Wednesday, February 25, 2015

Geek: Oopsie. Major oopsie!

Geek:  Oopsie.  Major oopsie!  The basic sort library functions in Java (including Android) and Python are broken.  Broken as in they can sometimes crash, given the right inputs.  Java's sort implementation is over ten years old, and (to the best of my knowledge) this bug hasn't been reported before.  One particularly interesting aspect is that the bug was found during an attempt to formally prove the algorithm's correctness.  If you're not a computer science geek, you might not realize just how difficult proving correctness is.  In fact it is so difficult that almost exactly zero of the software you use everyday is proven to be correct...

The consequences of affirmative action, part 72,663...

The consequences of affirmative action, part 72,663...  This isn't really a new story, but more a recent example of an old and familiar one.  The so-called “affirmative action” programs are really about dictating college admission and graduation outcomes in a race-neutral way.  I'm over-simplifying here, but the general idea is that if 15% of the population is black, then 15% of the admissions to a university should be black, and 15% of the graduates should be black.

If, in fact, the qualifications of applicants were race-neutral, then this would make perfect sense.  However, the reality is that blacks on average are less prepared for university than whites.  There are a variety of reasons researchers have put forward to explain this, and many of them aren't based on genetics at all.  For example, some postulate that black culture de-emphasizes the importance of education, and even stigmatizes it.  Whatever the reasons actually are, the reality is that blacks of college-admission age are less prepared than whites.

Both blacks and whites are less prepared than Asian students (or Jewish students, from a religious perspective).  Once again, most researchers postulate non-genetic, usually cultural, explanations for this.  But whatever the reasons, Asians of college-admission age are, on average, more prepared than either blacks or whites.

To implement an affirmative action program that dictates admission and graduation outcomes when preparedness isn't race-neutral, obviously the standards for admission (and graduation) also cannot possibly be race-neutral.  There is no other way to get equal outcomes when the “inputs” are unequal.  Therefore admission standards under affirmative action programs are most difficult for Asians, and least difficult for blacks.  That's pretty simple and straightforward – but it leads to outrage amongst the students held to a tougher standard, and amusing denials from the affirmative action crowd (who don't like to admit that they're fighting racial discrimination with more racial discrimination).

The solution seems obvious to me, and it's the same solution propounded by both Martin Luther King and Clarence Thomas: end discrimination of all kinds.  As Thomas famously said:
...there is a 'moral and constitutional equivalence' between laws designed to subjugate a race and those that distribute benefits on the basis of race in order to foster some current notion of equality. Government cannot make us equal; it can only recognize, respect, and protect us as equal before the law. That affirmative action programs may have been motivated, in part, by good intentions cannot provide refuge from the principle that under our Constitution, the government may not make distinctions on the basis of race.

Extortion trumps cooperation?

Extortion trumps cooperation?  This article talks about a new solution to the classic game theory problem usually dubbed “Prisoner’s Dilemma”.  This new solution says that in some cases a strategy they dub “extortion” can beat the traditional cooperative solutions.  This is an interesting result, and the piece is fun to read – but that's not what caught my eye.

One of the authors of the new solution is Freeman Dyson, the famous physicist with a fascinating career.  He also happens to be 91 years old, and it still (obviously) a productive scientist.  In the last decade or so he's been a vocal skeptic of the anthropogenic global warming hypothesis, and his skepticism was one of the factors driving my own.

Now here he is publishing a paper that's rather far afield from his usual haunts.  I'll take that as evidence of a curious mind, a health skeptical outlook, and of those little gray cells still working just fine despite his advancing years...