missile. They can knock out their entire navy.” Ken’s comment: “It won’t happen again.” The Russians were eager to know why he thought that, but he was silent. Later I asked him, and he told me, guardedly, that the British, who even knew the serial numbers of the French missiles, had adequate defences. The Exocets had friend-foe signals to allow or prevent ships from taking them out. Apparently there was a mess-up and the British destroyer had not recognized the missile as a one fired by the enemy. That was immediately fixed. And true to Ken’s word: it didn’t happen again.
