Error: I'm afraid this is the first I've heard of a "comments" flavoured Blosxom. Try dropping the "/+comments" bit from the end of the URL.

Wed, 30 Dec 2009

An Engineer's Holiday Wish List

While the country at large is engaged in polarized debates over health insurance reform, I have been thinking about other problems -- problems I know a bit more about -- and wishing for their solutions to appear under our collective Holiday tree. In this essay, I'll describe a few of the solutions I want Santa to put in his bag.

I'm fully aware that the things us engineers cherish can be turn out to be both a solution to an old difficulty as well as a totally new unintended problem. When we envision for the perfect engineering solution to some acute inconvenience -- like when we believe that the new Red Ryder BB gun on our wish list would neatly solve the problem of backyard squirrels near our bird feeder -- we tend to discount dire warnings: "You'll shoot your eye out!"

As an example, mankind's 19th century wish for compact sources of energy and the development of the fossil fuel powered engine ushered in the first industrial age, bringing a better life to millions of people. As factories sprung up, and the earth was dug down in an orgy of cheap energy production, we largely discounted the fact that by-products from this activity would pollute our land, water, and air. In time, the contamination caused by continued use of fossil fuels may destroy the very conveniences it created. We have very nearly "shot our eye out" with that "solution".

Of course, it's often difficult to appreciate all the significance of unintended negative consequences of even something familiar. It's therefore doubly difficult to predict consequences when a technology is first introduced. To our credit as a species, we are better at appreciating the complete, extended consequences of our actions than any other species. We still do miss a lot of things, but fortunately we are also pretty good and devising clever ways out of jams.

In any case, for better or worse, here is my list.

Fusion Power

When I was a kid in the late 1960's I regularly watched the CBS show 21st Century. The great Walter Cronkite was the narrator, and in an inimitable voice that allowed not an particle of doubt to taint his pronouncements, Cronkite assured us all that scientists and engineers were working on harnessing a form of nuclear power that fueled the stars: nuclear fusion. When fusion energy was finally harnessed -- it would only take a few decades to work out the details -- a virtually limitless and totally non-polluting source of energy would be at mankind's beck and call. Dirty, finite energy sources, like coal and oil would be abandoned. There would be no need for problematic nuclear fission reactors either, with their nasty radioactive waste disposal problems.

What happened? The last I looked we were burning up our finite supplies of coal and oil at an even faster rate than we were in 1967. Fission plant construction, at least in the USA, has been indefinitely stalled by irrational fears of atomic energy tracing back to Jane Fonda and the "China Syndrome" panic. As somewhat of a consequence, work on fusion power is practically nowhere to be seen. Funding for fusion research is minuscule in comparison with other Washington budget line items and earmarks.

In 2010, the US budget for Fusion Energy Sciences (FES) support is supposed to be about $400 million. That seems like a very small number to me, considering the trillions we spend facilitating the availability of oil, coal, and gas.

I've heard politicians argue that billions in bailout money needed to go to gasoline powered car companies so that they could develop the next generation of "plug in" electric hybrids. Although I'm a big fan of electric cars (not of bailouts), I was still rendered incredulous by the assumptions implied. For one, I couldn't keep from wondering where the energy for those next generation plug in cars would come from.

Yes, there is some significant work on fusion power, just not in the USA. ITER (originally the International Thermonuclear Experimental Reactor) is an internationally funded tokamak (magnetic confinement vessel) that should go on line in 2018. ITER is a big deal. The vessel is designed to produce approximately 500 MW of fusion power sustained for up to 1,000 seconds. Assuming it works as planned, the yield will be far past break-even, releasing 5-10 times more energy than what will be needed to heat the plasma.

I know that ITER may fail. It'll take decades to complete and is unlikely to succeed as planned, being big and ugly and unprecedented it its potential for messy disasters. As a research reactor, the resulting energy yield will not be used to create electricity. They'll just dump the heat. And this thermonuclear earth warmer will cost around $10B to build and maintain. Not cheap for a toy.

But it's a step in the right direction if you ask me.

Let me put it this way. If we are willing to send troops to Afghanistan at the cost of $100B in an operation that'll take decades to complete and is unlikely to succeed as planned, being big and ugly and unprecedented it its potential for messy disasters, why not spend 10% of that same money to fully pay for something like ITER? Success in Afghanistan could bring peace and stability to critical energy producing regions. Success with ITER could do the same, and then some.

