Monday, September 29, 2014
This month's Scientific American devotes twenty pages to articles on diversity in science—the shameful lack thereof, and what can be done about it. One piece is a kind of confessional by a Lockheed Martin engineer who quickly moved into management and found that as she chose staff from widely varying backgrounds, the quality of her group's work increased. Other articles cite social-science studies that show diverse organizations are not only more socially just; they do better science and engineering too. The reader of these paeans to diversity could not be blamed for taking away the impression that diversity is like goodness: you simply can't have too much of it. Is that true? Or are there limits to diversity?
What does diversity mean? It has both an objective aspect, and a subjective or political aspect. In the strict sense of diversity meaning merely "difference," one can objectively measure diversity in genetic makeup, diversity in hair color, or diversity in virtually any other measurable characteristic that a group of things or people has. This scientific aspect allows statisticians to crank out reams of charts showing the degree of gender diversity in the number of Ph. D.'s granted, ethnic diversity in hiring practices, and so on. So in this scientific sense, diversity is a quantifiable, measurable thing.
But when we ask what kinds of diversity are significant in the sense addressed by the Scientific American authors, the list narrows to political and cultural hot-button matters: gender, race, socioeconomic status, ethnicity, sexual orientation, and so on. Scientists and engineers must deal with these matters not as mythical objective professionals, but as human beings. And in so doing, the issue becomes an ethical, political, and even philosophical one.
The idea of virtue is not a scientific concept, but it is one of the best ways to describe a certain class of characteristics involving choice, as Aristotle says. I think Aristotle would class diversity as a type of virtue because a diverse organization is better with regard to social justice than a non-diverse one, and (as the social scientists have shown), diverse scientific organizations do science and engineering better than non-diverse ones. Making something intrinsically good and also better at what it does are the two main aspects of a virtue (again, according to Aristotle), and diversity qualifies on both counts.
The next question is this: can you have too much diversity? Most virtues represent a mean or rough average between the two extremes of excess and deficiency. Assembling a competent technical organization with a mind to diversity represents a compromise between extremes. In the decades before diversity in its modern sense was recognized as an organizational goal, those in charge (usually white males) picked the best people they could while following the cultural norms of their time. These norms generally (but not always) excluded women and minorities, and tended to perpetuate the demographic makeup of the organization, while making it extremely hard or impossible for non-whites and non-males to enter. This was bad.
However, you can imagine an opposite extreme. The perfectly diverse organization would have diversity statistics identical to those of the largest applicable sample group: the state, the nation, or even the world. William F. Buckley is supposed to have said he'd rather be governed by the first hundred names in the Boston phone book than by the entire faculty of Harvard University, and in this proposal, at least, he was favoring the opposite extreme of diversity I am talking about. But if diversity is the only criterion of selection, the specialized competencies that a research or engineering organization needs will be absent, except by chance, and it will fail to achieve its objective, unless its only objective is to show that it is acceptably diverse.
The U. S. National Science Foundation has in recent years spent a substantial portion of its resources encouraging diversity in various ways. To the extent that these efforts have righted previous injustices committed either consciously or through unconscious bias against certain groups, they are to be applauded. But there is nothing scientific about the choices of which measures of diversity to work on.
In a secular democracy, these choices are made politically. And making politics your ultimate authority can land you in unpleasant places, as scientists in Russia and Germany have found. A crackpot biologist named Lysenko got his hands on the political levers of control in the old USSR in the 1920s. Lysenko thought acquired characteristics could be inherited, and for the next forty years, any Soviet biologist who disagreed with Lysenko about evolution was liable to disappear into the Siberian work camps. And the Nazi party in Germany took delight in calling Einstein's theory of relativity "Jewish physics." Such blatant overruling of science by politics can always happen if those in charge value political goals more than the integrity of science.
I am personally about as un-diverse as you can get: an old white male conservative Christian Texan. An organization composed of people like me would score close to zero on any diversity index you care to name. I view the diversity project as an attempt, however flawed, to show the type of love that wills the good of the beloved to people who would otherwise be kept from flourishing to the best of their abilities. There is nothing wrong with this type of love. It is the type of love Jesus Christ exhorted his followers to show to each other. But implementing diversity in a way that helps those who need that type of help without inflicting harm or the loss of opportunity on others is an inherently complex task. The only way to do it perfectly would be to have perfect insight into the problem of social justice, and only God has that. Any human attempt at diversity represents a compromise between using resources to increase diversity, versus using resources to address the task at hand. And those promoting diversity should remember that it is possible to have too much of a good thing.
