Votomatic card

Votomatic was a brand name of a punched card-based voting system produced by IBM starting in 1965 based on an invention by Joseph P. Harris. Earlier ideas of using punched-card ballots went back to the 1890s, but this was the first system to achieve success. By the 1990s, some version of punched-card voting was in use by over one-third of U.S. voters. However, the problems with the 2000 U.S. presidential election, in which some of the notorious Florida recounts were of Votomatic-style ballots, led to punched cards going out of favor for ballots in favor of other systems including touchscreens and optically-read ballots.

A Votomatic card had a punch position for each candidate or option in each race or referendum up for vote. These were pre-perforated so that a press of a stylus at the appropriate position would cause the central chip of paper (known as a "chad") to fall out, leaving a hole which could be read by a tabulating machine. In general the candidate names weren't on the punched card itself, but on an external ballot that the card was placed beside in order to punch the holes that lined up with the desired names, so it is necessary to know the layout of the ballot for a particular election to interpret the vote cards resulting from it. (In the case of the infamous Palm Beach County "butterfly ballot", even many of the voters didn't interpret them correctly.)

The 2000 election and its lengthy recounts put the word "chad" in the public consciousness. This word goes back in data-processing to at least the 1930s, and is of uncertain origin, possibly related to "chaff". An urban legend that the term is a back-formation from a device named "Chadless" after its inventor's surname is false. The election recount in 2000 focused on indeterminate ballots in which the chads were not fully punched out in the proper way; "hanging chads" were still partially attached to the card, while "pregnant chads" were fully attached but dimpled from a partial attempt to punch them. These ballots were counted inconsistently on different runs through tabulating machines, and were the subject of intense debate about how they ought to be counted in a manual recount.