Philip Greenspun’s Weblog

We MIT 1982ers derided the Civil Engineering majors as learning about “bridges and sewers.” Harvard thought Civil Engineering was so dull that they supposedly fired their entire (top-notch) department in the early 20th century. The administrators didn’t think that civil engineering was going to be interesting going forward (and debacles such as “Longfellow Bridge repairs will now take about as long as the original construction” may have proved them right).

I attended the renamed MIT Department of Civil and Environmental Engineering “new research breakfast” recently and learned what they’re up to.

MIT has a strong bias toward “engineer and build to solve all of the world’s problems.” This works best if you first gather the world’s smartest people. These are easy to identify because they are MIT students, graduates, or employees. solve.mit.edu (A TED-style event, October 5-8, 2015) offers a succinct summary of this philosophy: “Solve will inspire extraordinary people to work together to solve the world’s greatest challenges.”

This bias was apparent during the department head’s opening slides. The world population was going to grow to 10 billion people pretty soon and we were going to engineer transportation, food, and climate solutions for them all. There was no slide noting that we’ve built a lot of highways in the past 2000 years and yet still are plagued with the traffic jams that bedeviled contemporaries of Octavian (see “The Embattled Driver in Ancient Rome” by Matthews). Nor was there any slide noting that improved food production yields and medicine had been accompanied by a huge increase in human population so that at least some people remain without adequate food and medical care.

Is there any evidence that there are limits to engineering accomplishments? That we might not be able to build our way out of any uncomfortable situation? Right in the room with us it seemed to me that there was ample evidence. The speaker with the Windows 10 laptop was interrupted by a “not connected to network; backup not started” message over the slide (i.e., the programmers at Microsoft have yet to come up with a way of suppressing unimportant alerts when PowerPoint is in “show” mode). The speaker with the Macintosh notebook computer struggled for quite some time even to get the machine connected to the projector. Fifteen years ago I wondered “Given that these classrooms have all been recently remodeled at tremendous expense, why couldn’t there be a permanently mounted camera at the back of the classroom so that live streaming could be offered with a simple switch? Why do we have to call MIT A/V to send a person down with a tripod and camera taking up space at the back of the room if we want to have a video record?” What was the situation today? The room that we were in had been beautifully redone recently, complete with super expensive projector. The folks running the event wanting to stream it out live so they… called MIT A/V to send two people down. An enormous tripod and camera took up valuable space at the back of the room while a second area was set up in the corner to monitor and push the video out to the web.

Professor John Ochsendorf gave an interesting talk about fairly traditional civil engineering. He talked about the value of physical 3D-printed models, some of which indicate that a structure can support itself in situations where computer modeling predicts failure. This technique was used in the structural engineering for Professor Meejin Yoon‘s design for a memorial to Sean Collier, the MIT campus police officer killed by the Tsarnaev brothers. Yoon designed the form of the memorial but there was a question about whether steel pins were necessary to hold together the stone that she had spec’d. The physical 3D model led to a 66 percent reduction in the number of pins. Ochsendorf talked about the achievements of the Romans in building the Pantheon, e.g., using lighter concrete toward the top (see this article in Nautilus).

The opposite side of the field was demonstrated by Professor Lydia Bourouiba, who showed videos of people coughing and sneezing and discussed fluid dynamics models of disgusting human behavior. Why does it matter? Health care environments are set up on the assumption that you need to be about one meter away from a person who is coughing/sneezing. In fact, her research shows that seven meters is the safe distance. She noted that hospitals put most of their effort into protecting workers from disease; preventing patient-to-patient transmission is something for which a good case must be made.

How come we’re always sick? Professor Martin Polz explained that we’ve been running a massive evolutionary experiment with bacteria by pumping farm animals full of antibiotics. “The bacteria compete with the animal for food,” he noted, “so you get faster growth if you use antibiotics in quantities that are many orders of magnitude greater than what are used with humans.” He reminded us that the vast majority of bacteria are harmless, that we ingest billions of bacteria with every glass of water, and that the bacteria that are in and on us outnumber our own cells by a factor of 10. Now that we have antibiotic-resistant bacteria taking antibiotics can be a crazy bad idea. The drugs are broad-spectrum so they kill off a lot of harmless bacteria, thus opening up a big ecological niche within your body. That niche gets filled with the pathogenic antibiotic-resistant bacteria so by taking antibiotics you’ve hugely increased the quantity of nasty bugs inside your body. What can be done about it? Stop using antibiotics on farms, for one thing, but after that treat humans with bacteriophages, specific viruses that target harmful bacteria. The idea of designer phages goes back to the 1930s but people in the West lost interest when antibiotics became available. The Soviets were skeptical about the long-term effectiveness of antibiotics and the shotgun approach so they kept advancing the state of the art, but that work stalled with the collapse of the Soviet Union. Polz’s Lab does stuff that looks a lot like biology.

