Main | April 2005 »

March 30, 2005

Autism and Google versus Math, or A Little Research Is A Dangerous Thing

Tyler Cowen at the economist blog Marginal Revolution complains that the rise of autism diagnoses in America from 1 in 2500 to 1 in 166 seems like an urban legend: he can't find any reputable citations for the "1 in 166" figure on Google.

Tyler, that's probably because you didn't think to perform the secret decoder trick where you rewrite the 1/166 as 6/1000. Looking for "autism 6 1000" gets you this citation, among others.

There's some kind of point here about unspoken knowledge: when I, with an excessive math background, see "1 in 166", I immediately recognize that as a rewriting of "6 in 1000". But Tyler didn't, despite his smarts and despite being an economist (which is plenty mathematical). So what in my past let me clue in to the secret mathematical origin of the 1-in-166 figure?

The other point is the one we all know: never confuse "what I find on Google" with "everything there is to learn". Not this decade, anyway.

Posted by danielstarr at 01:39 PM | Comments (0) | TrackBack

No Government in Iraq -- and That's Good News

The Iraqi National Assembly keeps failing to form a government for the country. This is backhanded good news: it means the leading Iraqis are confident that Iraq's insurgency is dwindling.

The winners of the Iraqi elections -- the Kurds and the moderate Shiites -- both have reason to take their time before cutting a deal. The Kurdish leaders want de facto independence for the Kurdish provinces -- and guess what? They already have de facto independence for the Kurdish provinces. They'd like Kirkuk annexed to the Kurdish provinces, too, but that's just a bonus. The Kurds have been self-governing ever since the United States forced Hussein to withdraw from northern Iraq after the original 1991 Gulf War. Every month that goes by without an overall Iraqi government is another month that the Kurds get to continue having their goal -- an almost total independence.

The moderate Shiites, for their part, have reason to wait for as long as it looks like the insurgency is fading, because the more peaceful the country is, the more the Shiite Arabs -- who make up perhaps 60 percent of Iraq's population -- can insist on having their way as the majority, as opposed to making concessions to the Kurds -- or, especially, the favored folk under Saddam Hussein, the Sunni Arabs.

But both Kurds and moderate Shiites would lose in a civil war: that would hand power to the radical Shiites like al-Sadr, and make mixed-ethnicity cities like Mosul or Kirkuk into slaughterhouses like Sarajevo during the Serb-Muslim-Croat wars of Yugoslavia/Bosnia in the 90s. So if the Iraqi politicians are being slow to get their act together, that tells us they think the insurgency is fading, civil war is not going to happen, and the insurgents are eventually going to settle for an "exit strategy" of peaceful politics.

And the Iraqi political leaders are probably a lot better judges of the insurgents' strength than most American soldiers, journalists or politicians.

Of course, the Iraqi politicians could be overconfident: in the 1850s, few American politicians believed the United States would fall into civil war, and so they refused to compromise with each other on slavery. And we all know exactly how that turned out.

[E1] Added link to Beltway Traffic Jam.

Posted by danielstarr at 12:42 PM | Comments (0) | TrackBack

March 29, 2005

Traditionally, They Kill Japanese Storytellers Just As They're About to Write the End

I have a happy weakness for sci-fi television, and the best science fiction television series these days are all Japanese anime shows. (Thank heavens for Netflix.) Japanese anime series tell some pretty fine stories... with just one problem. They don't have endings.

I exaggerate. A few of my favorites, like The Vision of Escaflowne and Angelic Layer and Full Metal Alchemist, have pretty darn dramatic and well-thought-out endings. But I just came to the end of Martian Successor Nadesico, and there was no end there.

Nadesico tells the story of a battleship in a war between Earth and Jupiter, where the main character has mysterious ties to an ancient Martian culture, and also can't make up his mind which of his crewmates he's in love with. (There's a subgenre of anime called "harem" shows, featuring one clueless male and a bunch of women who may or may not be romantically interested... but I digress.) You might expect Nadesico would end with, say, the end of the Earth-Jupiter war, or the secret of the Martian culture, or the hero maturing enough that his love life would actually make sense to an outside observer...


