Tuesday, September 3, 2019

Software Problems, Gastric Reflux, and the Cult of Eschewing Blueberry Yogurt

If there's a replicable problem with some software, no problem! You write to the developer, reporting that when you hit the button while pressing the key the duck's supposed to quack, but...no quack. The developer tries it himself, and is able to replicate your replicable problem. The bug is squashed, and presto. All fixed!

But some problems are intermittent. You search for a reliable trigger, and the developer struggles to determine the source, but it's all maddeningly unpredictable. Essentially random! Over time, a few dozen, or hundred, or thousand, or million people come to share your anguish, all searching for a cause and a workaround. This can persist for years, even decades.

As a wonky Mac nerd, I've watched this scenario unfold in online forums hundreds of times. The exact same patterns predictably appear each time.

There are the moralists. They, themselves, are not experiencing the problem, which clearly means the others have only themselves to blame. They're morons who are using their computers wrong. Right-users are just fine.

Then there are the drones who offer the same mindless, rote bundle of suggestions for any unexplainable problem. It makes them feel/seem smart, and hey, it might help (this mindless scattershot approach is like asking everyone in a bar to sleep with you on the odd chance someone says yes). Repair permissions! Zap the parameter ram! Reinstall your system software! Clean your keyboard with a Q-Tip and isopropyl alcohol! Folks take the day off from work to diligently work down the list, and not once in the history of computing has it ever fixed a widely-experienced intermittent problem.

Eventually religions form. Human minds cannot tolerate randomness. We resourcefully find some sort of order in even the most staticky white noise. This is what humans do; what we've evolved for (among other things, it makes conspiracy theories attractive). We need to know, and if we don't know, and can't know, we'll concoct some flimsy explanation and make ourselves believe it, religiously, even if the theory keeps disproving itself.

Someone announces that they've finally found The Answer. If you press the button seven times, turn up the volume then turn it back down again, fluff your pillows and eat a danish, then restart your computer twice in 17 second intervals, that seems to work!

Others will follow these arcane instructions and, praise cheeses, find blessed relief. Why? Because intermittent problems are intermittent, which means literally anything you do seems to fix it. Thus a cult is formed, and, like any incipient religious movement, it will reflexively disregard contrary evidence. Yes, the problem came back, but I might not have waited exactly 17 seconds. Also, it's raining. Anyway, it's better now, but there's some other x-factor we haven't quite pinned down.

Meanwhile, other religions arise, offering various incantations and mythologies. Leff's Law of Intermittent Software Problems: the moment when the most people are the most confused is inevitably the moment of the greatest unflinching certainty that it's all been pinned down (non-coincidentally, this is also Leff’s Law of Religion).



Gastric reflux is notoriously hard to pin down. Droning doctors will hand you a long list of things to try doing (and to not do). The scattershot approach. No one in the history of reflux has ever been fully fixed by following these rules. Meanwhile, moralists who don't have reflux will tell you (however politely) that you're fat and slovenly and eat poorly; that you're an alcoholic and a general moral failure.

As you dive into online discussion, you'll find a rich array of religious cults built around reflux. You may even create one yourself! "I've figured it out!" you'll at some point announce. "Blueberry yogurt! That's the problem! When I eat blueberry yogurt, I get reflux, and when I don't, I'm fine!"

Lots of problems here. First, "always" getting reflux means it happened twice (three times, tops). And since you don't eat blueberry yogurt all the time, and don't get reflux all the time, odds are high that on any given day you won't eat blueberry yogurt and you won't get reflux, giving a false impression of success. What's more, you're not factoring in all the times in the past when you've eaten blueberry yogurt without a problem. That's not a tabulation you've been running all these years because it's not super interesting. I've probably scratched my left ear twelve million times without triggering a migraine.

Here's how you really get roped in: Amid the vastness of the Internet and its infinite typing monkeys, you'll discover 4 or 5 or 26 fellow travelers who also believe that blueberry yogurt is the answer. Sweet corroboration! This evidence locks in your conclusion, and a new congregation has formed: the cult of eschewing blueberry yogurt.

Faith will soon be tested. On some days without blueberry yogurt, you'll still get reflux (but NOT AS OFTEN!!!!). On some days without reflux, you will have eaten blueberry yogurt (a FLUKE!!!! There's also an X-FACTOR!!!!). These results would challenge the faith of non-believers, but you are fervid, standing by your conclusion while acknowledging that the problem is subtle and multi-faceted. However: you're making progress in sussing out the full truth!

The intermittence of the problem has sucked you in. It's like a reverse slot machine that pays out continuously and briefly stops once in a while. No matter what you do, you'll feel smart - you're doing it right! - most of the time. Each new move feels like it's working, while every queasy bout feels like some nagging x-factor you've not quite closed in on.
In a bullish market, day traders feel like geniuses, because the overall rising market proves every move wise. But markets always dip!
Progress seems to be made. Most of the time you feel good! Hard-won insight like the blueberry yogurt revelation has helped a lot! If you’d made even a cursory effort to track and log your condition - nobody does! - you'd observe that the outcome has always been more or less constant. Nonetheless, we're getting on top of this thing. We're all considerably above average.


Intermittent problems are the big peril with refurbished tech. Easily detected, well-understood problems will generally be addressed, but items may be returned due to intermittent issues, and there's no protocol for checking/fixing them because they're inherently hard to detect (much less repair). Of course buying new is also a risk, but I'd imagine that returned items turned around via refurbishing tend to have a greater frequency of such issues.

No comments:

Post a Comment