February 2019
Volume 34 Number 2
[Don't Get Me Started]
Our Event Horizon
By David S. Platt | February 2019
.jpg "David Platt")
This column is running in the February issue of MSDN Magazine, but I had to submit it before Christmas. The turn of the year, with darkness stubborn at the edges of the day, leads me to contemplate life, the universe and our shared profession. Today I’m pondering how this crazy computing business started—with Alan Turing, of course. I’m re-reading Charles Petzold’s book, “The Annotated Turing” (Wiley, 2008), an excellent guide to our first guru and his foundational paper.
Turing wasn’t originally thinking about real computing machinery. Though he later went on to help build some hardware, the concept of cat videos in everyone’s pocket probably never crossed his mind. Rather, his Turing machines were thought experiments imagined in response to Kurt Gödel’s First Incompleteness Theorem (1931). Gödel proved that if mathematics were consistent, theorems would have to exist that were true, but could never be proven—not because we weren’t smart enough, but because of the basic structure of the universe, like Heisenberg’s contemporaneous Uncertainty Principle. Turing wanted to find ways to identify these unprovable theorems. Could any finite algorithm do that? (Spoiler: No.)
What does that have to do with today, I hear you wondering. Here’s what. I was having dinner with my parents, and my mother asked me to check the score of the Patriots (American football) game. As I wrote last December, accessing this information is difficult for seniors, and shouldn’t be (see msdn.com/magazine/mt832865). But I still have enough dexterity to Google “New England Patriots” on my phone, and enough visual acuity to identify them beating up on some hapless opponent. Then my father asked me to find the blessings over the Chanukah candles that we would start lighting that night. Again, a quick Google, and bang, right there were my answers, in Hebrew, English translation and transliteration; even videos of professional cantors singing them, with and without cats (youtu.be/AQvt5mMzJVA and youtu.be/YZD0QXPj_XI, respectively).
That encounter started me thinking: Just about any question of fact can be answered by Googling. But, similar to Turing, I pondered which questions cannot, and how we can identify them.
I’m not asking whether Google currently contains a particular item, as Turing wasn’t investigating the proof of any particular statement. Google doesn’t know where my car keys are right now (damn), but if I put a Tile tracker on my keyring, it could find out. What kinds of questions can Google never answer, no matter how big or how good it gets? Where is its quantum limit, its event horizon?
Google can’t easily decipher the spin of human languages. So it can’t tell me if a government payment for installing solar panels on my roof is an incentive (good) or a subsidy (bad). Am I foolishly taking out a second mortgage, or wisely adding a home equity line of credit to my comprehensive financial plan? Is this annoying software behavior an issue or a bug (see msdn.com/magazine/ff955613)? Am I a cranky old fart, or a noble silverback? Google can’t say.
My daughter Lucy pointed out that it’s hard to Google one’s own medical symptoms, except in trivial cases. (“OK Google, my nose is bleeding!” “Hmm, I think you have a nosebleed.”) The signal-to-noise ratio of the medical world is too low, and humans too suggestible.
Knowledge of horrible but obscure diseases was once limited to heavy, expensive, paper medical textbooks. Students would scare themselves silly with induced hypochondria so common that it earned a name: “medical student syndrome” (see bit.ly/2GVcRlZ). “Tingling in my left nostril, only on Tuesdays, hmm, must be McCloskey’s disease. It says here my nose will fall off next week.”
All of that information is now online, instantly searchable by any civilian. “Wait, does my pinky hurt too? Incipient Ebola, has to be.” Plus, as a cancer-surviving reader pointed out to me, your search results are swamped by people trying to make money off you. Google can’t help much here, either.
I’ve only been working this problem for about a month, so I haven’t fully generalized these results. What do you think so far? Gödel, Heisenberg, Turing and Plattski?
David S. Platt teaches programming .NET at Harvard University Extension School and at companies all over the world. He’s the author of 11 programming books, including “Why Software Sucks” (Addison-Wesley Professional, 2006) and “Introducing Microsoft .NET” (Microsoft Press, 2002). Microsoft named him a Software Legend in 2002. He wonders whether he should have taped down two of his daughter’s fingers so she would learn how to count in octal. You can contact him at rollthunder.com.