ITER is being funded mostly by Japan and the EU. Last year, the US contributed only a measly $11M to ITER. I think John McCain told me that we spent roughly the same amount studying pig odor in Iowa.

This year, maybe the US will pony up about $140M -- that'll be subtracted directly from the total sum available domestically.

There's no question that the design of a practical fusion reactor is a thorny engineering problem, but it is merely an engineering problem none the less. Look at the sun. Fusion exists. It seems to me that with political will and adequate funding drawing the brightest minds to the problem, progress can be made and success would be inevitable. Walter Cronkite's promise would be fulfilled!

The payoff for success is immense; it's an ultimate game changer. The burning of coal could stop. Fossil oil could be replaced by synthetic liquid fuels. Peace would reign in the Middle East. What could possibly go wrong?

Fusion, the nuclear reaction that lights the stars, is about as sustainable and clean a form of energy production as anything in the universe. I fear that the flea power from windmills and photovoltaics will never scale, sustainably, to power the endlessly growing energy needs of any industrial nation. Only fusion can do that. Funding for fusion research needs to increase.

I want a "star in a box" under my Holiday tree.

Uncrashable Cars

The death toll from the 9/11 attacks is often given to be about three thousand souls, but I don't think that's the complete total. Since the attacks, more people are now driving instead of taking a plane. Because cars are much less safe than planes, it has been estimated, since 2001, that there have been a several thousand more auto fatalities, directly a consequence of this reaction to the attacks. These auto fatalities are the hidden victims of 9/11.

This extra death toll is a sad irony, of course. Part of the reason people are driving more is because driving gives them a feeling of control, and that control "feels safer" to them, even if it is significantly less safe, in fact. Worse yet, some people drive to avoid the inconvenience of intrusive airport security. Extra security designed to keep us safe actually kills people when its inconvenience pushes travelers to use their unsafe cars. The recent "underwear bomber" did not crash his targeted plane, but he's certainly killed some people on the highway.

Cars kill people. There's no other way to look at it. About 40,000 people die in car crashes in the USA every year. Nearly a million people die in car accidents worldwide. Yet, where is the outcry? Why aren't pundits (other than me) yammering about how we are "less safe" because of some detail in auto safety standards? Why don't people demand the president "do something" about auto safety? Why aren't congressional committees convened to investigate failures auto safety? And most of all, why do billions of bailout dollars go to these merchants of death -- the auto makers -- without a single requirement to make their product safer?

Sometimes I feel that I'm the only person that sees this as a critical problem needing to be fixed.

I hear people who'll argue that the safety risk in driving is acceptable to society given the reward of benefits that accrue from the automobile. It's a choice; no one is forced to drive. In fact, driving is the quintessential expression of freedom in the USA.

Fine: driving is freedom. I don't want to take anybody's freedom away. But can't we see auto fatalities as a flaw in the design of cars? It's an engineering problem to solve. What we need are better cars that make us just as free, but don't kill people -- or at least don't kill quite as many. I think this can be done.

Drinking and driving is also linked in our minds with freedom and control. Although we endure invasive searches every single time we board an airplane, people would balk if sobriety checks were set up at the exit from every bar. Arguably, such checks would cut highway fatalities in half, but because autos imply personal control not present in airliners, we see drunk driving as a personal responsibility issue. To use Lawrence Lessig's terminology, we use right coast code, the criminal justice system, to "fix" the DWI bug, rather than left coast code, fixing the technology itself.

Me, I think a car that can be crashed by a drunk driver is a flawed design. We need to come up with cars that drunks can't crash, or maybe alchohol that's fun to drink but doesn't impair your driving ability.

One of the stumbling blocks preventing the development of safe cars, in my opinion, is the strict conservatism surrounding the current design of the automobile. Why do cars have, say, a steering wheel? The way I see it, they have design features like a steering wheel because they've always had them. Nobody sees a compelling reason to change the guidance user interface design to, say, a tiller. Back when cars were first developed, steering wheels were a mechanically simple user interface to the mechanism used for steering. Now with the many improvements in auto technology, including steer-by-wire, the user interface to the steering mechanism can be virtually anything. We should ask the question, at least nominally, if a steering wheel is the thing we really want.

