International Policy Digest

Phil Roeder/Flickr
Tech /06 Feb 2020
02.06.20

Digital Nightmares

When I sat down in front of the TV to watch the results of the Iowa caucus, the memories came flooding back, all of them bad. There were no results. It was déjà vu all over again as my mind raced back to 1984 when I was a relatively young punk systems analyst in a certain Minnesota County who had been given the thankless task of digitally tabulating ballots. Oh, I didn’t know the job was thankless at the time, but I soon found out just how thankless it was…and that wasn’t even the worst part of the experience. The County’s ballot-counting system was archaic, just barely computerized. Punch cards, with those infamous “hanging chads” that would become so important in 2000. Setting it up took forever. ‘How can this thing possibly work?’ I asked myself. But once I did get it set up and I tested it thoroughly it amazingly did work! It was slow, but accurate, if you paid close enough attention to its care and feeding.

On election night I was confident it would work. The only weak link as I saw it was the punch card reader which had been retrieved from storage where it had been gathering dust since the last election as it was no longer used for anything else. But I even had that covered as I had acquired the services of a grizzled old ex-computer operator from the previous computer age when punch cards and plugboards reigned supreme. He stood, or rather sat as his knees were bad, at the ready in his garage mechanic’s outfit toolbox close at hand. And he got the job done too, gently coaxing the balky old card reader slowly through yet one more election.

The polls closed at 8:00 pm. Election judges began bringing in their steel-boxed cards around 9:00. All they had to do was count the cards by hand and agree on a total. At this point the “human factor” kicked in. Most of the judges were in their 60s to 80s. They had been doing punch card elections since Ike was president, so they knew the drill, but it had been a long day for them-6:00 am setup to 8:00 pm closing time and beyond. They were tired and cranky, and a bit confused at times. There had been a big turnout and some precincts had taken over 2,500 ballots. The law said that three judges had to agree on the total, and that total had to agree with the card reader count. Aye, there was the rub. Everyone was screaming for results, but they were only trickling in. When we finally finished it was 5:00 am. We had a good, accurate result. I breathed a deep sigh of relief, and thought things were over, but they weren’t, and wouldn’t be for many moons.

If you want to meet all the soreheads, crackpots and conspiracy theorists in any given city, state or county just run an election count. I spent a lot more hours trying to defend my actions on Election Night than it took to run the count. And in the end, nobody was satisfied. Some were so dissatisfied that they decided to take matters into their own hands. One municipality decided to go back to the “good old days” of paper ballots and hand counting. They finished their 1988 count two days after the election. That count was disputed and had to be counted again by the state. Another town hired a nerdy high-schooler equipped with a $500 TRS-80 personal computer to do their own counting software. That model was known as the “Trash-80” and though the kid worked cheap his software didn’t get approved by the Minnesota Secretary of State, so they had to do a glacial hand count too. But before November ’88 I had moved on. The prospect of having to do another election count spurred my decision to seek other employment.

If you can get dumped on for running a correct count that has the sole disadvantage of being slower than expected (even though it was faster than the County’s 1980 count) how badly can you be crucified for a count that’s an unmitigated disaster? We’re now going to find that out, though I suspect we won’t find it out soon (see Florida 2000). Some of my 1984 bogeymen are still around in 2020. The expectation of “instant returns” is even stronger than it was then. The old refrain of “if the results aren’t in yet something’s up!” is still alive and kicking.

I strongly suspect that many of the caucus chairs and election officials are still elderly, especially as this is Iowa. Old and tired after a long day of balloting. They don’t have punch cards to deal with now, but they have smartphone apps which need to be uploaded. My own experience uploading apps is rather spotty. Some have worked just slick. Others have tested my patience. Still, others have annoyed me to the point where I decided that I was better off without them. And then there’s always the problem of knowing if they’ve loaded successfully. Many caucus chairs either didn’t get their apps loaded or weren’t able to fight their way through the apps’ confusing security to use them. Any other folks out there ever have problems with systems security? What if you had those problems in the middle of an election count where the whole country is watching? Does anyone remember Alan Arkin’s line as Lt. Rozanov in The Russians Are Coming! The Russians Are Coming!: “There is World War Three and everybody is blaming you!”

So, whoever you blame please don’t blame the troops in the trenches. An app that was either untested, or not fully and realistically tested. A backup system that was overwhelmed when the app failed. An app that was not needed in the first place. What’s the matter with calling in the results? Not fast enough for you? Not up to date? I’ve heard those refrains from some of my cutting-edge friends. Friends with wonderful new apps they insist that I use. I generally pass on them; and if I was in the business of selecting software that had anything to do with elections and ballot counting, I would be extremely cautious. Don’t throw anything that works out until you are sure its replacement works. And don’t forget about the people who end up having to run your brave new apps. They need more training than just the assurance that “it’s so simple a child could use it.”