Thursday, September 18, 2014

Steven Pinker and the Ivy League School

I nearly didn't read this article, because i haven't read Deresiewicz’s article. And, Deresiewicz’s article doesn't sound like a good use of time. I also haven't seen the movie Love Story, probably for the same reason that i haven't watched Titanic, despite owning a copy (not entirely my fault). But i may have to watch Love Story, since the wiki plot synopsis doesn't seem to have enlightened me enough. I'm likely missing details of the plot.

But, it's a good article. I'm glad i read it. And, it matches my experience working for four Universities, which includes Harvard, two state schools, and a selective university (i have additional experience at another selective university and a state school, and they match prior experience, more or less). Yes, i admit, i don't really want to spend time reading an article that doesn't support my biases.

I could have applied to go to Harvard. I considered MIT enough to order a course catalog. But they didn't look like a good deal, in fact, they seemed to offer less than the state school, and in ways consistent with this article. It seemed obvious, but in retrospect i spent great gobs of time on it. At the time, MIT was considered a great graduate school.

The "teach you how to think" line was in vogue where i went to school. No one seemed to have any idea how to do it, or measure the effectiveness of any ideas that might come up. That seems to be changing. But in any case, they did, for the most part, make an attempt to teach stuff one might need. And above all, they came up with a way to get the students to demonstrate competence. This last bit is important. Most schools don't do this (I'm not aware of another that does). And yet, it's why i went to school.

I love this: "Perhaps I am emblematic of everything that is wrong with elite American education, but I have no idea how to get my students to build a self or become a soul." Not many would have the courage to say something like that. But the emperor doesn't have any clothes. And it turns out that courage is something that is important for both student and teacher to have. Perhaps it should be taught. In any case, it should be easier than teaching "building a self".

His list of stuff students should learn is very good. Perhaps it should start in middle school, dumbing down nothing, and including the bits under "a liberal education". And the goals should be introduced even earlier. The idea that i'd ever "appreciate that people who disagree with me are not stupid or evil" is going to be hard to swallow. I've seen people who are obviously evil who act like they're stupid. While i try not to go off the deep end on conspiracy theories, Watergate really did happen. And, i think, the goal is in conflict with knowing how to distinguish vetted fact from superstition, etc.

At the end of the article there is a conclusion. Should Harvard adopt merit based admissions? Perhaps they could do it as a pilot project. Perhaps they've already got pilot project statistics, and they could go with the results. Perhaps they could look to see if someone else has already done this experiment, and go with the results. In summary, these ideas are Run an experiment, look at data you already have, and learn from others. As these are skills that Harvard should be teaching their students, perhaps they could simply turn these ideas into student projects, then evaluate the projects and use the results.

Monday, May 19, 2014

Management and The Art of War

In Sun Tzu's The Art Of War, one of the lessons is that a good strategy involves teaching your staff good tactics, then using the good tactics to advantage. So, today's Harvard B school management, with its emphasis on how much money is made each quarter, and stupid slogans totally misses the point. Following this plan dooms your business to slow death. Toyota seems to have noticed this. Toyota replaces robots with humans. this problem is not confined to the automotive industry.

Readers of this blog may have noticed that there is a bias towards education. So, let's say that you're a leader of a major organization, involving multiple technologies. How do you educate your staff? You could simply hire college graduates. But the pool of graduates is getting smaller. College costs too much except for those lucky enough to be born into wealth. The middle class is becoming part of the poor. It's not enough. And besides, college doesn't teach people how to make your business work. They still need to learn that on the job. Who do they learn it from? Well, they need to learn it from people already doing the job. But doing this means that your business will tend to simply learn how the business operates, not how it should operate. You can get some more input from contractors. After all, these people very likely have worked for your competition. And even if they aren’t currently as good as you are, they have something to teach you, even if it is by obvious bad example. Contractors are currently treated as second class staff. This is not the way to build a smoothly working organization. It’s a way to alienate part of your staff. It was my opinion that engineers should spend at least some of their time on the factory floor figuring out how to make things work easier, faster and safer. And by engineers, i mean everyone who could learn something for the organization.

In order to create strategy that incorporates good tactics, management must learn all the best tactics. But even if the CEO is already an engineer, that doesn't mean (s)he knows all kinds of engineering, or computer science, or networking, or communications security, or telephony, or building design, or logistics, or finances, or marketing and so on. How does one do it? Well, IMO, the CEO should have depth in at least something, but likely many such things. But also, the CEO should have a skills team covering all bases that can provide advice. And the CEO has to know the language of all the skill team members so that the advice can be evaluated. I'm not saying that the CEO should micromanage the whole system. I'm saying that the CEO should be able to build and tune strategy that allows the whole company to work efficiently. One of the things that management of larger companies currently does is build fiefdoms or silos that insulate groups from different disciplines making communication and cooperation between groups slow, difficult, unresponsive and otherwise dysfunctional. These structures build in inflexibility, and one hears "that's not the way we've always done things" (a phrase that is nearly always something that is demonstrably a lie).