Monday, September 22, 2014
In an article in this week's New Yorker magazine, Nick Paumgarten contemplates the wider effects of GoPro, the sports-oriented wearable cameras that have inspired viral videos of amazing stunts watched by millions on the Internet, and things like the GoPro Mountain Games, a venue where mountain bikers, rock climbers, and even ten-year-old girls on ziplines record every second of their exploits for fun, and sometimes profit. (It turns out that GoPro sponsors certain athletes with things like free cameras, or a monetary reward for getting a million hits on a GoPro-made YouTube video.) While admitting the obvious entertainment value of the small portable video cameras both for the users and the viewers, Paumgarten looks at the potential downsides of this new technology, and provides some useful food for thought.
First, he worries that people will increasingly fall victim to what I'd call "camera-itis", which I have suffered from on numerous vacations: the temptation to transform a live novel experience, whether of a ski slope, a wedding, or the Leaning Tower of Pisa, into just another shot to be captured. This is in contrast to the traditional reason to see remarkable sights, which is to let them soak into you and transform you over time. For example, the way Henry Adams visited Chartres Cathedral in France.
One of the most profound appreciations of the artistic merits of Chartres Cathedral was penned by the American historian Henry Adams. A reader of his book about that medieval monument to faith can sense the hours of study, contemplation, and reflection that Adams put into his work, both in time spent at the cathedral and in historical research. Perhaps Adams used photographs to jog his memory, but his musings on the cathedral are about as far as you can get from the exploits Paumgarten describes toward the end of his article: BASE jumping from famous buildings while wearing a GoPro. Yes, the One World Trade building does show up in a shot filmed by a clandestine leaper from that building, who managed to survive. But the point of that video wasn't artistic appreciation.
For readers like me who have never heard of BASE jumping before now, it stands for "building, antenna, span [of a bridge, presumably], earth [meaning a cliff]"—four types of places from which a jumper whose courage sometimes exceeds his judgment leaps in the fond hope that his parachute will deploy in time. So far, 242 haven't—that is the total number of deaths recorded in an online BASE fatality list since the sport began around 1981. A GoPro or other means of documenting your exploit is a necessary part of this fringe semi-suicidal cult sport. While it would be unfair to shut down an entire camera industry on account of its abuse by a few kooks, it must be admitted that the availability of cameras like GoPro have encouraged this sort of daredevil activity.
But one of the most serious and wide-ranging issues Paumgarten identifies in connection with GoPro-type cameras springs from a matter I blogged about a few weeks ago in connection with the shooting of Michael Brown in Ferguson, Missouri. Paumgarten writes "a world in which the police film every interaction is not all sweetness and light. You may catch some bad cops, but you'd also hamstring the good ones. . . . It deprives the police of discretion, and the public of leniency. There are many things we'd rather not see or have seen."
Paumgarten's chief concern here seems to be that once a cop films an infraction, he or she will have no choice but to proceed with an arrest. He says a video record "has the effect of a mandatory sentence" and enforces "uninterpretable standards of exchange."
I have spent some time trying to figure out what he means by uninterpretable standards of exchange, and haven't made much progress. Every image has to be interpreted somehow. Objectively speaking, images are simply arrangements of color on a screen, and require perception and interpretation by a human mind to convey meaning at all. The question is not whether a video can be interpreted—not only can it be, it must be if it is to convey any information—but who does the interpreting, and what principles the interpreter follows in translating the raw images into conclusions about pertinent matters of fact.
Paumgarten may be thinking in legal terms that filming police encounters effectively brings the whole legal system—judge, jury, prosecuting and defense attorneys, you name it—onto the street corner along with the cop and the public. And of course, such a situation would markedly change the interaction between law enforcement and citizenry. But unless we become an actual police state, with every action, word, and gesture not just potentially, but actually scrutinized by a hostile Big Brother, video recording of police work need not change the routine activity of cops who deal with the public. For every arrest, there are many lesser interactions of the "break it up, folks, there's nothing to see here" variety. And unless some lawyers find a way to exploit the presence of recordings of this sort of minor interchange by charging police brutality where none exists, there is no reason to think that cops with good judgment will be any less able to deal with the public in these minor ways than they are now. But never underestimate the ability of lawyers to squeeze profit out of a situation.
