The Natural Order of the Automotive Universe Prevails Once Again

[FoMoJo]

Ford closing in on Toyota as favorite of buyers

The Ford brand is closing in on Toyota as the favorite of retail auto buyers, a sign of rising popularity of the Fusion mid-size car and Escape sport-utility vehicle, and of growing dismay with the Japanese manufacturer.

Toyota Motor Corp.'s namesake brand, a default choice of a generation of car buyers who admired its high quality, lost ground during the past six years amid recalls, natural disasters and stiffer competition.

Its share of the U.S. retail market fell to 13.5 percent last year from 16.3 percent in 2008, according to data provided to Bloomberg by IHS Automotive using Polk vehicle-registration records. Toyota's 2.8 percentage point loss matched the gain of Ford Motor Co.'s main brand, to a 13.2 percent share, over the same period.

Retail registrations are the best measure of the tastes of individual car buyers because they exclude bulk fleet sales to corporate and government customers. Toyota dominated the retail market before the recession, with top sellers such as the Camry sedan and Corolla compact. Those models now face tougher competition as car shoppers favor Ford models including the Fusion, which draws frequent comparisons to an Aston Martin, and a restyled, fuel-efficient Escape. Hyundai Motor Co. and Subaru, the auto unit of Fuji Heavy Industries Ltd., also gaining ground.

"Before 2010, Toyota's image was bulletproof, and while it is still strong, it's not rock solid and as perfect as it was before," Tom Libby, auto analyst for IHS Automotive, said in an interview. "It now appears their march forward has been slowed."

Read more from link...

[Laconian]

It's funny that they highlight Toyota's fall from grace with quality, because Ford isn't exactly doing well in that arena these days, either.

[FoMoJo]

It's funny that they highlight Toyota's fall from grace with quality, because Ford isn't exactly doing well in that arena these days, either.

Toyota's problems have seen greater mention in the media due to recalls and associated litigation; though not nearly as harshly as Ford was trashed over Firestone's tire problems a dozen or so years ago. 

[Laconian]

Oh, BTW - did you see the final resolution on the Toyota issue?

The problem was unbounded recursion leading to a stack overflow, which corrupted the heap. The failure mode should have been for the controller to be rebooted (it would've happened in <1sec) but instead, things froze up.

Personally, I'm really surprised that recursion is A Thing in safety-related systems. I would banish that shit upfront.

[FoMoJo]

Oh, BTW - did you see the final resolution on the Toyota issue?

The problem was unbounded recursion leading to a stack overflow, which corrupted the heap. The failure mode should have been for the controller to be rebooted (it would've happened in <1sec) but instead, things froze up.

Personally, I'm really surprised that recursion is A Thing in safety-related systems. I would banish that shit upfront.

So bad code was responsible for the unintended acceleration?  Never saw it mentioned in the media.

[Laconian]

So bad code was responsible for the unintended acceleration?  Never saw it mentioned in the media.

http://www.edn.com/design/automotive/4423428/Toyota-s-killer-firmware--Bad-design-and-its-consequences

[FoMoJo]

http://www.edn.com/design/automotive/4423428/Toyota-s-killer-firmware--Bad-design-and-its-consequences

With some of the UA incidents, it seemed that it must be related to the software.  I'm glad it was finally pointed out.  I, also, wonder how they managed to keep it out of the mass media.

[ifcar]

http://www.edn.com/design/automotive/4423428/Toyota-s-killer-firmware--Bad-design-and-its-consequences

This seems to quote the plaintiff's expert without any rebuttal. Not saying it's not accurate, but it's rather one-sided.

[FoMoJo]

This seems to quote the plaintiff's expert without any rebuttal. Not saying it's not accurate, but it's rather one-sided.

How do you defend against something like that?  Obviously, Toyota couldn't.  What is more disconcerting is that it is likely true of other manufacturers.  Nobody wants to admit that the car's computer can freeze and you can't shut the car down.  They need a kill switch.