I'm not in any way claiming that steering cars with tillers rather than hand-wheels will save lives. No doubt such a change will cost more lives, as drivers will need to be retrained. You can make similar arguments about the brake pedal and the accelerator. They need to stay the same for fear that it will confuse people and create liability for the manufacturer. It seems to me that this form of reasoning extends, bit by bit, to the entire vehicle. In this way, the design of an auto is cast in stone. Only minor cosmetic changes are possible: new colors, more cup holders, decorative fender lines are OK. But in general, we leave the bloody thing alone.

What about seat belts and air bags? Certainly they have made cars safer. Indeed they have, but they are the exception that proves the rule. Seat belts and air bags were not a result of consumer demand for safer cars influencing the market. Auto manufacturers fought against each of these changes, as did many drivers. In point of fact, hardly anybody wanted them. The market, acting on its own, would never have put effective seat belts in cars. No, these safety devices were the result of government regulation forced by a small group of very determined people (most notably, Ralph Nader) who made an irresistible moral argument that gradually won over our conscious.

In our modern, 21st century society, except for residents of the largest cities where mass transit is a viable alternative, it's difficult to earn a living or otherwise function normally in our society without driving, or driving in, a car. It's absurd to think people employ any sort of accurate safety trade off analysis and consciously choose to risk their lives in cars. Most people simply don't have another realistic choice. Society offers the Hobson's choice: "The automobile, love it or leave it."

Well, I want that to change. I want an uncrashable car under my Holiday Tree. This pipe dream will still give us the freedom of personal transport delivered by the cars of today, except that it will be nearly impossible to kill yourself, or anyone else, in the practice of operating this uncrashable vehicle -- drunk or sober.

This new miracle doesn't need to be perfect. It merely needs to have a safety record similar to aviation.

This new miracle doesn't need to duplicate the paradigm of 4-wheeled autos. It merely needs to give the driver the same feeling of freedom and control. In fact, short of government intervention, I'd argue that the conservatism surrounding 4-wheel auto design effectively blocks any change that could make the current things significantly safer. Therefore, I predict that an uncrashable car will take a different form entirely. Maybe it'll be a rocket car. Maybe it'll be like a Segway. Maybe it'll have a tiller. These are engineering design details. The requirements are clear and the payoff -- over a million lives saved annually worldwide -- is worth it. Isn't it?

Sink, Tub, and Toilet Robots

A while back I received a Roomba robot as a gift. The thing works pretty well. We deploy it in the front lines of our rectangular foyer and it sweeps up dirt near the entrance door. This advance guard sweeping keeps the the marauding dirt from being tracked into the rest of the house, an area that is far more topologically complex and difficult for a robot to navigate. All in all, even considering this limited deployment, it's a great gadget. It's not a toy.

But sweeping floors is a relatively minor chore in the home. The alternatives of swinging a hand broom or pushing a vacuum cleaner are pretty easy too. Although the Roomba helps out, it does nothing to relieve the worst of the home chores -- chores like cleaning the sink, tub, and toilet in the bathroom.

Why can't a robot do that unpleasant yet very necessary chore? I want a Sink, Tub, and Toilet (STT) robot under my Holiday tree.

The amount of money people would pay for such a thing is considerable. Maid service costs near $100/mo and a large part of what they do is cleaning bathrooms. If people are willing to pay $1200/yr for a maid, it seems to me that they'd easily pay twice that for a STT robot with a three year warranty. For a couple of grand, you could sell a pretty sophisticated robot.

Those of you that have some familiarity with the programming of autonomous robots are no doubt anxious to tell me how hard a problem cleaning toilets would be for a robot. Really? Why? Because bathrooms are all so different and a robot would get confused or lost? I think you lack engineering imagination. What if Kohler or Moen sold such robots, specifically programmed to clean the shape of their exact fixtures? Wouldn't that make the problem tractable?

I think it can be done, and getting the market to buy self-cleaning sinks and toilets doesn't take government action or a Nobel prize-worthy discovery in physics. People would buy these things by the truckload and put them under their Holiday trees.