Take-away: (1) “Bridges and sewers” per se still don’t excite people in academia, but the department has reinvented itself; (2) it never hurts to add biology to any research project!

Full post, including comments

I’m reading The Oregon Trailand the author reminds us that real estate speculator-to-president is not an entirely new path:

George Washington was America’s original maharaja of mules. Historians have long been squeamish about acknowledging that General Washington, like many of the American founders, was a voracious land speculator. Few academics and high school history teachers want to risk their careers by suggesting to their students that the father of their country worked the same day job as Donald Trump. Washington was a land developer, often described as the richest of his generation. By the end of the American Revolution, General Washington controlled about sixty thousand acres of land, more than half of it in the promising frontier country west of the Alleghenies, in what we today call West Virginia, Ohio, and western Pennsylvania. Wresting clear title to this rich bounty of soil from the English crown may not have been a principal motive for fighting the Revolutionary War, but Washington knew that he would profit mightily if independence was achieved.

How does this relate to mules? It seems that Washington needed a way to get around and collect rent:

The traditional draft horses imported from Europe or bred on colonial plantations were magnificent equine specimens, weighing up to a ton apiece, their marbled thighs glistening under the sun as they pulled plows and farm wagons over the flat corn and tobacco fields of eastern Virginia or Pennsylvania. But these agrarian mastodons were enormously hungry at the end of the day, and, like so many “purebred” species, suffered the common defects of animals mated too often within the same bloodlines. The big, beautiful drafts were prone to lameness and chipped hooves, they lacked stamina, and essentially they could perform only one job— yanking a plow or a wagon across level cropland. Heavy draft horses were notoriously ungainly on the kind of steep slopes and rocky ground that would be encountered while conquering the Alleghenies.

Washington and his fellow Virginia planters had long known about the plucky, kick-ass little mules developed for pack trains and for pulling light freight wagons in the Spanish territories of the lower Mississippi and Texas. These “crosses” were bred from horses and small Mexican donkeys, usually producing a mule that stood only four feet at the withers, the part of a horse or mule where the neck joins the body. What the young republic needed now was something much bigger— sturdier, draft-quality mules that stood at five or six feet. In Spain and France, where farming required pulling loads up the steep paths of terraced vineyards and wheat fields, mules of this size had been bred for centuries out of tall donkey sires called “Mammoth Jacks.” Mammoth jacks were any of several long-legged, large-boned studs selectively developed for draftlike qualities, probably from Middle Eastern donkeys brought back from the Crusades. The mammoth jacks had eventually branched off into several discrete European breeds: the Andalusian, Catalonian, Majorcan, and Maltese lines. But the courts of France and Spain, reluctant to share such prize breeding stock with the colonies of their rival Britain, had always banned the export of mammoth jacks to America. After the American Revolution, however, Washington was a global hero, and the Europeans were glad to help the man who had trounced their old British foes. In 1785 the king of Spain, Charles III, dispatched to Mount Vernon a shipment of mammoth breeding stock that included an Andalusian jack named Royal Gift. The shipment included two “jennies,” or female donkeys, suitable for mating with Royal Gift to create more mammoth studs. In the meantime, Washington’s old fighting companion during the Revolution, the Marquis de Lafayette, had shipped from France his own gift, a Maltese jack named Knight of Malta and four jennies.

…

By 1810 the region’s initial breeding stock had yielded an estimated 800,000 mules distributed throughout the South and beyond the Allegheny frontier.

Full post, including comments

The son of Canada’s 1970s leader, Pierre Trudeau, was recently elected to be the new leader. At first glance this would seem to make a mockery of the idea that Canadian society is a meritocracy. How can it be that among the 35 million Canadians the most qualified person to lead the government is the son of the old leader? Won’t this example of nepotism discourage young Canadians from striving their hardest?

On the other hand, Justin Trudeau‘s parents are well-known to Canadians and genetics determines 50 percent of behavior (see The Nurture Assumptionfor where the rest comes from; it will be a rude shock to helicopter parents!). So perhaps for Canadians who thought that Pierre Trudeau did a good job the most rational choice, more or less regardless of qualifications, is any child of Mr. Trudeau.

What do readers (esp. those from Canada) think? Is electing the son of the old leader a good or bad idea?

