You certainly don't have to be an idiot to fail to handle a stuck throttle, most people will never have the experience, and if it becomes a problem starting at highway speeds, many drivers may feel the need for both hands on the wheel. I would want to know in advance that I could turn the engine off without engaging the steering wheel lock. And that it kills both the electric and gas power in a Prius. Shifting to neutral is likely to to be the last resort, but most effective.

There could well be an issue with the anti-lock brakes as well, if the braking power is being limited, they may well not have the stopping power needed to overcome the engine, the recent police assisted stop was made made after slowing with the emergency brake which probably is mechanical and will actually lock the wheels. That would explain the claims that the brakes were full on but the car didn't slow down, and the odd signs of only partial brake application noted in some cases. Apply full engine power and limit brake effectiveness, if that bug could be proved it would explain many things.

It would also probably exonerate the man who is in prison after unintended acceleration.

There are black boxes, and several stories circulating about them. The first says there is only one laptop in the entire USA which can read the black box codes. And the other claims that the black boxes are now not recording brake and throttle settings, or so Toyota claims. Call me paranoid, but if that data showed something which proved the throttles didn't stick or the brakes weren't used, I bet Toyota would give the NTSB a hundred of the laptops. Of course I would promptly crash a Toyota with brake and throttle on and see what the readings were, "I am not a trusting person."

The link I gave to the Associated Press story notes both accusations, and a number of other steps Toyota has taken to hide the black box data. As Nixon found, the damage from being caught in a coverup is worse than the crime.

I see that Toyota apologists are in full cry, trying to prove that Toyota unintended acceleration is really due to a bunch of old people who don't know how to drive, young people driving badly, and almost anything other than a real defect in the throttle software. Steve Wozniac (Apple founder and designer) says he can replicate the problem at will, and he claims that Toyota is ignoring him. Wonder why.

Other things to consider:
- The same old and young drivers don't report unintended acceleration for other brands.
- The unintended acceleration in other cars was a real human design defect[1]
- The press (other than Slashdot) has ignored Wozniac as well, for the most part

[1] It seems that if pedals are put in a "stick shift" configuration, throttle on the right, brake in the middle, and (space for) clutch on the left, people don't hit an unintended pedal. Putting the brake and accelerator in the middle space, under the wheel, is more likely to result in error. Since the elderly and young men are somewhat more likely to have driven manual transmission, they are more likely to stomp in the middle of the floor looking for the brake under stress. Call that human error exacerbated by design factors. But those were not situations lasting for minutes.

The issue of sexual preference is not related to morality, but to blackmail. If someone is openly gay, that shouldn't be an issue. If someone is secretly gay, bi, cross dresser, etc, then that employee might be subject to blackmail attempts. Other secrets apply, affairs, fake degrees (or cheating to get a degree), etc.

There is actually a reason for this, people who can be pressured are additional risks. It is an unpleasant part of having clearance, having been through DOD and DOE I have seen the first few rungs of the investigation, I'm sure checking is greater for people who do top secret work, as opposed to IT people who might get a change to see some.

The problem with an effective tarpit is that it ties up one of your sockets as well as the originator's socket. I played with a program to open a socket, and for each incoming packet wait some time and NAK. Too effective, kept the attack socket up for hours.

What are the odds of a legitimate using the same source IP?

The problem with certificates is that there are stored somewhere easier to find than inside someone's head. A certificate is a key anyone can use. But, you say, "I can password the certificate!" So the security comes back to a password again.

For really security conscious users there will be a password on the filesystem, one on the account (login), and one on the certificate. They will be non-trivial and different.

My point is that a certificate by itself is less secure than a password, since anyone can use it. Only by combining the certificate to identify the system and the passwords to prevent misuse will you have acceptable security. The hardest part is finding the user who will actually faithfully use all the available security.

Combine this with human factors, like fingerprint readers, and port knocking and you can reach whatever combination of levels matches your requirements.

One solution would be to have a countdown clock on the error message, so it couldn't be cleared with a click for some number of seconds. Having that time get longer would improve attention. Having the error send or log the user override would also increase the attention paid to the error.

If you want to really get attention, make them fill in a CAPCHA field, or call the help desk for a password. It all depends on just how hard you want to push, and how much management wants you to push.

If an error really is serious, an unclickable warning, like "you have ignored an error. This action has been logged and you will be personally responsible for any loss or compromise of company assets." But that's not the path for "your disk space is getting low" that's the path for "you are sending a confidential document unencrypted" type warnings. Logging user action to the help desk will improve attention no end, if and only if management backs you.

Of course Microsoft has figured out a way to deal with this, they will charge the legitimate user again. The object of the program is to generate revenue, not to protect the rights of legitimate users. The only question is if the false positives are accidental, or a case of guilty unless proved innocent.