Return of the Prodigal Son: The Rehabilitation of Steve Dunwell
Harwood Kolsky
2002
Your query was:
Another responder to the book questionnaire stated "it is my understanding that[Watson's] public apology ... came only after Harwood Kolsky, upset as we all were by Watson's statements to the press, had the nerve to request an audience with Mr. Watson. This was after several months during which the bugs were worked out of the Los Alamos system and the AEC expressed enthusiasm about its performance. During this time Steve, who had been appointed Systems Development Manager for the Poughkeepsie lab, was removed from this post after about three days. I heard that Dr. Kolsky's session with Mr. Watson was extended to a full day and after this an educated Watson issued his public apology to both Dunwell and Stretch. Steve then was awarded a richly deserved fellowship. I can't verify this story but it is what we all understood to be the facts at the time."
Well, Dr. Kolsky, will you verify that something like the above actually took place?
Answer:
Well, there is some truth in the story, but it has become more "heroic" in the retelling. There were two stories. The bare facts are:
In 1961 an investigation was launched at IBM HQ on why Stretch did not meet its goals. I became upset over the negative implication of the term "investigation" and wrote a 14 page memo entitled "Stretch Investigation" to E.R. Piore, Chief Scientist. I spelled out the Stretch project's accomplishments and objected to there being a witch hunt. He assured me that this was not the intention of the investigation (although it probably was). In June 1961 when the "Stretch Evaluation Reports" came out. They were ambiguous, but not as bad as expected.
July 11-12, 1961, there was a Stretch Users' Group meeting at Los Alamos. The users and prospective users were enthusiastic about Stretch and were looking forward to a "Stretch II", which they assumed was inevitable.
LRL Livermore was still negotiating for a Stretch. They argued that since the machine was not 100 times as fast as the 704 they should not to have to pay full price. IBM caved under pressure and cut the price. LRL got their machine, but also doomed the 7030 to never be profitable.
T.J. Watson made his famous speech at the 1961 Western Joint Computer Conference in Los Angeles. Although he didn't actually say the words, everyone understood, "IBM is getting out of the super computer business." The effect was the same. IBM lost momentum and opened the gates for Control Data CDC and other competitors. Steve Dunwell had been banished to a "dungeon" office in the basement of Yorktown.
By 1963 the CDC 6600 was a real threat. On August 23, 1963, I wrote a strongly-worded memo "The AEC Computer Situation" saying IBM's inaction was losing our top customers, and urged the establishment of a new Stretch-like project to recapture them. Again I sent it to my old friend Manny Piore. (This memo also got me on the witness stand of the IBM antitrust trial in 1974.)
It so happened there was a meeting of top executives the very next week at Jenny Lake, Wyoming. Piore showed my memo to Watson, who reportedly hit the roof. Those who were there told me that at one point he held my memo rolled up like a club and shook it while he blasted THEM for losing the super computer market.
My telephone began to ring immediately. Everyone wanted to know what was going on. So did I. One week later, I had a lengthy meeting with T.J. Watson in Santa Monica. Two nervous aides were with him. The discussion was very serious but friendly. He really wanted to know what was going on...
Later I got a letter from him saying that our meeting was "most useful and interesting" and that "you will see sufficient changes taking place"...