JF Ptak Science Books Post 1917
Leonardo wrote backwards and from right to left, Benjamin Button lived backwards at the hands of Scott Fitzgerald, Rene Magritte's man in the bowler saw the back of his head, Herrimann's Ignatz the Mouse I am sure saw the back of his head looking around the world with the world's most powerful telescope, rugby passes are all done backwards, paper images of vue optiques appear backwards, lightning for all intents and purposes starts backwards from the ground up, reverse mathematics are worked from theorems to axioms, and the Chicago River (1900) was engineered to flow backwards for the foreseeable future, while the Mississippi River famously flowed backwards for just a bit in the New Madrid Earthquake of 1812.
I can only imagine what audiences must have felt when they saw the first moving pictures played backwards--seeing them played forwards was a novel-enough (and revolutionary) idea, but the simple idea of reversing the direction of the film would have proved to be equally fascinating.
Imagine the first time you witnessed a staged train wreck on film, back there in 1897, and imagine being able to see it played over and over again, until you were filled. I'm not so sure that there were even any still photographs of a train wreck as it occurred to this point, even with advances in film speed and lens, so seeing the even unfold in front of you at leisure must have been overwhelming. Now imagine these same folks seeing the event and watching the locomotives reconstitute themselves. It would have been an extraordinary event. Even observing the Etienne Marey sequences and seeing what actually happens when a person bends over to pick up a pail of water would have revealed almost as much in new detail as when Galileo was in the middle of his earliest observations.
Looking at things backwards is a good idea so far as thinking about engineering problems and of course in checking experimental results in the sciences--its not so good an idea though to change the results produced by the scientific method because they're not a good intuitive fit to expected parameters.
Such was the cased with the first (and successful) employment of a computer to predict the outcome of a presidential election. THe computer was the UNIVAC (the world's first commercial computer and a blazingly fast machine at 10k operations a second, nearly six orders of magnitude lower than "superfast" by contemporary standards), which was brought in by Remington Rand to CBS News to crunch the numbers on the tight race between General Dwight Eisenhower and Gov. Adlai Stevenson (II) on 4 November 1952. (Stevenson was the son of a former U.S. Vice President and would run again against Eisenhower in 1956.) Pioneers Pres Eckert and John Mauchley, along with Max Woodbury (and programmer Harold Sweeney, who is seated at the UNIVAC's control panel and who seems never to be mentioned in the iconic photo at top, with Eckert at center and anchorman Walter Cronkite at left). CBS News Chief Sig Mickelson and Cronkite were not comfortable with the proposal, but ran with it anyway, sensing a moment of the-future-is-now.
The Eisenhower/Stevenson race was seen by the large majority of pundits to be too close to call, so when the UNIVAC's results pointed to a landslide for Eisenhower (438 electoral votes and 43 states to Stevenson with 93 electoral votes and 5 states) folks got very sweaty and nervous, not trusting the outcome. As this was still a very early age in human-machine interaction, and the computed results fell far away from perception and expected response, changes were made in the UNIVAC's programming to determine a more "reasonable" response by the machine, the new results making the race very tight and fitting human expectations and giving Eisenhower a very slim margin of victory. As poll results started to sweep in an hour or so later indicating that Eisenhower was showing with a huge victory, the UNIVAC was again reprogrammed and at about midnight the announcement was made that the UNIVAC had indeed been correct in the first place. The final results were 442 electoral votes for Eisenhower and 89 for Stevenson. In the next presidential election in 1956 the three networks all had computers working for them--with them--and a different perception had been formed on working with computers.