Beyond law-enforcement concerns lies the greater question of how life will change as recording cameras become more nearly ubiquitous. There have always been foolhardy persons willing to risk life and limb for the chance to do something that will get their name in the paper—even if it's the obituaries. We may have a few more of these folks now that GoPro has come along, but they probably would have gone ahead and done something foolish anyway without a camera. In the hands of private citizens, GoPro cameras seem to be mostly a benign influence, encouraging the sharing of remarkable experiences by those who do not have the descriptive verbal abilities of a Henry Adams. And while wearable cameras hold out the promise of better evidence in police work, we need to adopt rules that preserve the ability of the cop on the beat to use his or her discretion in enforcing the law. The introduction of radar speed-detection devices did not eliminate the warning ticket. And the use of wearable cameras need not transform a police department into an array of RoboCops that automatically mete out punishments for all infractions, however minor or technical.
Monday, September 15, 2014
By now, almost everybody with a cellphone and a car knows that it's a bad idea to text while you're driving. But people still do it, and some of those people die in text-related car crashes and take innocent victims with them. What if technology existed that simply prevented people from texting from a moving car at all? Wouldn't that solve the problem?
Scott Tibbetts thought so. Tibbets and his company Katasi were profiled in a recent New York Times article for developing a promising technology that would simply block texting from any phone that was in a moving car. While there are several technological solutions to this problem that are already on the market, they all have various problems.
Some text-blocking apps work by using the phone's GPS to figure out if the phone is moving faster than walking speed. If it is, the software concludes that you're driving, and blocks texts. This one turns out to be a battery hog, because the GPS system has to run all the time. It also might present problems for train and bus passengers. Another system uses the car's speed sensor and links it to the phone with a Bluetooth wireless connection. But it costs over a hundred bucks, and there aren't that many people who are both concerned enough about texting while driving to buy it, and also willing to shell out that much money for something they could do for free with a little more willpower, perhaps.
Mr. Tibbetts' solution is cleverer than these. It involves connecting a wireless box to the car's OBD-II port—the on-board diagnostics socket that the auto technicians use to figure out what the "service engine" light means. When the car's moving fast enough to be dangerous, the wireless box sends that information to the cellphone network, which then asks the phone—once—where it is. Then, if the network is using the software developed by Mr. Tibbetts' firm Katasi, the software uses the location data to figure out things like who is driving the car. You don't want a whole family's text service blocked just because Mom is driving to the grocery store, for instance. That way, the GPS battery-drain problem is minimized, and the computational heavy lifting is done in the cloud, so to speak, rather than by the phone.
Mr. Tibbetts, an aerospace engineer and entrepreneur, has persuaded both an insurance company and a cellphone provider (Sprint) to cooperate in test trials, which have worked fine. But it appears that the largest player, Sprint, has gotten cold feet lately, and has stalled further tests. In the Times interview, Wayne Ward, vice-president for business and product development at Sprint, expressed concerns about product liability. Currently, if a driver texts while driving and gets in a wreck, it's the driver's fault. Mr. Ward asks what might happen if Sprint sells the Katasi system that claims to prevent such accidents, and then some glitch happens and somebody sneaks through a text and crashes anyway? Why, Sprint could be sued!
Pardon me, but it appears that there's more going on here than meets the eye. Any time a small independent company comes up to a big firm and offers the big guy new technology, the not-invented-here problem can raise its ugly head. Short of buying the small upstart outright (which happens a lot, by the way), if the big firm adapts the small company's technology, they will be on the hook for royalty payments or other forms of obligation that big companies don't want to be tied down to. And there's also the simple pride factor expressed by the phrase "not invented here"—if we didn't think of it first, it can't be that good.
Besides, it's not clear who would make enough money to offset the expenses of the added hardware and software—and lawyers' fees, if Mr. Ward's fears turned out to be correct. The existing GPS-based solutions for text blocking in cars aren't exactly selling like hotcakes, even after all but five states have adopted no-texting-while-driving laws of one form or another.
