Thursday, October 27, 2016

Kicking "Angels and Demons" while it's down

First, the ever entertaining Cinema Sins:



Now for some science. Let's assume that the stolen vial contained 1 gram of antimatter. Then, the explosion would release $E = mc^2 = (0.002) \times (3 \times 10^8)^2 = 1.8 \times 10^{14}$ Joule. At 4.184 petajoule per megaton of TNT equivalent, that is an explosion of roughly 43 kiloton.

(The more observant readers will notice that there's two grams in the energy computation. That's one gram each matter and antimatter.)

The operational ceiling of helicopters is around 25,000 ft, but the helicopter piloted by Ewan McObi-Wan Kenobi is nowhere close to that altitude when he jumps. An air burst of 43 kiloton even at say 10,000 ft would create a lot more damage than shown in the movie. (For comparison, Hiroshima's burst was at the yield-optimized height of 2,000 ft and with a yield of about 15 kt.)

So Ewan McThe Ghost Writer would probably be a carbonized carmelengo rather than a usurper to the Vatican throne. And probably so would the faithful in St. Peter's square and the Cardinals in Busch Stadium The Sistine Chapel.

- - - - Fait divers - - - -

Here's the page that CERN put up to address the errors in the movie.

Nikolaj Lie Kaas, who plays the assassin, plays the corrupt CEO of the energy company in the recent Danish series "Follow The Money." I guess he's typecast as the sociopathic type now.

I claim extra nerd points for using RStudio to do the computations (was already open; it's pretty much always open these days):



Sunday, October 23, 2016

Gravity "batteries"

When there's too little demand for electricity, certain grid operators (like the Portuguese one) use excess capacity to pump water from downstream of dams to the dam reservoir. This is a way to store energy for peak demand.

I understand that some mountainous region is studying the possibility of replicating this with a funicular that would operate as the water in the dam. The losses involved in moving the funicular imply low roundtrip efficiency (the ratio of the energy recovered to the energy entered into the "battery"). And, of course, the funicular can't be used for passengers, unless there's some special discount for unpredictable schedules.

At least two people have told me about a start-up (I forgot its name) that wants to solve the battery problem by using the same approach, only with dedicated masses on vertical tracks.

The tragedy of engineering is the murder of beautiful illusions by ugly numbers.

Let's say this company can use $100\%$ roundtrip-efficient motor/generators, that is, all the electrical energy that is converted into potential energy of the moved mass can be recovered as electrical energy with zero losses in the whole process. (Yes, this is a ridiculously generous assumption, but it won't matter.)

Say this company has a 1000 metric ton mass that can be raised up to 10 meters. It can therefore accumulate $98$ megajoule (MJ) or $27.44$ kWh. Sounds ok-ish for a battery, except:

1. If that mass is made of lead (density = $11.34$ kg/l), a cheap-ish dense material, its volume is 88.2 cubic meters. That's large for a battery: it's a cube almost 4.5 meters on the side. Remember that this assumes $100\%$ roundtrip efficiency motor/generators.

2. Gasoline has an energy density of $32$ MJ/l and jet fuel has an energy density of around $30$ MJ/l; using a readily available commercial-grade combined-cycle generator with around $65\%$ total efficiency, 98 MJ can be generated with less than 4 liters of jet fuel or gasoline.

Okay, the combined cycle generator takes some space, but so do the motor/generators and the support frame for the 1000 ton mass. And the space for the vertical track, of course.

Numbers. Killing illusions. No wonder so many people avoid them.

- - - -

To make up for the bursted bubble of delusion, here's the feel-good video of this week:



Sunday, October 9, 2016

Aging engineers versus experienced engineers


There are growing complaints that Silicon Valley companies discriminate against middle-aged engineers. But it might not be just ageism, it might just be aggregation error.

Engineering comprises mainly two things: a body of knowledge and a problem-solving mindset. To be a good engineer one needs an up-to-date body of knowledge in the relevant field and a facility with different problem-solving approaches used in the field (and possibly outside it as well).

(For the moment let's leave aside the problem-solving mindset; its dynamics are complicated and very situation-dependent: while some engineers acquire and develop problem-solving skills with experience, other fossilize their thinking, for example due to organizational practices.)

As part of what I do is continuing education, I have observed the dynamics of the body of knowledge as engineers' careers progress.