Thursday, March 06, 2014

Edmunds on fuel economy

Edmunds has done a fairly complete set of tests of fuel economy in a sort of Myth Busters way. They mostly match my own results.

I use cruise control aggressively, if i have it. That is, i use if even if it only marginally makes sense. That's because it gives me 4% to 5% fuel economy at the same speed. And, it generally reduces my workload.

I consistently get 17% better fuel economy for going 62 MPH rather than 70 MPH. Why 70? Michigan's highest speed limit is 70. I try very hard to not speed. Why 62 MPH? My speedometer also shows KPH, so there's a handy reference - 100 KPH. Naturally, Edmunds gets a higher percent improvement. Often overlooked is that driving slower is safer. We tend to increase our risk to some threshold. I try to minimize my risk unless there's some corresponding benefit. Usually there isn't.

Drafting: I've tried drafting manually. I found that drafting a truck fairly got me worse fuel economy than driving at the same speed. And, driving at the same speed with cruise control is better yet. Drafting is risky, and pointless, so i don't do it. This test was done on a cross country trip, covering about 400 miles each. Wind and terrain were similar. My guess is this. Constant speed changes, through brakes or engine drag, kill fuel economy. Engine drag is the same as the brakes.

Roof luggage. I lost between 5% and 10% for putting roof luggage on my car. Cruise control was used at the same speed on the same course for thousands of miles. My luggage carrier at least looks fairly aerodynamically efficient. The way i look at it is that you put roof luggage up there if you need it. I'd love to be able to drag a trailer. It would be better than driving a pickup and paying the cost all the time.

A/C. I've not done this test. I'm not into noise, so i don't drive with the windows open on the highway. I've read articles about just this topic. The answer seems to depend on the details; such has speed, the kind of car, and so on. It's not a huge difference anyway.

Tire pressure. I underinflated the tires on my car. It did not change my fuel economy in any measureable way. I overinflated my tires, and again, no change. But i view it as a safety issue. For at least one of my historic cars over inflation leads to over steering.

Manual transmissions. I've heard otherwise, but my tests show that manual transmissions save you at least 5% in fuel economy over automatics. I'm not a gear head who likes to shift gears. What i'd like is a series hybrid. That is, and electric transmission. With no differential, we'd get an additional 7% to 15%. With no engine drag, we might get another 20%. There shouldn't have to be any gears, so there's smooth acceleration always. It should last longer than hydromechanical automatic transmissions. It should have redundancy that will get you home even if a wheel can't be powered. It should be cheaper to replace if it does fail.

My conclusions match too. The main issue is psychological. I used to have an hour commute to work. Driving faster could save me, at most, about five minutes. But since there were traffic slowdowns, it was impossible to measure five minutes. Psychologically, it's difficult to let everyone pass you. This is the biggest hurdle. But i can use the cruise control much more often at 62 MPH than at even 65 MPH. Everyone else has to pass me. I have the right of way. It's worth something.

Edmunds fuel economy tests

Monday, February 24, 2014


I have a land line for a phone. I often get messages on my answering machine where the automated message machine gets the complete message onto my answering machine. Today i got "Press 2". That's an entire message. I also got "Press star now to repeat this message". These were different voices, so presumably, different senders. Total fail.

For a list of the ways that technology has failed to improve our lives, please press 1.

Wednesday, February 05, 2014

Caffeine: last part with any luck

Last May, my left shoulder started hurting really bad. I recognized the symptoms. I had them in 2001, though in my right shoulder. Back then, i noticed that it didn't hurt quite so much on Sunday, especially Sunday afternoon. What did i know about Sunday? Well, that's when i'd get a headache from caffeine withdrawal if i'd had too much that week.

I never acquired the taste for coffee, but i can drink Mt. Dew in the morning. Mt. Dew is also an acquired taste, especially in the morning. It's not as vile as beer in the morning, though pretty close. I'd get withdrawal headaches on Sunday because i don't drink soda at home, just at work. So, the obvious experiment was to switch to Sprite, which has no caffeine. After two weeks, there were two things that were obvious. First, there was a clear connection. My shoulder pain had backed off considerably. I still had restricted motion. But it wasn't getting worse.