One could imagine a legal solution: make something like the Katasi text-blocking system mandatory by government fiat. Nobody has seriously put forward that idea yet. But it might happen. There was a time when ordinary window glass was used in automobiles, with the result that otherwise minor wrecks turned deadly when razor-sharp knives of glass flew around and sliced—well, enough said. But when the technology of laminating glass with a plastic inner layer was developed around 1920 to keep the shattered pieces together, auto companies adopted it, partly motivated by fear of lawsuits. Eventually, most countries made it a legal requirement for all glass in automobiles to be laminated or safety glass, but it looks like the firms were ahead of the government in that case.
Safety glass is a different kind of thing than automatic text-blocking. An auto company could start using safety glass and just raise the car's price incrementally, and hardly any customers would notice the change. But as soon as you stop a person from doing something that they're used to doing, like texting while driving, you create a sharp negative impression. And that's something that cellphone providers are reluctant to do as long as there are competitors ready to take business away.
My hat is off to Mr. Tibbetts, who put five years and millions of dollars into developing a clever technological fix for a significant problem. But as many engineers turned entrepreneurs have learned, building the better mousetrap—or text trap—is only part of the problem. Convincing people to buy it and use it is often harder than coming up with the invention itself. If everybody used something like the Katasi system on their cellphones, we would all be safer, no question about that. We would also lose a little freedom of judgment which we can now exercise, which is whether to text while driving. Perhaps some telecomm industry leaders will get together and agree to adopt Katasi, or something like it, but such inter-company cooperation for a non-financial thing like safety is a rarity. It could happen, though. I bet Mr. Tibbetts, for one, hopes that it will.
Sources: The New York Times article "Trying to Hit the Brake on Texting While Driving" by Matt Richtel, appeared in the online edition on Sept. 13, 2014 at http://www.nytimes.com/2014/09/14/business/trying-to-hit-the-brake-on-texting-while-driving.html. I also referred to Wikipedia articles on on-board diagnostics, windshields, and safety glass.
Monday, September 08, 2014
Plutonium is nasty stuff. It's highly radioactive, so breathing plutonium dust is not a good way to live to a ripe old age. And did I mention it's an essential ingredient in most thermonuclear weapons? For these and many other reasons, nuclear waste contaminated with plutonium is not something you just toss in the ordinary trash can. That is why, at great trouble and expense, the U. S. Department of Energy built the Waste Isolation Pilot Plant (WIPP) about fifteen years ago, a few miles outside Carlsbad, New Mexico. It is the nation's only federally operated "permanent" disposal facility for nuclear waste. I put permanent in quotes, because, well, something that happened last Valentine's Day showed that so far, putting stuff there is anything but permanent storage.
WIPP is in a salt mine, but salt happens to be a byproduct. The reason WIPP was constructed in the middle of a large salt deposit is that over geological time scales, salt acts more like Silly Putty than rock—it bends and flows instead of breaking, and seals any cracks that might develop. So the scientists and engineers who designed WIPP chose to site it thousands of feet underground in a salt deposit so that even after 10,000 years, underground water would be unlikely to penetrate to the still-radioactive byproducts of nuclear-weapons manufacturing, which comes from a number of national labs dating all the way back to World War II. And for most of the facility's history, things went more or less according to plan. After they dug tunnels in the salt and opened up an area the size of several soccer fields, they began filling the space with 55-gallon drums full of nuclear waste from places like Los Alamos Nuclear Laboratory and elsewhere.
Then, on the night of Feb. 14, 2014, when no one was actually underground but some monitoring personnel were standing their watches on the surface, a radiation alarm went off alerting technicians to high levels of radioactivity underground. The expert who knew what to do about such an alarm was not on duty. They tried to contact this person, without success. This went on till early on the morning of Feb. 15, when some workers began to suspect that the radiation released underground might be coming up through the ventilation system to the surface. After trying to change some ventilation filters, managers finally ordered the WIPP personnel to go to a safe location, but by that time they had been exposed to low levels of radiation, as a later investigation showed.
