Comments on The Week in Geek

Use of that cellie jammer will land you in the pokey, but I swear to God, I think it might be worth it.

In other words, I think it might be worth it.

At least one waiter in every restaurant needs to have one.

A cellphone jammer would only be more dangerous for bicyclists. Once you get 60 feet from a cellphone wielding jackass, his signal cuts out. Then the driver starts paying even more attention to the phone, while trying to reestablish a connection, and less attention to the road.

Well you also could report that my first report of Bungie leaving Microsoft was 100% right on the money.

@3

Fucking right on.

I took a class on knot theory my last semester of college. I've got some pretty funny but SUPER SUPER nerdy stories.

Well, you little tease Ben... Share.

I think the Human Pool Table would confuse Seattleites worse than a 4-way stop sign WITH pedestrians ready(?) to cross.

I will try the caffeine nap. My personal research concludes that 2.5 hours of sleep is the dumbest move one can make.

So one of the characteristics of a knot is its crossing number, which, obviously, is the number of times the strand crosses over itself. There's no known formula for predicting the number of distinct prime knots (knots that cannot be split into other knots under particular conditions) of a given crossing number, but we do know how many up to a certain point, i.e. there are no one- or two-crossing prime knots, one each of three- and four- crossing knots, etc. up to 1388705 for sixteen-crossing prime knots.

We covered this pretty early on in the course, and some of the other students decided to try and estimate how many seventeen crossing knots there were by basically looking at the numbers up to that and guessing. I was taking a numerical analysis class at the time, so I decided to try my hand at it, too, but via another route.

I took all the seventeen known data points (Zero crossings through sixteen crossings) and numerically interpolated the sixteenth-degree polynomial that matched all the data points. The next day in class, after the other students have given their results, I mentioned that I had been looking at the table they had in the book and 'noticed' that all the values corresponded exactly with this particular sixteenth-degree polynomial.

IF YOU DO NOT GET THE JOKE: 1) It's probably the most straightforwardly stupid way to go about doing it. 2) The formula was fucking hideous. Beyond having seventeen terms, the coefficients were the most godawful fractions you'd ever seen, with forty-five-digit numerators and denomenators. Just fucking awful. And I just 'noticed' that this hideous abomination of a formula fit the datapoints. Uh-huh. 3) The way interpolation works is it gets you an okay-ish approximation within your data set, but once you move outside that dataset, it goes wildly out of reasonable range, and the more datapoints you have, the higher the degree of your polynomial, and so the more rapidly it shoots away from the actual value of whatever you're estimating. So my final estimate for seveteen-crossing knots was something like a million bajillion fafrillion.

So yeah, pretty fucking nerdy.

Pilot Wings!

Got to get me some for the Wii!

I'm a motorcycle rider and get just as nervous as you bicycle riders when I see a motorist jabbering away, and had earlier given some thought to getting a cell phone jammer, but came to the same conclusion as keshmeshi above. Those folks whose calls got dropped would be paying even less attention to their driving, as they tried over and over to redial. I think that shooting them would be safer (and more fun).