The second observation was unexpected. It turns out that i don't really like Sprite. So, i switched to water, and i ended up consuming about a half-gallon every day. A side effect of this switch was that i lost 37 pounds (17 kg). That's because water has about 700 fewer calories than Sprite or Mt. Dew in the quantities i consume. This lost weight wasn't hard to find. I don't remember, maybe it was under the bed.

In any case, this started in February 2001. Just in case i wasn't convinced, in March 2001, i went on a job interview. They offered me a cup of coffee. Again, i don't drink coffee. But on interviews, i seem to lose about 50 IQ points. It's amazing i ever get hired. It was one of those small Styrofoam cups. I drank it. My shoulder hurt like hell for four days. I still reach for things with my left hand, even though now that's the shoulder that hurts. The pain held pretty steady until the month of October. Then the pain ramped down to zero, and my restricted motion became unrestricted. But fast forward to now. This is month nine. Every day, my left shoulder's pain noticeably decreases and the range of motion increases. I fully expect to be totally cured by month’s end. That is, if i don't fall off the wagon. It really hurts to fall off the wagon.

Sunday, January 19, 2014

Engines Under Ursus, SciFi podcast review

I've been listening to Engines Under Ursus, by Marin Brady. I've just finished the 26 episodes i had downloaded. There are at least two more. But my ancient 32 bit computer died, more or less. That's another story. This more modern machine i'm using is in a sort of zombie mode, booted from thumb drive, while it gets a brain transplant. Apparently, Marin has written a bunch of short stories. He says he wrote Engines, sent it to Tor, and got something unintelligible back. I don't usually write reviews before i've finished them, but this is an exception. If Tor rejected this piece, it's a mistake on their part.

I've heard Scott Sigler's audio for Ancestor, Infection (now called Infected), Earth Core, and The Rookie. I mention this for two reasons. First, i've had ear buds in my ears for about ten years now. And second, to compare experiences. Scott has great stories, uses voices for the parts, has some quirky reading. I couldn't get over how he'd mispronounce words sometimes. Easy words. He wrote the book, and the words are used right, but...

Martin Brady has some sort of accent. Maybe it's Irish. Whatever. He has multiple voices, and many don't have an Irish accent. They're quite consistent. While many actors, including voice actors, over act, Martin has a more subdued underacting style. It's not a monotone. I can only call it deliberate. And he takes the performance an extra step. He adds pod-safe music to each episode. The music is typically at the end of the episode. Frankly, i wasn't much interested in it. That is, until episode 19. I no longer have two hours of driving commute per day. So it's more unusual now to drive while listening. But for episode 19, i was. And i listened to the music because i didn't want to fiddle with the mp3 player. Holy shit. I practically had a nose bleed, it was so good. I don't even get nose bleeds. And there are lines and bits from the song that are in the episode. So now, i need to go back and listen to everything. So, while Scott moved the field forward by using audio books to get the word out (and it seems to be at least part of what's really working for him), Martin has added a dimension to audio story telling. The audio version is going to be, in my opinion, better than the printed version. I've had to go back and listen to the song, uhm, repeatedly. Apparently, Rockbox (open source firmware for my Sansa and other mp3 players) has a feature called "bookmarks". I figured out how to use them, and it has let me conveniently play this piece at will. I really hope Martin has high quality wav files so that his finished product can be released in higher quality mp3s, OGG, or FLACC files eventually. The piece deserves it. I'd buy it.

What about the story? Well, at first, i had no idea what to make of it. I mean for at least ten or fifteen episodes. For some books, at this point i'd stall, and either drop it, or pick it up later, often years later. But this was different. It was something like the movie Brazil (my favorite movie, BTW). Engines seems to be going somewhere. But in Brazil, even after watching the movie several times, and reading the screen play, i still have no idea where it's going, or for that matter, what happened. There's an unpredictability in Engines that i like. When i finished episode 26, i noticed that episodes 24 and 26 were on my player. Rockbox has a feature where you can delete a track before you've finished listening to it, if you're near the end. I must have listened to episode 25 out of order. I think it was in the low 20's. I recall wondering what Martin was up to. Martin wasn't up to anything. Sometimes authors move stuff around. But here, i did it, and it didn't matter as much as you'd expect. The story isn't over, but it's looking like a real winner.

Ursus is a planet other than Earth. There are aliens. Humans are aliens elsewhere. There are androids, and lots of tech stuff. It's not hard science fiction. But the science rules broken wouldn't completely change the story if they weren't, at least not so far. The story is about the story. I'd say more, but at this point, you should have all you need to decide if you want to read it. I wouldn't have written this review if i didn't like it.