He kisses one of the girls. I have no idea why, except they were childhood friends. The war goes on. The Martians remain unexplained. And that's... it.

In fairness, Nadesico is a comedy. But lately I've run into a whole bunch of series where you watch the end and you really feel the writers made up the ending at the second-to-last episode, or had a big ending idea that exceeded their talents and came out as twenty minutes of viewer confusion. (Rahxephon... Berserk...)

Frankly, it all reminds me too much of what I don't like about Neal Stephenson. But I watch them anyway, for the same reason I watch Stephenson: the ending may be a train wreck, but the middle is often a heck of a good ride.

Anime as the supreme guilty pleasure -- that's for another post.

Posted by danielstarr at 02:01 PM | Comments (0) | TrackBack

Future Combat Systems: Why the Army Wants its Own Aircraft Carriers

The Army wants to spend a huge amount of money to replace some of its battle-tested heavy-armored tanks with a bunch of untested robot- and missile-carrying vehicles that will each have about as much armor plating as the average can of chicken soup. Kevin Drum doesn't see the point; the New York Times points out that the General Accounting Office is predicting they'll be hugely expensive; and Fred Kaplan in Slate doesn't think they'll be useful. But the truth is these new "Future Combat Systems", or FCS, are probably one of the best investments the Army could make to win wars faster with fewer dead soldiers and fewer dead civilians. With FCS, the Army is finally hoping it can do what the Navy did two generations ago, and switch from fighting with battleships to fighting with aircraft carriers.

A battleship -- the navy kind, not the kid's game -- is, of course, a floating mass of armor plate and big guns, whose purpose in life is to steam up close to bad guys and pound the crap out of them. And an aircraft carrier, of course, is a giant flat-topped ship bearing a suspicious resemblance to a floating cheese tray a quarter-mile long. An aircraft carrier has hardly any armor or guns. All it does is provide command, upkeep, and landing space for several dozen aircraft.[1]

But in the years leading up to World War II the US Navy did some wargaming and realized that an aircraft carrier could be far more powerful than a battleship. In fact, the Japanese attack on Pearl Harbor did little damage to American sea power precisely because all three of our aircraft carriers were elsewhere. While a battleship only commands the waters in the range of its guns, an aircraft carrier can send planes to deliver bombs on enemies hundreds of miles away. And while a battleship is blind to enemies until it's practically on top of them, an aircraft carrier can spot enemies with any of its aircraft while it stays safely out of reach. Aircraft carriers can see without being seen, attack without letting the enemy attack back, and exert force over a far larger area than a battleship.

And that's the idea behind FCS: to let our troops see more enemies before they get a chance to strike, to bring more of our weapons to bear on enemies without letting the enemy do the same, and to let the same number of troops defeat enemies and protect friends over a larger area. Technology has reached the point where a swarm of missiles, sensors, and air and ground robots can let the Army duplicate some of the strengths of a naval aircraft carrier group with a land-based FCS "drone carrier" group.

Moreover, because "drone carrier" vehicles don't need to get close to the enemy to attack, conceptually they can operate with less weight of armor than tanks. Less weight means less fuel needed and smaller planes needed, which means the ability to operate FCS units in enemy territory further in and faster than heavy tanks can -- or to operate in terrain that's too rough to get tanks into at all.

If We'd Only Had FCS In...

We can talk about the technology that goes into FCS, and how exactly they'll pull off this "drone carrier" approach to war, and what kind of missions will go to FCS and what missions will remain with heavy tanks -- but honestly, it's going to take years and a ton of money to make FCS work, just as it did for the Navy with its first aircraft carriers. So what are we spending it for? It's not worth an unlimited price. On the other hand, if we had FCS in Iraq now, we could protect convoys from more ambushes and clear out insurgent strongholds like Fallujah with fewer friendly and civilian casualties. If we'd had FCS in Afghanistan, we could have caught more Al Qaeda members fleeing from Tora Bora, and lost fewer American lives in Operation Anaconda. If we'd had FCS we could have credibly threatened a ground invasion of Kosovo sooner, giving Milosevic an earlier excuse to surrender, and ending the killing of Kosovar Muslims that much sooner. If we had FCS we could quickly send ground troops to places like Sudan and Rwanda with much more strength and less vulnerability than now, which would increase our ability to stop genocides, or at least to credibly threaten to stop them. And if we had had FCS, we could have done a much better job reinforcing our Rangers under fire in Somalia.

