There were just a handful of comic strips that mentioned mathematical topics I found substantial. Of those that did, computational science came up a couple times. So that’s how we got to here.

Rick Detorie’s **One Big Happy** for the 17th has Joe writing an essay on the history of computing. It’s basically right, too, within the confines of space and understandable mistakes like replacing Pennsylvania with an easier-to-spell state. And within the confines of simplification for the sake of getting the idea across briefly. Most notable is Joe explaining ENIAC as “the first electronic digital computer”. Anyone calling anything “the first” of an invention is simplifying history, possibly to the point of misleading. But we must simplify any history to have it be understandable. ENIAC is among the first computers that anyone today would agree is of a kind with the laptop I use. And it’s certainly the one that, among its contemporaries, most captured the public imagination.

Incidentally, Heman Hollerith was born on Leap Day, 1860; this coming year will in that sense see only his 39th birthday.

Ryan North’s **Dinosaur Comics** for the 18th is based on the question of whether P equals NP. This is, as T-Rex says, the greatest unsolved problem in computer science. These are what appear to be two different kinds of problems. Some of them we can solve in “polynomial time”, with the number of steps to find a solution growing as some polynomial function of the size of the problem. Others seem to be “non-polynomial”, meaning the number of steps to find a solution grows as … something not a polynomial.

You see one problem. Not knowing a way to solve a problem in polynomial time does not necessarily mean there isn’t a solution. It may mean we just haven’t thought of one. If there is a way we haven’t thought of, then we would say P equals NP. And many people assume that very exciting things would then follow. Part of this is because computational complexity researchers know that many NP problems are isomorphic to one another. That is, we can describe any of these problems as a translation of another of these problems. This is the other part which makes this joke: the declaration that ‘whether God likes poutine’ is isomorphic to the question ‘does P equal NP’.

We tend to assume, also, that if P does equal NP then NP problems, such as breaking public-key cryptography, are all suddenly easy. This isn’t necessarily guaranteed. When we describe something as polynomial or non-polynomial time we’re talking about the pattern by which the number of steps needed to find the solution grows. In that case, then, an algorithm that takes one million steps plus one billion times the size-of-the-problem to the one trillionth power is polynomial time. An algorithm that takes two raised to the size-of-the-problem divided by one quintillion (rounded up to the next whole number) is non-polynomial. But for most any problem you’d care to do, this non-polynomial algorithm will be done sooner. If it turns out P does equal NP, we still don’t necessarily know that NP problems are practical to solve.

Bil Keane and Jeff Keane’s **The Family Circus** for the 20th has Dolly explaining to Jeff about the finiteness of the alphabet and infinity of numbers. I remember in my childhood coming to understand this and feeling something unjust in the difference between the kinds of symbols. That we can represent any of those whole numbers with just ten symbols (thirteen, if we include commas, decimals, and a multiplication symbol for the sake of using scientific notation) is an astounding feat of symbolic economy.

Zach Weinersmth’s **Saturday Morning Breakfast cereal** for the 21st builds on the statistics of genetics. In studying the correlations between one thing and another we look at something which varies, usually as the result of many factors, including some plain randomness. If there is a correlation between one variable and another we usually can describe how much of the change in one quantity depends on the other. This is what the scientist means on saying the presence of this one gene accounts for 0.1% of the variance in eeeeevil. The way this is presented, the activity of one gene is responsible for about one-thousandth of the level of eeeeevil in the person.

As the father observes, this doesn’t seem like much. This is because there are a *lot* of genes describing most traits. And that before we consider epigenetics, the factors besides what is in DNA that affect how an organism develops. I am, unfortunately, too ignorant of the language of genetics to be able to say what a typical variation for a single gene would be, and thus to check whether Weinersmith has the scale of numbers right.

This finishes the mathematically-themed comic strips from this past week. If all goes to my plan, Tuesday and Thursday will find the last of this year’s A-to-Z postings for this year. And Wednesday? I’ll try to think of something for Wednesday. It’d be a shame to just leave it hanging loose like it might.