  • The largest group by far (sadly), makes little attempt to keep up-to-date with their field after formal education ends. In conversation, after a corporate training event, a member of this group told me that keeping up-to-date was "very nice in theory, but we don't have the time." All of us would like more time; but this person spent tens of hours per week watching TV. One of those hours per week spent updating their skill set would mean 52 hours per year, which would be more than enough (most of the participants in that event had fewer than 20h/year of training or study, and self-paced learning can be much more effective than group events.)
  • Most of the remaining engineers realized their technical obsolescence would become a problem and were retooling themselves for a management job. The main problem with this attitude is that there will always be fewer management jobs than engineers who plan to go into management. Secondarily, firms have both partially replaced management jobs with consultancy engagements and started prioritizing management-trained applicants over engineers.
  • A few engineers fell into a third category: those who keep up-to-date either because they realize the job implications of doing so or because they really love their engineering field. The problem, for those in this group, is that their small number makes them liable to be categorized into one of the other groups.

Placing ourselves in the position of Google, for example, the decision to consider a candidate who's been out of formal education for several years versus considering one that's just graduated --- even if Google believes that the energy of youth can be balanced by the temper of experience --- comes down to which of the three groups above the older candidate will fall into.

In the absence of good information, statistically the older candidate will be in the first group, in other words, aged, not experienced, a distinction that most of the engineers can but will not make (as it defeats their case).

(The younger candidate's type is irrelevant, because being fresh from school means an up-to-date skill set, at least for the near future.)

There are obviously many confounds: consider a choice between a newly minted computer engineer from Idaho State - Tubertown with no code to show (not even from school projects) versus a 45-year-old Caltech graduate class of '95 who has code on GitHub that is particularly relevant to the job, for example.

For the other engineers, who have been lax in their updating of skills, there's a solution: it's never too late to learn. And then: show, don't tell.


Sunday, October 2, 2016

Much ado about "loving" science





Science identity products

Science identity products like t-shirts, mugs, posters, and computer wallpapers are used to signal that the owner has an interest in science; unfortunately, because this interest in science has become fashionable -- at least in some segments of the population -- poseurs also buy these objects, lowering the quality of the signal.

I've written often (one, two, three, four, five times at least) about the problems with using science as an identity product, as with the people who "love science" as long as they don't have to learn any.

These products aren't necessarily only appealing to poseurs, though. People with a real interest in science and in science education also like them for, among other reasons,

1. Identity signaling. Like the poseurs, except it this case it's a real signal. People want to communicate their interest in science and the beauty of some scientific results and natural phenomena. (I own quite a few science identity products myself.)

2. Recruitment. These products can be useful motivators for bringing newcomers into an appreciation of science. By showing that there are other nerds geeks people interested in science, they create social conditions for others to come out as nerds geeks people interested in science.

3. Mere exposure. People like or at least feel more comfortable with things that appear familiar. The more exposure people have to scientific concepts and images, even if as part of jokes or background material in sitcoms like The Big Bang Theory, the less aversion they may feel when science content is presented to them.

There's one possible disconnect undermining these three points, though: that people who are influenced by exposure to the science identity products only like the aesthetics:

Problem with science popularization



The big problem with the poseurs, which is a real problem not just my "I liked that band before it was cool" complaint, is not that they use the products to pretend to like science, though that would be bad enough. The real problem is that poseurs know that they don't actually like (or know) real science, so they feel threatened by those who do and take action to counter that threat, usually distracting from the science.

As my previous post showed, many poseurs in the media try to be "sciencey" but they fail miserably because in the end they don't understand that science is not like literature or art where the judgment of some other people is what matters. In science, reality is what matters. Poseurs don't get that, because to them reality is whether others buy into their pose.


Popular science content

Making science accessible to the general public is one of the most effective ways to improve society: it allows more people to partake of the benefits of knowledge (for example, avoiding junk science and quackery), it helps garner support for scientific enterprises that require public funding, and it creates the foundations for new generations with more and better scientists.

My ideal science popularization

The problem is that popularizers can be real science popularizers or they too can be poseurs. And the poseur popularizers tend to be more popular. The glaring exception is Carl Sagan, but that's because he was both a pioneer in popularization and a real research scientist prior to that.

The most obvious difference is that Carl Sagan's Cosmos was designed to impress people with the power of science, while many current popularizers design their programs to impress upon the audience (a) how special they, the audience, are; and (b) how smart, knowledgeable, and suave the popularizer is. There are some exceptions, but they aren't the most successful popularizers, at least not on TV.



A rule-of-thumb that works for me is to ask whether the popularizer is an active researcher (or was until recently active) in the field. People whose job is some variation of "science popularizer" tout-court, even if they have some scientific training (which many of them don't), tend to focus on people and events rather than concepts and principles. In other words, they popularize the story of science rather than the actual science. (In many cases they either avoid the science completely, or they get most of it wrong.)

This rule works for two reasons:

First, an active researcher will know the science better than a non-researcher popularizer. This IMNSHO more than balances any communication advantages the non-researcher might have. One of the hilarious examples of this advantage is The Igon Value Problem, where active researcher Steven Pinker takes on the intellectual lightweight Malcolm Gladwell. (But supporting my observation above, Gladwell is more popular than Pinker.)

