limelight79

I have already addressed that point twice. Why do you keep ignoring it? Some improvements are good.

I understand the game can be played only while parked. But guess what? That software is in the car all the time. It's another place there could be a bug that allows access to vital systems. It's another place where there could be some weird interaction with other systems in unintended ways.

I don't understand why you keep ignoring that point. A software developer should understand the issue better than most people. I've given example after example of why we should be careful about what we put in cars and weigh the benefits against the risks. You refuse to acknowledge that there might even be an issue.

I can only hope when this shit fails, it doesn't kill me.

Nope.

I don't know how I can explain this better, so I'll summarize.

• Cars are safety-critical items, being they weigh 3000+ lbs and travel at high speed and can kill people when something goes wrong.

• It's critical that the software that controls the drivetrain, brakes, etc. be as perfect as it possibly can be because of the first point.

• Adding more features increases the likelihood of something going wrong.

I don't understand why this is even an argument. It's common sense. Why would anyone disagree with those three points? (Unless they're the reason vehicle deaths are up...maybe they just don't care and see their car like a cheap appliance and ignore the "potential to kill" factor.)

Your comments about the hatch and the tires only prove my point: Auto manufacturers already make plenty of mistakes on things that have existed for decades, things that should be solved problems by now. Why would software be any different? They're going to fuck things up there, too. And already have; I gave examples before.

More complex software means more bugs; anyone in software development or testing can tell you that. These are known facts. What if that game has some bug in it that lets hackers take over the brakes remotely? Unlikely, sure. Impossible? Definitely not - again, look at the Uconnect 8.4 issue in FCA vehicles a few years back; a remote attacker could break into the car and do just that. It's an extremely scary bug. Fortunately they were white hat and FCA actually listened to them and worked with them to patch the bugs, but what would have happened if it hadn't been white hats that found it? Or the manufacturer didn't care to listen, as often happens with software vulnerabilities?

No one would accept a computer program that runs an X-ray machine overdosing people (which has happened). It's the same thing. The FDA would hopefully never approve an X-ray machine that has Tetris on it, either, for the same reasons we shouldn't accept it in cars.

The advantages of computer control are huge, we're getting more mileage and more power out of smaller engines all the time. I'm not recommending we get rid of computers in cars. I'm saying it's imperative that any additional features are weighed against the benefit. Playing Tetris on a dashboard screen is not a useful feature in a car that can't be easily handled by the smartphone you almost certainly also own.

I don't think there's any more to be gained by discussing this further. I can't make it any clearer. Good day.

The reason I didn't understand your comment is that your examples just proved my point - safety critical items WERE NOT VETTED correctly. What makes you think the auto manufacturers are being careful with software? They aren't, or are doing the bare minimum. I listed multiple software issues with various manufacturer that never should have happened.

Look, I'm not a luddite. I like having nav and satellite radio and all of that. But adding a game to the already questionable software development in a car is insane. Cars are a safety critical item and everything added to them should be carefully considered and thoroughly vetted.

Someone compared it to a cell phone, but cell phones aren't 3,000 lbs of heavy machinery barreling down a highway at 70 mph. A phone needing to reset or whatever is a minor inconvenience compared to what could go wrong with a serious software bug in a car. I'm not sure why people would argue against this, it seems self-evident. I provided examples.

I'm not sure I follow your point here. Even necessary parts of a car failed for you, and almost caused injury. Now people are advocating adding unnecessary parts to cars that may also fail and cause injuries or death.

I worry about features in cars.

For example, our Mazda has headlights that turn with the steering wheel (ala Tucker Torpedo's center light). Neat idea and it is a useful feature while driving at night on the rural roads by our house. But what will happen when it fails, and how much will they cost to replace? (I've been told they "fail straight ahead", but who really knows for sure. I'm hoping we get rid of the car before that happens.)

I saw a pickup with automatic folding mirrors having an issue with them folding and unfolding while the guy was driving. I followed him through several traffic lights and watched it happen a few times. Automatic folding mirrors would be a nice feature on my pickup, but I'd rather not have them fail especially when I'm towing a trailer and be completely blind to the rear-right.

I've seen pickups with the running bar that folds out. I'm not sure there's much value in that other than "oooh shiny" but if it fails to open while I'm getting out, it could hurt.

Our Mazda again has several software bugs in the infotainment system. None of these are critical, but it does make me wonder how much testing they did on the software that controls the brakes, for example. Are the brakes going to fail to release someday? I already know the computer has some control of them, because of the auto-hold feature that I usually keep turned off, and because I sometimes notice a slight delay in releasing the brakes when I take my foot off the pedal.

The FCA Uconnect 8.4 infotainment systems allowed an attacker to remotely take over throttle, brakes, etc. until they were patched. That's an obvious safety issue.

And that crash at the Peace Bridge last week, it seems very likely it was caused by an issue with the car, rather than the driver (there is evidence the driver was alert and trying to stop, and he swerved around another car that turned in front of him before the crash). Turns out the right-hand drive version of that car had a recall of an issue with the accelerator...which supposedly did not effect left-hand drive vehicles. But here we are with two people dead and a third injured from a vehicle that may have been accelerating out of control through no fault of the driver.

The point is that including additional features, even if only software, increases the complexity of the system and makes errors more likely. It increases the chances of some unexpected interaction or failure. It increases the surface of a software attack for a potential safety issue. It makes the code that much harder to test for bugs in general and security in particular.

Trump took Abe Simpson seriously.

Wait. They believe JFK is alive, too? I knew about JFK, Jr but not the JFK part.