All that said, my instinct is that the GAO is right, the technologies needed to pull off "drone carriers" are not mature enough, and the FCS schedule needs to be slowed down. And I sometimes agree with those who would like to see the Army commit to being as organized about fighting insurgencies and guerrillas as conventional opponents. You could certainly argue that FCS is another case of America spending too much on defense and not enough on, say, child care. But it's just plain stupid to talk about FCS as if it's another F-22 or Comanche, a specialized system that will only make us better in a few rare situations. If and when we make FCS work, those drone carriers will do our soldiers and the civilians good in every war we fight.

FCS' Tripod: a Swarm of Sensors, a Swarm of Missiles, and the Network of a Thousand Headaches

For FCS to work, a lot of technologies are desirable but three things are absolutely necessary: lots of robotic eyes and ears in the air and on the ground; lots of lightweight, smart, powerful missiles; and a network that lets every sensor and every missile be ready to help every soldier in the area. Ironically, the technology that's the hardest is the one we most take for granted in everyday life. The robotic eyes and ears are already hard at work in Afghanistan and Iraq, and the "rockets-in-a-box" smart portable missiles are pretty far along in development -- but making the network operate properly under battlefield conditions promises to be even harder than booting Linux on a dead badger.

You probably know a few places where your cell phone always loses reception. Now, what the military needs to make FCS work is a network that operates anywhere, even without cellular towers, and never loses reception, even when parts of the network are being attacked with large explosives or high velocity rockets, and always puts your call through, even if 50 people are trying to talk to the same sensor at once. If you know how to do that, I can tell you two things. Number one, the Defense Department would like to hire you. Number two, every phone company on the planet would like to hire you.

So it's not too surprising that the military's next generation communications system is, well, not quite performing up to specifications.

But let's suppose that we did have a brigade of lightweight vehicles, carrying hundreds of flies-around-corners missiles, surrounded by a cloud of all-seeing robots, and all linked up by a network proofed against not only zombie badgers but even cybernetic gorillas. What good does this technology tripod do for the soldiers? How would they carry out missions differently?

A lot depends on how close FCS gets to that kind of technology perfection. But we can certainly sketch in some examples, just to get the feel of what happens when people start taking this drone carrier approach to land warfare.

Ambush protection: Right now, the easiest way for a convoy in Iraq to detect an ambush is when explosions start going off around them. If you have enough robot sensors in the area, you can send one flying overhead in the convoy's path, which spots any ambushers who aren't skilled enough to hide themselves from above as well as from in front (and you'd be surprised how many that is). More to the point, most ambushers count on getting away through culverts or side streets to ambush you again. That gets a lot harder if you can send a flying camera along to track them from the air.

Rapid reinforcement: If you're a soldier under fire, and you don't have a tank with you, you may have to wait anywhere from 5 to 30 minutes for air support to arrive and drop bombs on the bad guys -- by which time a lot of people can be wounded or dead. But if you have hundreds of precision missiles on call through the FCS network, you can cut that 5 minutes to 30 seconds for the missiles to drop on the enemy. This is still not as good as having an Abrams tank with you, but it's a whole lot better than waiting for airpower. And missiles, being smaller and more precise than bombs, can give more options for killing bad guys without hurting civilians. Finally, if they're the new breed of loitering missiles, they can actually launch in advance, hang around overhead and wait for the enemy to come out before dropping on him. (In Afghanistan, Al Qaeda fighters were often canny enough to get under cover whenever the airplanes showed up, only to come out and shoot at our soldiers some more when the airplanes flew off.)