Second, an active researcher has to protect his/her reputation in the field. This adds motivation to get things right to the knowledge (the ability to do things right). When no one in Astrophysics takes you seriously (because you  call yourself a scientist but your career total citations of 150 mark you as a museum manager), you can say ignorant things on twitter about planes and helicopters. An engineer who wrote nonsense like this would be mocked at any future technical conferences he/she attended:



Personally I decided to read textbooks in lieu of popularization books,* but there are some popular books I've read that I found worthy of recommendation, so here are two for now:




Deep science (or other technical) content

Leaning technical material is something that requires audience (perhaps in this case "student" would be the better term) participation.



Lectures can motivate study and are a good introduction to the material, but only self-paced study and practice exercises can make technical material stick.

There's a qualitative difference between (to quote again from my old post about Heisenberg) understanding that this is a joke, i.e. popularizer-level understanding:
Police officer: "Sir, do you realize you were going 67.58 MPH?
Werner Heisenberg: "Oh great. Now I'm lost."
and being able to completely spoil the joke by computing the actual uncertainty (deep understanding):
A simplified form of Heisenberg's inequality, good enough for our purposes, is 
$\qquad \Delta p \, \Delta x \ge h $ 
Going by orders of magnitude alone, assuming that the mass of Heisenberg plus car is in the order of 1000 kg, and noting that the speed is given to a precision of 0.01 mi/h, an order of magnitude of 10 m/s, with $h \approx 10^{-34}$ Js, we get a $\Delta x$ of the order of 
$\qquad \Delta x  \approx \frac{ 10^{-34} }{10 000} = 10^{-38}$ m.
Only practice and study can create the kind of deep understanding that allows you to spoil people's fun at parties with numerical sidebars like this. Certainly something to aspire to...

That's not to say that lectures don't have value; I think of them as the warm-up sets you do before actually exercising. In that sense, they are very important, since they provide a passive experience that gets the material into context, setting up the active experiences of self-paced study and practice exercises.

Walter Lewin, shamefully retconned out of OCW and their official YouTube channels by MIT for undisclosed non-scientific transgressions, was one of the best Physics instructors online; even better than Feynman, since Lewin used actual in-class demonstrations and calculations matched to the examples. Here's a great class on standing waves:



As for most MOOCs, I already wrote my views in my post "MOOC-rize this," and the TL;DR is:

1. MOOCs have economies of scale in production and diffusion, but the difficult parts of education, personalized attention, for example, don't scale.

2. MOOCs can derive brand equity from the institutions associated with the teaching, but whether that brand equity is deserved is an open question: there are many components to education beyond what most MOOCs offer. I made some observations about that regarding the Kenan-Flagler Online MBA.

3. MOOCs built out of classroom teaching and associated materials are audience-targeted; a course like Lewin's works well at MIT and possibly CalTech, but the speed of exposition and the amount of off-classroom work that Lewin expected from his students will not work for most other universities. Other materials, like textbooks, may partially make up for this, but even so most students would probably prefer better match between materials and audiences.

4. The major weakness of MOOCs as they exist now is the lack of evaluation and, in many cases, of ways to check your exercises. Since audiences (students) learn from these exercises, done individually and then corrected by a knowledgeable instructor, this is actually a much bigger weakness than I noted on the "MOOC-rize this" post.


In conclusion

There's nothing wrong about being out and proud as a nerd geek someone who likes science; take care to avoid poseurs, both individuals and media darlings who don't have a track record of research; and if you want to learn more (kudos to you), there are plenty of MOOCs and other free resources to help you. One of those resources is called a Public Library and for the effort of getting a library card you can get a good education because in the end what matters is that you want to learn.


In the end what matters is that you want to learn. Poseurs don't want to.




-- -- -- -- -- -- FOOTNOTE -- -- -- -- -- --

* I find textbooks to do a better job than popularization books since I want to learn things at a more proficient level than a passing understanding. This requires time and effort, but I like it. (Hey, I lift heavy weights for no reason other than I like lifting heavy weights, so this isn't that different.)

The one enormous barrier to this approach is the ridiculous cost of textbooks in the US. I was interested in molecular biology, so I got Molecular Biology of the Gene, I believe for its weight in gold. There's now a new edition which costs its weight in diamonds, so I won't get that. Note that this is a personal interest in molecular biology; this is not work-related or anything monetizable, so the $\$200$ are a hobby expenditure. Which is fine, but still could discourage others from buying such an expensive book for a hobby.

I rationalize the cost by reminding myself of an old business associate who spent $\$150$ on a date with someone who, according to his later report, made Lady Macbeth sound warm and cuddly. So, that's about 3/4 of a textbook he could have bought there...