Related:

• Family Law in Canada (summary: a country that offers the possibility of making $1+ million in tax-free child support after a one-night sexual encounter will now have fully legalized marijuana…)

Full post, including comments

Back in 1993 I established a reputation for insanity by asserting that desktop applications, such as Microsoft Office, were dead. “Everything will run inside a browser and be backed by a web server,” I argued. People around me (EECS department at MIT) pushed back hard against this statement. I had to admit that not everything could run in a browser, even if the browser were augmented with the ability to interpret small computer programs downloaded from the server (browsers would get Java and Javascript capabilities about 1.5 years later). The application that I was most willing to give up on was computer-aided design (CAD). Plainly the performance could never be acceptable without running compiled software directly on the user’s personal computer (desktop Unix workstation, actually, at the time).

With Microsoft Office alive and well and iOS and Android apps being released every few seconds it seems that I was dead wrong about native applications being supplanted. And recently I learned that I was dead wrong about CAD being restricted to running in a native application. Onshape has released a collaborative server-based full-blown mechanical CAD system that runs in a browser. A lead engineer in Shanghai can design a widget and watch in real time as a junior engineer in Indiana tweaks the design. Either of them can spin a 3D model of the design around at any time. As with Google Docs, changes are saved back to the server automatically. The version control system is much more elaborate, however, and it is possible to name versions of a design.

What’s behind the wrongness of my prediction? Three things:

1. anything related to 3D graphics or display in a modern computer or smartphone is done by the graphics processor (GPU), not the CPU
2. companies such as Google have put tremendous resources into making Javascript execute fast within browsers
3. there is a 2011 standard that lets a Javascript program running inside a browser tell the GPU what to do: WebGL

The client side of Onshape is essentially the mother-of-all-Javascript systems telling the graphics card what to do. It turns out to be more than sufficiently responsive. Other than a thin shell of an iOS or Android app, the company offers no software to download and install.

Onshape has an interesting way of pricing their system. It is free if you don’t mind the rest of the world looking over your shoulder. I.e., you don’t have to pay if your design can be looked at by others. They charge for keeping a design private.

I’ll be interested to hear from Mech. E. readers what they think about this. Current distributed work in mechanical CAD seems to be done using standard programs that write big files and then sharing those files with Dropbox or a private equivalent. This has all of the same problems as using Microsoft Office in a distributed collaborative fashion, e.g., files named “20151007-widget-design-FINAL-v2-edits-philg-FINAL”. Looking at the US News list of the best global engineering universities I note that they don’t seem to be clustered. Among the top 20, 9 are in China, Hong Kong, and Singapore, 5 are in Europe, and the remaining 6 are in the U.S. I would think a company could realize a significant competitive advantage by facilitating collaboration across those continents.

Separately, what games are making heavy use of WebGL? This seems like a useful substrate for online multiplayer video games.

[Disclosure: I have a little personal history with CAD. I was co-author of the ICAD system. This enabled the specification of a family of designs such that it was possible to go from specifications to a 3D model automatically. It was the most popular application program on the Symbolics Lisp Machine. Given that it required typing specs in a machine-readable language, my standard retrospective description became “the world’s best mechanical CAD system for people with PhDs in computer science.”]

Full post, including comments

Here’s something Hillary Clinton said in the October 13, 2015 Democratic Presidential debate:

I remember as a young mother, you know, having a baby wake up who was sick and I’m supposed to be in court, because I was practicing law. I know what it’s like. And I think we need to recognize the incredible challenges that so many parents face, particularly working moms. (Clinton)

All of the Democrats agreed that Americans should be compensated for the burden of parenthood, e.g., with “family and parental leave to all of our families.” (Sanders; see also this posting on who should pay for said leave) And given the array of subsidies for parents already in place it seems that there is political consensus around the idea of paying Americans to reproduce.

My question for today is when and why this became necessary. At the time of the nation’s birth parents did not get any federal tax credits for having children. There was no income tax (the Sixteenth Amendment was passed in 1913). Parents also had to educate children themselves or pay a private school (Wikipedia says that it wasn’t until 1870 that “all states had free elementary schools.”). Real incomes were much lower than today. There were no disposable diapers, dishwashers, or washing machines. Yet “the average woman had over seven livebirths in 1800” (NBER). Suppose that five of those children survived through age 18. Is it truly the case that going to work in an air-conditioned office in 2015 and coming home to meet two children who have returned from their government-provided K-12 class and after-school program is actually “an incredible challenge” compared to doing home-based work while dealing with five children circa 1800?

