Profile For John L.

John L wrote:

There isn't enough paper. There aren't enough letters in the alphabet. There aren't enough words in the dictionary to list all the possible "points-of-failure" involved in putting a driver-less vehicle on the average US Interstate highway.

I've spent 30 years working in system automation (most recently in offshore drilling system automation), so I can speak with some small amount of authority. There is no way in hell that this little experiment will survive the first crash involving a fully-loaded, autonomous "semi" taking out a car-load of humans on a family vacation to "where-ever" America.

Truck loads of lawyers will be billing until the cows come home!

Well put John...

I did my time, 30 years of it on the delivery side; pharma, logistics and distribution, including two real time systems; one for Conrail (in-cab controls) and one that involved flow metering, movement, and control of combustible liquid. Been through my share of pre-production forced marches lasting for months.

Whenever public safety is involved software testing is incredibly complex and usually protracted until all regulatory oversight is satisfied. Quadruple redundancy and every little detail must be documented. And the testing...it never really ends, it changes to maintain compliance and functional upgrades.

Totally agree here with John...testing every possible scenario will be an infinitely difficult task. Testing things we know can happen is one thing...assuming you can surpass that. Then comes security. A whole other set of variables, a moving target that is virtually impossible to predict or prevent 100% of the time.

Recently FedEx was hit with a RansomeWare virus that crippled their operation for days...what is the risk of autonomous trucks being hacked for the purpose of committing a crime? I'd suggest it would be a very juicy target for a terror network. No one is talking about that...but make book on it, very difficult to prevent if this ever comes to fruition.

Although at least a decade or more away, I do believe there will eventually be systems that will automate trucking, in controlled and limited access corridors. But to the degree of eliminating the driver, I just don't see that as a possibility, ever. Too many things can and will go wrong. Eliminating the fail safe of an on-board human backup or assist...doubtful.

And Ed...just so we are all clear, computer systems and the infrastructure necessary to securely run them, specifically for the Fortune 500 Corporations, are housed in huge facilities called Data Centers, usually more than one for failover purposes. One of the largest banks in the US has two such facilities on the East coast near Philly, each over 300,000 square feet. Wall to wall, rows and rows of racked servers, data storage, EC gear, firewalls, network and telecom gear with overhead track work for the miles and miles of cabling. There are thousands of facilities like this throughout the US. We still need very large spaces for all of this equipment...

It's late and I don't wish to start into another long thread, but i will just say that, IMHO, we (they) are trying to solve the wrong problem. We don't need automated vehicles so much as we need automated roadways. It's the environment that these vehicles will operate in that is the real problem. It's total chaos out there and there isn't a computer built that could do it. At least not one that would fit under the hood of your average automobile.

No, we need to take control of the stream of vehicles on the highway essentially turning them into a virtual train. Oops, there I go off on a tangent that i said I wasn't going enter ... can we pick this up later?

There isn't enough paper. There aren't enough letters in the alphabet. There aren't enough words in the dictionary to list all the possible "points-of-failure" involved in putting a driver-less vehicle on the average US Interstate highway.

I've spent 30 years working in system automation (most recently in offshore drilling system automation), so I can speak with some small amount of authority. There is no way in hell that this little experiment will survive the first crash involving a fully-loaded, autonomous "semi" taking out a car-load of humans on a family vacation to "where-ever" America.

Truck loads of lawyers will be billing until the cows come home!

Is there some advantage to double clutching?

No, there really isn't.

But once in a while you'll try floating gears and get stuck in gear for whatever reason. Like say for instance you're starting down a hill and you're loaded heavy. You're not paying attention cuz a pretty girl was walking down the sidewalk and you let your rpm's hit the governor before you tried shifting. Now you can't get it out of gear. You try "kicking" the throttle hoping when you let off quickly it pops out of gear but you're already at the governor so hitting the throttle does nothing. And you're being pushed downhill by the load you're carrying so it won't slow down quickly enough on its own to pop out of gear when you let off the throttle. You're stuck in gear. The only way to release the pressure off the gears so you can get it out of gear is to push in the clutch.

So knowing how to double clutch at least gives you that instinct to hit the clutch if its stuck in gear. But I mean, that's a stretch. That's not even double clutching but at least I get a point for thinking up a reason to use your clutch when shifting!

Other than the fact that the manufacturer of the transmission (Eaton-Fuller) prescribes it in order to prolong the life of the transmission, no, there is no advantage to "double clutching."