Thursday Throughput: Leap Year Edition
[ThTh1] In a little over a week, we will have an extra day in February. Why do we do this? Because the time it takes Earth to revolve around the Sun is not exactly 365 days. There’s a little under a quarter day extra. So if you didn’t occasionally adjust the calendar, our seasons — which are defined by our position relative to the Sun — would start to change, moving in the calendar until January was the height of summer. Here’s a great illustration
This is why we have Leap Years pic.twitter.com/bCX31kymMW
— Dr James O'Donoghue (@physicsJ) February 6, 2020
Leap days actually over-correct slightly. So every century — except for those divisible by 400 — we skip leap day to get us back on track.
[ThTh2] Are medical errors really the third biggest cause of death? Orac breaks it down. The short version is probably not. Those estimates rely on second-guessing doctors and making very large assumptions of what medicine is capable of.
[ThTh3] Should you avoid having kids to fight climate change? No. Future generations will burn a lot less carbon than we do. And they will be needed to fix what we’ve broken Earth-wise. We should be alarmed not impressed when people don’t want to have kids because of climate change. Because what it really means is that they’re giving up hope.
[ThTh4] I have no idea if it is even possible to plant a trillion trees to combat climate change. That’s over 100 for every man, woman and child. It’s not even clear the biosphere can support it. But it’s worth a try. And if this moves the Republicans to a more proactive position on the climate, that’s great.
[ThTh5] A good primer on the tension in the Hubble Constant, one of the most fundamental measures of the state of the universe. I should note that things may be getting better on this front as the systematics in the measures are beaten down.
[ThTh6] Alphabet is shutting down their power kites experiment. It didn’t work out. But this is the kind of crazy idea for fighting climate change we should be looking at. High risk but high reward.
[ThTh7] Immune cells that might kill cancer.
[ThTh8] NASA reworked Voyager’s “pale blue dot” image and the result is stunning. It reminds you of how small we are in such a big universe.
[ThTh9] One of the debates raging in social science right now is the gender-equity paradox. In countries with more overall gender equity, women are a smaller percentage of STEM students than in countries with less equality (think Sweden vs. Algeria). The explanation for this has been that, in countries with little gender equity, science is one of the few fields where women can advance. But in more equitable societies, they are less restricted. It is further argued that while women and men are equally capable in science, science tends to be men’s best subject, while other fields tend to be women’s best. The Law of Comparative Advantage indicates that people will gravitate toward what they’re best at.
This idea has come under fire in recent papers, which question both the existence of the paradox and its interpretation. You can read an explanation and defense here. I don’t think we know nearly enough right now to explain the gender-equity paradox. But it’s throwing a spanner into what once seemed like a fairly straight-forward issue.
[ThTh9] the explanation I’ve seen is that countries with less gender equality don’t treat STEM as a prestigious occupation; the Men do Important Man Jobs like Lawyer and Doctor and Banker and Politician. Countries with enlightened and inventive populations recognize the value of STEM, and therefore treat it as an Important Job, which of course must be done by Men…Report
That seems as plausible to me as any other explanation. People like to point out that most computer programmers were women when it was regarded as secretarial work. Once it started to pay well, men took over.Report
People like to point out that most computer programmers were women when it was regarded as secretarial work.
Men have been the majority of programmers since long before programming started paying really well. I’d be interested in seeing the statistics for earlier, but this has never been true in the microcomputer era.
It’s possible that in the very early years of mainframe computing, women were overrepresented among programmers while men worked on the hardware, but this would have been a very small number of people.
Journalists have been waging a smear campaign against the tech industry for about a decade now, one prong of which is trying to push a narrative about white men (because Asians don’t exist) driving women out of the tech industry.
All the evidence points to the low representation of women being driven by lack of interest rather than sexism. Notably, in the US, women peaked as a share of CS majors (37%) back in 1983; the idea that either society in general or CS departments or the software industry in particular has become more sexist since then is dubious, to say the least. Female (and black and Hispanic) representation in the software industry is about what you’d expect given the demographics of the pipeline.Report
Oh, I could see it getting more sexist, but not as a driver, rather as a result.
If you have women exiting the field for reasons not related to the various sexist cultures (changing careers, starting a family, etc.), you can get a critical mass of sexist thinkers that move the culture.Report
ThTh1: Nothing about leap seconds? Those are turning out to be problematic for soft real-time applications, since they occur irregularly and can be either positive or negative in sign.Report
We haven’t had one since 2016. If we get one in the future, I’ll write about it.Report
Leap seconds strike me as something that’s important for astronomers and nobody else.Report
A surprising amount of software behaves badly if there aren’t 60 seconds in every minute. There were a number of unexpected software glitches when the 2016 leap second occurred. At least one large-ish service shut down for an hour before that one, just in case. It’s enough of a problem that Google (and others) don’t implement it as a single second, but “smear” it out over 24 hours with their time servers running very slightly fast or slow (depending on whether a second is being dropped or added).Report
I think the problem with leap seconds is the software engineering community just kind of decided to treat UTC as the single default time system, as it is pretty good for that most of the time, but ignored the fact that it actually isn’t ideal for timestamps, because of leap seconds.Report