[Separately, the candidates who showed up to the debate were apparently all in agreement that the way to alleviate the “incredible challenge” of dealing with a couple of kids is to pay mothers to stay home for three months following a child’s birth. Does that make sense? What about the remaining 17.75 years of supposedly “incredible challenge”? If I sit on a Carnival cruise ship for three months doing nothing am I then ready to face an “incredible challenge” for 17.75 years? As noted toward the end of my maternity leave posting, it seems as though it would be better to take all of the benefits given to parents (except for K-12 schooling) and load them into the 0-5 period. But even then I am not sure why extra benefits during the first three months will make all the difference.]

What do readers think? Compared to the days in which Americans were having kids without being paid, we have disposable diapers, a river of inexpensive imported clothing and shoes, home appliances, electric lights, take-out food on almost everyone’s way home from work, taxpayer-funded babysitting from age 5-18 (a.k.a. “K-12”), mostly taxpayer-funded babysitting from age 18-22 (a.k.a. “kollege”). Why is it that American parents bellyache about how tough it is and how they need to be subsidized by childless Americans?

Full post, including comments

“FAA to Let Airlines Seek to Delay Navigation Upgrades” is a Wall Street Journal story on how the FAA is going to allow airlines to wait until 2025 to install ADS-B equipment. ADS-B requires a GPS receiver comparable to what is found in consumer electronics devices retailing for less than $100. It also requires a radio to transmit GPS location out to the Air Traffic Control computers. The whole thing can be done for $699 for experimental airplanes (Flying magazine). How much does complying with regulation and FAA certification add to the cost when a transport-category aircraft is involved? This article estimates up to $700,000 to retrofit an older airliner and as little as $130,000 during production of a new airliner.

Full post, including comments

“Pipistrel Alpha Electro: The trainer of the future?” is an AOPA article about the electric airplane that is closest to realistic availability and practical use.

Some interesting facts that I learned about this Slovenian-made product…

The Siemens (German-made?) 80 hp motor isn’t all that cheap, at $10,000, but it costs about as much to overhaul as a light aircraft alternator ($400). The motor will last for 6,000 hours with two overhauls in the middle. With 80 horsepower this machine will have the same power as the original Rotax-powered Diamond Katana, a thoroughly awesome trainer with an engine that U.S. operators found tough to maintain (partly due to poor support from Rotax).

Perhaps the age of hard-of-hearing flight instructors will be coming to an end: “The aircraft is quiet. There is some noise, but it’s mostly from the propeller, and headset-free conversation is no problem, even at takeoff power.”

No innovation can survive an encounter with U.S. bureaucracy: “One big hitch in the Alpha Electro’s future is the United States’ LSA rules, which make no provision for electrically powered aircraft. So for the time being the airplane lives in a sort of regulatory limbo. It can be flown in the Experimental category, but this permits no commercial activity—including flight training, the airplane’s principal mission.”

The batteries can be recharged in just 45 minutes, meaning that a flight school scheduling students into 2-hour blocks might just barely be able to substitute this plane for current trainers.

Who else is excited by this development?

Full post, including comments

From Facebook:

• User 1: “Not sure a man could have pulled off the Theranos scam as he would have been asked questions.”
• User 2: “Nobody wants to ask ‘How is a young uneducated American woman more capable than 100 chemistry PhDs at Siemens in Germany?'”

Full post, including comments

Back in August I wrote a posting about the New York Times portrayal of Amazon as a sweatshop. Most of my friends who have worked at Amazon were there in the mid-1990s and have departed for various reasons (“called in rich” being one; didn’t agree with Jeff Bezos being another; wanted to help rid the world of Republicans being a third). Thus I don’t have personal knowledge that would support or contradict the Times story. As a database programmer, though, I find Amazon’s rebuttal interesting. Some of the response rests on information queried from database management systems:

Elizabeth Willet, who claims she was “strafed” through the Anytime Feedback tool, received only three pieces of feedback through that tool during her entire time at Amazon. All three included positive feedback on strengths as well as thoughts on areas of improvement. Far from a “strafing,” even the areas for improvement written by her colleagues contained language like: “It has been a pleasure working with Elizabeth.”

Chris Brucia, who recalls how he was berated in his performance review before being promoted, also was given a written review. Had the Times asked about this, we would have shared what it said. “Overall,” the document reads, “you did an outstanding job this past performance year.” Mr. Brucia was given exceptionally high ratings and then promoted to a senior position.

I’m wondering if this is the future. Companies who are attacked by journalists will be able to use DBMS queries to argue that information was presented selectively and/or incorrectly.

Full post, including comments

A new emergency procedure, in Year 70 of widespread helicopter flying… a technique for getting out of vortex ring state with a powered bank instead of by reducing power and lowering the noise: “Flying Through the Vortex” (Rotor & Wing, September 1, 2015) by Tim Tucker, reporting on a technique developed by Claude Vuichard, a Swiss pilot.

Full post, including comments