Discreet patrolling: another benefit of all those robotic eyes is that you can keep watch for bad guys in the city without sending tanks crashing through every street. Of course, really stealthy bad guys aren't going to be spotted by a robot camera floating through the sky; this won't get you free of people dying at checkpoints. But a lot of bad guys aren't really stealthy, and almost no one is really stealthy all the time. And for ordinary citizens, a floating camera is creepy but not nearly so oppressive as a 70-ton tank.

Fast deployment, flexible maneuver: one problem the generals had in both Iraq and Afghanistan was that our heavy equipment, being heavy equipment, had to come in by and receive supplies through major airports or seaports. Since Afghanistan's Al Qaeda hideouts were in the roughest part of that very rough country, that meant our soldiers went in with little more than the equipment on their backs; and in Iraq, our major forces were restricted to coming in by the south, not the north, and to go through shooting galleries at every major bridge. If the FCS units get down to their 20-ton-per-vehicle goal, they can literally drop out of airplanes into the mountains of Afghanistan, or landlocked northern Iraq; and with fewer fuel requirements, they could quickly penetrate much further into, say, central Iran than heavy armor units could. This would be very risky -- FCS units aren't going have the armor to "charge!" with minimum loss the way tanks and Bradley fighting vehicles can -- but sometimes surprising the enemy is the way you get him to give up fighting.

Outnumbered but never outgunned: the most common way for American soldiers to take losses is when they find themselves suddenly vastly outnumbered -- an Abrams tank is an amazing machine, but its main gun still only shoots one round at a time. A missile carrier, on the other hand, can certainly launch 20 missiles at once, if there's a network to assign targets for all of them. A company with FCS support is going to be able to risk lopsided engagements in a way that current forces can't. This is especially true if America ever has to send ground troops in without first dominating the airspace.

FCS units, the new cavalry; tanks, the new infantry?

Back in the seventeenth and eighteenth century, the role of cavalry was to be good at scouting deep into the enemy's terrain, at making trouble for the enemy in unexpected places, and at providing sudden and decisive reinforcement to rescue a friendly unit or defeat a wavering enemy force. Meanwhile, it was the infantry's job to be tough and reliable, a force that could stand up under attack in a way that the easily-disarrayed cavalry could not.

By that standard, even though tank-heavy units are often designated "cavalry", it's the FCS units -- if FCS works -- who will be the Army's 21st-century cavalry, and the heavy armor forces will be the 21st-century infantry. Even when FCS works, these "drone carrier" groups are still going to lack the armor of heavy tanks, and that means that when the key issue is reliability and toughness, the Army's answer will involve Abrams tanks and Bradley fighting vehicles. But when the goals are speed, knowledge, surprising the enemy or reinforcing friends, the FCS units will do the job better.

Of course, by the 1800s cavalry units were a luxury for many armies: valuable but not a center of power. And maybe the FCS, too, will stay too expensive for a long time for us to buy it -- there are other things to spend money on, both in and out of defense. But FCS would make every war we fight go better for both our soldiers and the civilians. So we need to take the idea of drone carrier units seriously, whatever we decide about the price tag.

Have you seen any battleships lately?

[E1] The aircraft carrier folk can have my body after the battleship folk are done with it.

[E2] tag =

[E3] And a link to OTB's Traffic Jam.

Posted by danielstarr at 04:49 AM | Comments (0) | TrackBack

March 28, 2005

China's Edge Over Russia: More Lawyers

Why is Russia sinking into the ranks of the left-behind countries? Why is China joining the wealthy? Well, it sure helps to have law and order.

In China, the government has been trying to crack down on local officials and erode their habit of treating a town or village's citizens as so many cattle to be milked; this Washington Post article about women buried alive makes clear that they've got a long way to go, but they're trying. These days Chinese citizens actually will file lawsuits to try to enforce their rights, even against those with government connections. But in Russia, far from establishing the rule of law, the Post reports that under Putin crime by the police themselves has exploded in recent years.

China's other secret weapon of economic growth, of course, is the huge savings rate of ordinary families, which means that any economic opportunity that opens is instantly filled by a mom-and-pop business...

Posted by danielstarr at 08:08 AM | Comments (2) | TrackBack