Misc: August 2009 Archives


August 28, 2009

The LA Times reports that Raymond Azar, a Lebanese citizen was rendered from Afghanistan to the US:
Reporting from Alexandria, Va. - A Lebanese citizen being held in a detention center here was hooded, stripped naked for photographs and bundled onto an executive jet by FBI agents in Afghanistan in April, making him the first known target of a rendition during the Obama administration.

Azar, 45, pleaded guilty Tuesday to conspiracy to commit bribery, the only charge against him. He faces a maximum of five years in prison, but a sentence of 2 1/2 years or less is likely under federal guidelines.


"The FBI followed standard operating procedures when transporting prisoners to the United States," Gina Talamona, a Justice Department spokeswoman, said Friday. She said restraints "were used with the sole purpose of ensuring the safety of the defendants and the agents."


They said he was photographed naked and subjected to a cavity search to ensure that he did not carry hidden weapons and was fit for travel. Court records confirmed that Azar was shackled at the ankles, waist and wrists and made to wear a blindfold, hood and earphones aboard the plane.

This seems like a pretty aggressive set of security measures: presumably the FBI agents are armed and given that the prisoner has been restrained, one has to wonder exactly what the FBI expects him to do in order to escape. I'd be interested to hear what the standard procedure is for transporting known violent offenders inside the US. A little research finds this document from Virginia:

When an officer transports a prisoner in a non-caged vehicle, the prisoner shall be placed in the right front seat and secured with a seat belt. The prisoner shall be handcuffed with his or her hands behind the back, palms outward. A lone officer shall never transport two or more suspects in a non- caged vehicle unless directed by the on-duty supervisor. [Note: Some agencies require officers to place a suspect in the right rear seat, rather than the right front seat. Officer safety considerations can be argued to support either procedure.]

With that said, if you're transporting people you intend to aggressively interrogate, one might imagine wanting to isolate them to make them more vulnerable to future interrogation. Note that I'm not endorsing this, and even if that is the intention, it's hard to see why you would need to do this for someone suspected of bribery who's going to serve a couple of years.


August 12, 2009

I just got back from the EVT/WOTE 2009 conference. More accurately, I should perhaps say that I just got back from a conference held by United and Air Canada. Out of the 72 hours or so I spent out of California, I spent about 33 hours waiting for planes to board, sitting on a plane waiting for it to land, sitting on a plane in the air, sitting on a plane circling the airport, or sitting in a hotel waiting for the next plane after I missed my connection.

My routing out to Montreal was SFO-BOS-YUL, but when I got to Boston and finished walking the 5 or so miles to the satellite terminal of the satellite terminal where Air Canada takes off from. There I was informed that the flight to Montreal had been cancelled and that my new routing was through Toronto and then to Montreal. Suddenly my 9:30 arrival turned into a 11:30 arrival (if I can get on the full 10:30 flight) or more likely a 1:30 arrival (if I end up on the 12:30 flight). We get most of the way to Toronto and the pilot announces that Toronto is closed due to weather and that we're going to circle around. After circling for a while, we land in Buffalo to refuel, but by then the weather system had rolled into Buffalo and we spent about 2 hours on the ground waiting for it to pass so we could take off. Remember at this point that Buffalo is something like 100 miles from Toronto by car. In fact, two of my fellow passengers decided to rent a car and (after arguing with the FAs quite a bit) were allowed off the plane. We eventually took off and arrived in Toronto at around 12:30. Luckily, everything else was backed up and so our 12:30 flight had turned into a 2:00 flight, or, as it eventually turned out, a 2:45 flight getting into Montreal around 3:30.

On the way back things looked a little better until my flight from YUL to IAD was delayed in takeoff by 30 minutes and then in the air by another 20 minutes or so (we aborted one pass on the runway and then had to approach from the other direction.) I got to the gate for my IAD-SFO flight right after they had closed the door and even though the plane was still at the gate and the guy right behind me was United Global Services (their highest status level) the GAs refused to let us on. As this was the last flight out of the night, I spent the evening at the beautiful and classy Herndon Holiday Inn Express and took the 6:57 flight to SFO, arriving 9 hours after my scheduled arrival.

Obviously, this stuff is annoying, but I'm not sure there are any really useful lessons here. Obviously there's not a huge amount to do about the weather. On the other hand, in Buffalo it was really clear we were going to be on the ground for several hours so it seems like Air Canada could have let us off. For some reason they really hate to do that though. As for my return flight, I suspect we're victims of management by objective: airlines care a lot about on time statistics and delaying the flight for us, even though they knew we were connecting, threatens that. If you always held connecting flights then any single delay would have a ripple effect throughout the system. That said, missing connecting flights often has a much bigger effect on passengers than having a flight be 10-30 minutes late, so it would be nice to have statistics that measure this inconvenience. This actually seems like something you could do: the airlines know what passenger's ultimate destinations are and if/why they get rescheduled, so maybe you could replace a flight measurement with a measurement of the number of passenger-hours delayed or maybe the average number of passenger-hours delayed.

For instance, if I was flying SFO-ORD-IAD and my SFO-ORD flight is delayed but I still catch my connection and that's ontime, then this is a 0 delay. On the other hand, if my SFO-ORD flight is delayed and I miss my flight out of ORD and have to catch one that arrives two hours later, then this is two passenger hours of delay even if all the flights out of ORD are ontime. On the other hand, someone just on the ORD-IAD flight doesn't count as delayed at all. I'm not sure if this is a good metric, but the airlines should have all the underlying information and the software to compute this shouldn't be too hard.


August 3, 2009

The cash for clunkers program presents an interesting security challenge: a lot of the point is to take cars off the road, so this means we need to be sure that cars that are turned in don't end up back on the market. The natural solution is to put the vehicles "beyond use permanently", as they said of the IRA's weapons. But you want a simple and safe method of disabling them The WSJ covers the solution (literally), some stuff called "sodium silicate" (pointer due to Terence Spies):
What Mr. Mueller discovered is that sodium silicate is the designated agent of death for cars surrendered under the federal cash-for-clunkers program. To receive government reimbursement, auto dealers who offer rebates on new cars in exchange for so-called clunkers must agree to "kill" the old models, using a method the government outlines in great detail in its 136-page manual for dealers: Drain the engine of oil and replace it with two quarts of a sodium-silicate solution.

"The heat of the operating engine then dehydrates the solution leaving solid sodium silicate distributed throughout the engine's oiled surfaces and moving parts," says the National Highway Traffic Safety Administration publication. "These solids quickly abrade the bearings causing the engine to seize while damaging the moving parts of the engine and coating all of the oil passages."

Sounds like it beats sugar in the gas tank.