What happened? According to a recent report in the Los Angeles Times, one of the drums stored underground spontaneously ruptured, spewing out several cubic feet of white foam laced with plutonium. Some of the foam or vapor from it got into the ventilation system that exchanges air between the underground rooms and the surface. This system had radiation detectors, and in the event of a release of radioactive material, it was supposed to divert the ventilation air to filters that would catch all radioactive particles. But the dampers assigned to do this leaked, and lots of contaminated air got to the surface anyway. Over six months later, WIPP is still in a partial-shutdown mode, and estimates of what it will take to restore it to safe operation range up to $100 million or more.
Opinions on the propriety of nuclear technology range all the way from "no way, José" to "nuclear energy is our best weapon against global warming" and everywhere in between. Dead-set opponents of nuclear energy will see in the WIPP accident evidence that plans to keep nuclear waste safe for thousands of years in an underground facility have now been revealed to be a sorry joke. The disabling of WIPP for receiving nuclear waste has not only put the whole idea of underground disposal into doubt, but has also caused a chain reaction (so to speak) of delays in cleanups of nuclear labs around the U. S.
For those who still believe nuclear energy is a good long-term option for our future energy needs, the WIPP accident shows how even the best-laid plans can be upset by a failure of management integrity. Even now, no one knows exactly what happened chemically inside that drum to cause it to rupture. Investigations have revealed lapses in the procedures used to transfer information about each drum's contents to WIPP operators. In other words, WIPP managers are not sure what went into that drum in the first place, so they don't have a basis for duplicating it and maybe finding out how to prevent other similar ruptures. Finding one rattlesnake just hatching out of an egg strongly motivates you to wonder where the other eggs are, and the WIPP people may be sitting on dozens of plutonium rattlesnake eggs. And you thought you had problems.
All this talk about 10,000-year lifetimes makes me wonder what will be left of our own civilization even a thousand years from now. Egypt has its pyramids, Greece has its temples, and maybe all we'll have is WIPP?
A few days ago, a relative of mine sent me a video of the opening of a time capsule that was buried only fifty years ago, in 1964, at the founding of a bank in Fort Worth, Texas, where my father used to work. Whoever designed that time capsule did a good job: along with the perishable newsprint and film reels, they packed a sock full of desiccator and sealed the whole thing with an air-tight lead seal. As a result, the stuff inside looked like it had just been put on the shelf yesterday.
A 50-year time capsule is a far stretch from a 10,000-year nuclear waste repository. But when we are talking about stuff that could kill anyone it touches, the highest standards of engineering and safety must be followed, from the minute that hazardous waste reaches WIPP to the end of the 10,000-year warranty period. There will be pleas for more money for WIPP as a result of this accident, but money isn't the only answer.
Money can't buy integrity, and money alone can't bring into being a cadre of dedicated individuals to whom their duty with regard to safety is their highest calling. About the only place in government you can find this attitude consistently these days is in the military. I'm not saying we should call in the Marines to take over WIPP. But if they did, I bet you wouldn't have any more twelve-hour delays between the time an alarm went off and the time appropriate actions were taken.
Sources: One of the first reports of the WIPP accident was carried by National Public Radio at http://www.npr.org/blogs/thetwo-way/2014/02/28/283773449/13-workers-exposed-to-radiation-at-n-m-nuclear-waste-dump. The most detailed news report I have found was from the Los Angeles Times, which published it online on Aug. 23 at http://www.latimes.com/nation/la-na-nuclear-waste-accident-20140824-story.html. I also referred to the Wikipedia articles on "Waste Isolation Pilot Plant" and "Plutonium." For those interested, the opening of the 50-year time capsule at the former City National Bank is described in the Fort Worth Star-Telegram at http://www.star-telegram.com/2014/08/27/6072685/banks-50-year-old-time-capsule.html.
Monday, September 01, 2014
Inventions and engineers go together like songs and musicians, and consequently every engineer should know something about patents. The practice of patenting inventions to give the inventor a temporary monopoly as a reward for ingenuity is something that most countries do today, and by and large, we are probably better off with patent systems than without them. But patent systems can be abused, and some of the biggest abusers of the patent system in the U. S. are "nonpracticing entities" (NPEs), otherwise known as patent trolls, as pointed out in William J. Watkins' recent book Patent Trolls: Predatory Litigation and the Smothering of Innovation.