US Treasury Cash Cards

I have a Wawa gift card. I can load the card up with some cash, say a $20, and then buy coffee with it the rest of the week without messing with change and paper cash. It's a great convenience as Wawa doesn't round its prices to a nickle or dime. Coffee and a donut is $2.31, but if I use the gift card I can swipe it through the reader and be out the door a lot quicker than if I handed over a $5 bill and wait for the clerk to count out my two singles, two quarters, dime, nickle, and four pennies.

When I load up my Wawa gift card with a $20, nobody asks me my name, social security number, or does a credit check. If I lose the card, I lose the loaded value, of course, but this is no different than if I had lost the $20 bill directly.

The only thing I don't like about the Wawa gift card is that I can only use it at Wawa. The value stays in it forever, but that value is forever dedicated to the Wawa corporation. They already have my money in their bank account, earning interest, till I get around to visiting a Wawa store to reclaim my value in the form of a Peruvian Organic with hazelnut cream.

Why can't the US treasury mint our currency in a convenient form like the Wawa gift card? Then I could use plastic dough anywhere. Why do we still need to mess with the inconvenience of physical copper pennies? And more significantly, when countries like North Korea fund their economies by counterfeiting US paper money at a cost of billions to the US taxpayer, why do we still print large quantities of paper money?

Someday, I'd like to see a US treasury cash card under my Holiday tree.

Yes, I know about Visa and MasterCard gift cards. They've sold over a million of them. And, of course, I wouldn't just discard a Visa/MC gift card that I found under my tree. Such cards are private, stored value cards, similar to what I'm wishing for. They are similar, but not exactly the same thing. The differences are significant, particularly the fees charged. My Wawa card is free of fees, and does not ever lose value spontaneously. Not so with Visa/MC gift cards, which leak value all sorts of ways, most notably a monthly maintenance fee and a per-transaction service fee.

Another difference is anonymity. Although you can buy a relodable Visa/MC gift card for cash, you'll need to "activate" it, and that means giving up personal information. I don't have to reveal so much as my cat's name to get a Wawa card.

It seems to me that the solid popularity of flawed Visa/MC gift cards, despite their many costs and privacy shortcomings, is strong evidence that people would readily accept stored value cards in lieu of paper and metal currency should those shortcomings and costs be removed. If the treasury rolled out a system of cash cards, I predict it would be immensely popular and, ultimately, save our country a mint, or two.

There's also considerable evidence that private enterprise, on its own, would develop true currency innovations like a cash-like stored value card or on line e-cash if the US Treasury (and the treasuries of most other countries) didn't have a legal monopoly on the minting of cash. As things stand, private enterprise is very limited in what it can do, legally. There have been many attempts to create on line e-cash and smart-card cash equivalents, yet the US treasury has uniformly blocked these developments. Currently, it seems to me that the creation of private currencies or private stored value systems that approach the characteristics of paper cash is legally impossible. The only systems that have been allowed to flourish are credit card equivalents (e.g. PayPal) or checking account equivalents. Neither of these have the convenience, liquidity, or privacy aspects of cash.

Another flaw in these quasi-e-cash systems permitted by the treasury is the fact that transactions can be revoked. Payment received by the seller can be "charged back" to the buyer without the seller's permission. This lack of transaction finality sounds good to consumers as "purchase protection", but it's rarely used by them as such. Virtually all merchants have a warranty or exchange policy that consumers will turn to first before they ever think to call their bank to reverse the credit card charge. No, the vast majority of charge backs are a result of stolen credit cards. Effectively, sellers bear the cost of credit card theft. This patently unfair risk and cost externalizing by banks is why many merchants only accept cash. True electronic cash, with guaranteed irrevocable payment, would be accepted by all merchants instantly.

What I'm wishing for is a stored value system with the same liquidity, broad acceptance, anonymity, transaction finality, and freedom from fees as paper currency. I see no engineering reason this can't be achieved. The roadblocks are all legal and political.

Someday, if my wish is granted, the US treasury will come to the belated conclusion that redesigning the paper $20 bill yet again to thwart the North Koreans is just not worth it. They'll realize that the manufacture of millions of copper pennies is a waste of resources. For most purposes, metal coin and paper bills can be replaced by a stored value card. And, with some more luck, somebody will put such a card (funded) under my Holiday tree.