[Soup DeVille]

Oh, BTW - did you see the final resolution on the Toyota issue?

The problem was unbounded recursion leading to a stack overflow, which corrupted the heap. The failure mode should have been for the controller to be rebooted (it would've happened in <1sec) but instead, things froze up.

Personally, I'm really surprised that recursion is A Thing in safety-related systems. I would banish that shit upfront.

In automation, there are specific standards (IEC-1131-3 if you're interested) related to the safety of the machinery being independant of the programmable code. THe DOT has up until recently been adamant about certain systems retaining direct mechanical control (there must for instance, be a mechanical connection between the steering wheel and the wheels), but has been relaxing those standards for some time now.

[FoMoJo]

In automation, there are specific standards (IEC-1131-3 if you're interested) related to the safety of the machinery being independant of the programmable code. THe DOT has up until recently been adamant about certain systems retaining direct mechanical control (there must for instance, be a mechanical connection between the steering wheel and the wheels), but has been relaxing those standards for some time now.

The throttle/gas pedal for instance?...though I'm not sure if 'throttle' is appropriate anymore.

[MX793]

The throttle/gas pedal for instance?...though I'm not sure if 'throttle' is appropriate anymore.

Lots of cars have no mechanical linkage between the throttle and the pedal.  A number of vehicles don't even really have throttles anymore...  And, obviously diesels have never had throttles.  The pedal previously fed more fuel to the engine (and practically all of those are are completely electronic systems now too).  I believe brakes and steering are the two main systems that the DOT requires a physical connection.

[Soup DeVille]

The throttle/gas pedal for instance?...though I'm not sure if 'throttle' is appropriate anymore.

Throttle position request switch might be more appropriate for a drive-by wire system.; although "acceleration request encoder" might be more appropriate still.

[MX793]

Throttle position request switch might be more appropriate for a drive-by wire system.; although "acceleration request encoder" might be more appropriate still.

I call it the "go pedal".

[Byteme]

How do you defend against something like that?  Obviously, Toyota couldn't.  What is more disconcerting is that it is likely true of other manufacturers.  Nobody wants to admit that the car's computer can freeze and you can't shut the car down.  They need a kill switch.

We had 'em.  They were called ignition keys.  Turn it to the first position before run and it shuts the engine down without locking the steering column.

[FoMoJo]

We had 'em.  They were called ignition keys.  Turn it to the first position before run and it shuts the engine down without locking the steering column.

I expect that the old manual key switch would be a bit more intuitive and comforting to kill the engine if your car is uncontrollably accelerating, due to a frozen computer, towards a brick wall, as opposed to pushing a button and waiting 3 seconds to see if the button still worked.  It might be the longest 3 seconds of your life.

[GoCougs]

Ford closing in on Toyota as favorite of buyers

C'mon, man. That is fanboyism run crazy.

[GoCougs]

http://www.edn.com/design/automotive/4423428/Toyota-s-killer-firmware--Bad-design-and-its-consequences

That is awful if I'm honest. That's a code "review" not a system analysis. Without testing the hypothesis of "bad" code to get the desired effect no conclusion can be drawn. Firms vastly larger and more capable have researched the issue and have found no conclusive evidence.

No code is perfect but that doesn't mean it's bad. Even outside all that it's relatively easy to build a fault-proof control system; software doesn't control hardware first and foremost - a system does.

[FoMoJo]

That is awful if I'm honest. That's a code "review" not a system analysis. Without testing the hypothesis of "bad" code to get the desired effect no conclusion can be drawn. Firms vastly larger and more capable have researched the issue and have found no conclusive evidence.

No code is perfect but that doesn't mean it's bad. Even outside all that it's relatively easy to build a fault-proof control system; software doesn't control hardware first and foremost - a system does.

The article is very clear.  Full of bugs.