If you're a patent troll, here is what you do. First, you acquire a lot of broadly written patents on some technology that either has a lot of wealthy corporations active in it already, or is likely to in the future. Software is a favorite, as are wireless technologies and even business practices involving the Internet. You either file the patents yourself or, more likely, buy them up at a dime a dozen at a bankruptcy sale of some startup. Then you wait till a well-heeled firm like Apple or IBM is doing something vaguely related to your patents, and you send them a letter threatening to sue for patent infringement. If they call your bluff, you file suit in the U. S. District Court of the Eastern District of Texas, headquartered in the town of Marshall near the Texas-Louisiana border. You carefully select a jury of twelve rural citizens who probably wouldn't know an ethernet cable from a rattlesnake, and your lawyer (who is a member of your small NPE firm) exhorts them about the evils of out-of-state big-business interests who try to crush the innovative spirit of the small, struggling, independent businessman/inventor, namely you. More likely, the defendant corporation you threaten to sue simply settles out of court for millions of dollars rather than face an even more expensive jury trial, and you walk away richer and better-funded to do it all over again next time.
If this unsavory process smacks of blackmail, you are on the right track. Unfortunately, a number of legislative and bureaucratic circumstances have created a large loophole through which millions of dollars each year is drained from productive firms into the pockets of patent trolls. The U. S. patent law was written back at a time when most inventions had a lifetime of at least several years, so the term of fourteen years (later extended to twenty) was a reasonable one. But now that product lifetimes, especially in the software field, are measured in months rather than years, twenty years is plenty of time for a patent to turn into a kind of zombie, rising from the dead when revived by a patent troll to haunt a legitimate firm with dubious claims of infringement. The U. S. Patent and Trademark Office is severely understaffed, with each of its few examiners having to work through hundreds of filings a year. Rather than let a backlog pile up, they have defaulted to issuing patents freely, including many which really shouldn't be allowed. And finally, judges and juries in the Eastern District have proved so friendly to the patent trolls and hostile to the defendants that the American Tort Reform Foundation has put it on their watch list of "Judicial Hellholes."
The well-heeled giant tech firms are not the only ones harmed by patent trolls. Watkins cites an example of a small software startup in the late 1990s headed by one Brandon Shalton, who teamed with two nuns to develop a way that pastors could record their messages and play them back through a website. After testing the product with several churches, Shalton prepared to market his product, only to be confronted by a challenge from Acacia Research Corporation, which held a patent that they claimed would be infringed by Shalton's product. Shalton knew the claim was false—the general idea of recording audio on a website had been around for years. But lacking the resources to fight Acacia's threat, Shalton threw in the towel, and his company died before it was even born, a victim of a patent troll.
What can be done about patent trolls? One idea being considered for legislative action is to impose some kind of practice requirement on a plaintiff in a patent infringement lawsuit. Unless the plaintiff suing for infringement can show evidence of actually intending to use the patent to make something—manufacturing facilities, investment plans, etc.—the plaintiff would not be allowed to sue. Under current law, simply owning a validly issued patent, no matter how flimsy, entitles you to file a suit for patent infringement, even if you never intend to use the patent yourself to make or sell anything.
In making such a fix, lawmakers would have to be careful to protect the rights of the truly independent inventor, who has a good idea and patents it but lacks the resources to exploit it fully. This is the mask that most patent trolls don when they go before juries, but the reality is different in most cases. Watkins thinks there are ways to make sure that small independent inventors can still obtain valid patents and protect them, while putting the true patent trolls out of business.
Patent law represents a balance between protecting the rights of the inventor and allowing the benefits of the invention to spread beyond the inventor's initial monopoly. No patent system will ever make everybody happy. But Watkins makes a good case that patent trolls are currently stifling innovation and exploiting a loophole in the present system that ought to be closed.
Sources: Patent Trolls: Predatory Litigation and the Smothering of Innovation by William J. Watkins Jr. was published by the Independent Institute of Oakland, California in 2013. A complex case of a true inventor who eventually veered toward patent-trolling but left a legacy encouraging independent invention was Jerome Lemelson, who I mentioned in my blog of May 30, 2007, "Engineering Altruism: Two Paths."