In 2003, California Governor Gray Davis faced an October 7 recall election which pollsters predicted he would probably lose. The Guv – no doubt concerned only for the public weal – wanted the courts to delay the election until March, 2004. He said this would give local election offices time to replace “antiquated” punched-card equipment with a modern, computerized voting system.
The request seemed entirely reasonable, since even Hottentots in Darkest Africa knew that Al Gore lost the 2000 election because Florida was still using “creaky” old card readers left over from the Cleveland administration. (Or was it the Grant administration?)
During the 2000 Recount, officials and fresh-faced media types denounced “outdated” voting equipment in the Florida counties – especially those “antiquated, unreliable” card readers used to count and recount the ballots. Pundits repeatedly asked – with eyes piously raised to heaven – when Florida would come into the 21st century.
Scene-change: November, 2002; Montgomery County, Maryland. Always zealous to be completely up-to-date, MC rolled out a fully computerized voting system for the 2002 election, designed to avoid those nasty problems that cost Al Gore Florida’s electoral votes. However, if post-election comments by citizens – both in print and on-air – are a guide, it seemed that there was still a problem.
Numerous county residents voiced doubts that their votes had actually been cast, and worried about the lack of “artifacts” – i.e., a physical ballot. I, too, had wondered when someone would finally notice that high-tech voting – much touted as the Republic’s salvation – is non-verifiable, non-traceable, and non-duplicable. The voter takes it entirely on faith that his ballot was cast as he desired and can be verified.
Yes, yes – I see those hands raised by our computer whiz-kids. Of course, backup files can record how electronic ballots were cast. But Dude! How secure is that? Haven’t the hackers shown us, by this time, how easily files can be changed, erased or replaced? How hard would it be for an able hacker to break into an election-records computer and change an election result without leaving a trace to betray the deed?
The punched-card ballots we used for years in Montgomery County, where I lived for 33 years, might have looked old-fashioned. (No! Not that! Anything but that!) But they were actually solid artifacts, repeatedly usable for recounting. Moreover, the small, mechanical devices MC provided for punching our votes were – as I recall – entirely dependable and easy to use. A child could have operated them. The card did not contain pre-perforated punches that the voter pushed out with a stylus. One can only assume that those Florida tales of balky, unreliable punched-card ballots were either much exaggerated, or were indicative of a contract filled by some election official’s brother-in-law.
An electronic ballot seems sexy and very modern, but it is non-substantial – “virtual,” as they say – and immensely susceptible to fraud and manipulation to a degree that punched-card ballots are not, provided the latter are guarded and handled with reasonable care. Electronic voting is so rich with opportunities for chicanery that those old stories of unplugged voting machines in Chicago precincts on election-day will seem as childish as Romper Room, in retrospect.
Old computer pros like me pounded our armchairs and shouted at the TV set during the 2000 Florida-recount, as nonsense poured out from ignorant youngsters who had no idea what they were talking about. “Antiquated card-readers?” The internal combustion engine is over 100 years old. Is it “antiquated?” High-speed card readers represented a technology zenith of the late 1970s, when 2400-card-per-minute machines read hundreds of thousands of cards (and millions of actual punches) per day, nearly error-free.
At that time, I worked with programs stored on 10,000 cards and containing some 1,000,000 punches. It was extremely uncommon to encounter a single error caused by machine-reading of these gigantic decks. Had these machines produced read-errors regularly, we could not have done our work.
This is the equipment Florida and other states used for vote-counting. NASA put a man on the moon with this same “creaky” gear. How did we get Neil Armstrong there (and back) if the technology was so unreliable?
We phased out punched-card media – not because the technology was unreliable, but because it was too slow. And card-storage was too bulky and too susceptible to dampness, mice, fire, flood, and clumsiness. (A dropped card-deck was the nightmare of every early programmer.)
But with ballots, the issues are different. Solidity and replication matter far more than storage or speed. The slander of Florida’s voting equipment was entirely specious. It was really a media “bums-rush” to explain why Earth-tones Al lost the election. And in that rush, nobody noticed that when you abandon punched cards, you lose the artifacts which provide proof, enable recounting, and ultimately inspire public confidence that an election has been conducted correctly and without fraud.
The real problem in Florida was not poor technology. It was improper handling of ballots – on live TV, no less. My daughter, a Maryland attorney, called me one morning during the 2000 recount.
“The TV is showing people fanning packs of ballot cards,” she exclaimed. “Unguarded card decks are lying on tables. Some cards have fallen onto the floor. Reporters are walking all over the place. This is evidence!” she practically shouted. “You can’t handle evidence that way!” (Of course, you can – I reminded her – if you hope to change an election result.)
Running scared from media hype about “antiquated equipment” is not the same as a thorough, informed examination of a process and an expert evaluation of its equipment. If we think fraud lurked in every chad of punched-card ballots, just watch for the possibilities with electronic voting.
Early in my career I worked at a Navy technical shop which analyzed tactical and strategic military problems – often using computer programs for complex determinations. One of those programs, actually supplied by Air Force analysts, compared the tactical needs of the Navy and the Air Force for a certain aircraft. The program indicated how many planes each branch would need for various scenarios. We ran it against varying sets of parameters (inputs), finding that Air Force needs varied widely over different scenarios. But the answer for the Navy always read, “The Navy gets 944 airplanes.”
Finally, someone checked the actual program-code – written in a computational language called FORTRAN. We found the direct print-statement: “The Navy gets 944 airplanes.” The answer was not “parameterized”; i.e., “944” was not a computed quantity. Those calculations (and “results”) had been used in strategic and tactical deliberations at high military levels (hopefully, only by the fly-boys). Some sly Air Force analysts had played a joke on us which they knew we would quickly discover.
Prank or not, I learned something valuable about computers the day we discovered that bit of tongue-in-cheek computational chicanery. I saw that computers are tools, not oracles. They should not be trusted, implicitly, as they are only as trustworthy as the people operating them. That goes double (maybe quadruple) for elections.
Eventually the rest of the country will learn this, too. But our headlong rush into electronic voting – which, indeed, has already occurred during the past decade – could produce a lesson more far-reaching than a few airplanes, more or less, for the Navy.
Governor Davis failed in his desperation bid to delay the recall-election by raising the (dreaded) spectre of punched-card ballots. (He lost the election, too, but not because of hanging chads, etc.) Take it from an old computer pro – the issue was, like, totally bogus.
Computerized voting has become the proverbial fox in the hen-house. Every election now produces new scandals involving it. Those new, improved virtual systems will eventually become so hopelessly hacked and vulnerable to compromise that punched card voting will actually return as an “anti-voting-fraud innovation.” In fact, this is already occurring.
Fraud is the dark cloud hanging over this key element of our republican democracy. Fraudulent registration, voting by mail, and political resistance to showing ID at the polls are bad enough problems. But if citizens become convinced that computerized voting has made our entire process untrustworthy, we’re going to be in real trouble.
Today, old computer guys like me are sitting in our easy chairs, thumping our canes and wheezing about “944 airplanes” to anyone who will listen. Not every “modern improvement” turns out to be sound or beneficial. If we want honest elections, we need to take a giant step backward from electronic voting.
Of course, this assumes that we all actually want honest elections.