[Laconian]

Your stack is 94% full, and then decide to invoke an unbounded number of function calls on top of that...

That's awful.

[GoCougs]

Toyota turned the code over to at least NASA and Carnegie Mellon (I think on the latter) and they had their way with it with 1000x the intellectual juice of a "consultant" and concluded NPF. But again, that's a red herring. Code is only part of the system. There are HW backups too, and it's EE201 easy to do so (parallel sensors, circuits, processors). Code alone (nor HW alone) can't make the system crash to cause these problems.

Of all the material Toyota complaints - where there was a crash - NHTSA data popped open the data recorders - every single one showed throttle depressed at time of crash (save for some that concluded the throttle got jammed by a floor mat or shoe or w/e).

Someone (Popular Mechanics?) punished the Jesus out of the system (static, EMF, big voltages, etc.) and could only make it "fail" by wiring into the sensors and potentiometers on both the pedal on the floor and the TPS on the engine and with some sort of signal generator and precise triggering of signals - a processes that is statistically unlikely to happen over the course of billions/trillions of years.

[Xer0]

If Ford Sync continues to be as crappy as it is now and EcoBoost isn't putting up its mpg numbers, these new customers will leave it just as quickly as they came back. 

[FoMoJo]

Toyota turned the code over to at least NASA and Carnegie Mellon (I think on the latter) and they had their way with it with 1000x the intellectual juice of a "consultant" and concluded NPF. But again, that's a red herring. Code is only part of the system. There are HW backups too, and it's EE201 easy to do so (parallel sensors, circuits, processors). Code alone (nor HW alone) can't make the system crash to cause these problems.

Of all the material Toyota complaints - where there was a crash - NHTSA data popped open the data recorders - every single one showed throttle depressed at time of crash (save for some that concluded the throttle got jammed by a floor mat or shoe or w/e).

Someone (Popular Mechanics?) punished the Jesus out of the system (static, EMF, big voltages, etc.) and could only make it "fail" by wiring into the sensors and potentiometers on both the pedal on the floor and the TPS on the engine and with some sort of signal generator and precise triggering of signals - a processes that is statistically unlikely to happen over the course of billions/trillions of years.

A telling statement from the article...Toyota claimed the 2005 Camry's main CPU had error detecting and correcting (EDAC) RAM. It didn't.

[FoMoJo]

If Ford Sync continues to be as crappy as it is now and EcoBoost isn't putting up its mpg numbers, these new customers will leave it just as quickly as they came back. 

We know that will not be the case.

[ifcar]

A telling statement from the article...Toyota claimed the 2005 Camry's main CPU had error detecting and correcting (EDAC) RAM. It didn't.

The article simply quotes everything that the plaintiff said about Toyota. That is hardly unbiased proven fact. Maybe it's true, maybe it's distortions, maybe it's simply inaccurate -- who knows.

[FoMoJo]

The article simply quotes everything that the plaintiff said about Toyota. That is hardly unbiased proven fact. Maybe it's true, maybe it's distortions, maybe it's simply inaccurate -- who knows.

Reviewing code and hardware functionality is not a theoretical exercise.  It is not like pondering about the big bang theory; though it may seem like it to some.  It either works under all circumstances or it has holes in it.  This examination by an objective expert in the field has, obviously, found holes.

[ifcar]

Reviewing code and hardware functionality is not a theoretical exercise.  It is not like pondering about the big bang theory; though it may seem like it to some.  It either works under all circumstances or it has holes in it.  This examination by an objective expert in the field has, obviously, found holes.

It is not an objective expert. It is an expert hired by someone trying to get money from Toyota, and presenting its findings in a way that would look persuasive to a non-expert.

[FoMoJo]

It is not an objective expert. It is an expert hired by someone trying to get money from Toyota, and presenting its findings in a way that would look persuasive to a non-expert.

You think the information provided was false?

[MrH]

[FoMoJo]

Another headache?