One wonderfully, special part of my writing career has been the awesome people it has let me meet. I will never forget Bell Labs flying me to Pennsylvania to have lunch with J. Presper Eckert, one of the early computer inventors. He was key in working on UNIVAC.
That was a dinosaur of a computer. It took up a large room which had to be kept cool so UNIVAC didn't overheat. Plus, the computer needed a big team of workers to tend its many parts and keep it operating.
One of the stories J. Presper Eckert shared with me was being in charge of UNIVAC predicting the outcome of the 1952 Presidential Election. That was the election that pitted General Dwight D. Eisenhower against Illinois Governor Adlai Stevenson. The press and polls all favored Stevenson. UNIVAC would be shown on CBS during the lengthy reporting of incoming state results. It was hoped, though seriously doubted, that the computer could predict the election outcome ahead of the West Coast polls closing.
There was a lot of pressure to have UNIVAC accomplish this never-bef0re-tried task. It was viewed as possibly turning the public's view to believing computers could be worth companies and government agencies investing in them. Remember, back in 1952 no one had an iPad, a laptop, or even a cell phone.
In fact, I have to admit that when I interviewed J. Presper Eckert I was just beginning to work on a home computer, which took up half my desk and required me to use special codes to make it indent and capitalize. Text was green and I couldn't imagine color or pictures being on my screen. This was even before I wrote my first book about computers and kids creating their own computer programs.
Back to my story.....
The next part never seems to be reported in historical accounts. But it's key. Eckert explained that for UNIVAC to be able to make a prediction for the Presidential Election it needed data for comparison. So Eckert had his team collect what he called "Bellwether" data on past results from across the US. Something is called Bellwether if it shows what tends to lead a trend. In other words, UNIVAC was given a LOT of past election results it would then be able to compare against incoming data on Election Night. It's goal--LOOK FOR A TREND.
OK, the big night came. The famous news reporter Walter Cronkite was there. So was J. Presper Eckert. The data on polls reporting votes from each state came in and was entered into UNIVAC. The computer churned away--actually it was a matter of LOTS of parts called transistors switching on and off. Then the moment came.
IMAGINE A DRUM ROLL!!!
Before the polls were even closed on the West Coast, UNIVAC reported Eisenhower would have 483 electoral votes and Stevenson only 93.
Someone from the CBS news team took Eckert aside and said basically, "NO WAY! Impossible. Try again."
Eckert told me he was really sweating and nervous. He didn't want to let his company down by having UNIVAC fail and show computers really couldn't be counted on. He had his team try again. AND this time the results were....THE SAME! But CBS would not report UNIVAC's results and predict an Eisenhower win.
However, when the night was over and all the results from across the US were in--all the votes were counted--the results proved UNIVAC was right. Well, overall. The actual final number of electoral college votes was Eisenhower 442 and Stevenson 89.
Dwight Eisenhower was President.
Because of UNIVAC's success, computers became trusted data processors worldwide. Computers have, in fact, been counted on to predict the outcome of every presidential race since.
NOW, when you're watching reports of this year's computer prediction of the Presidential Election results, pause for just a moment. Remember UNIVAC, the computer that did it first, and wasn't trusted to be right.
No comments:
Post a Comment