A Brief History of Reliability
Abstract
Chris and Adam discuss a ‘brief history of reliability.’ There are many significant milestones in the history of mankind. The dawn of the iron age meant that instead of having tools that lasted a matter of days, humans could (in a very short period of time) create tools that lasted several human lifetimes. The Industrial Revolution of the 1800s taught us that we could ‘over-engineer’ a device so that it would last longer than its ‘useful life.’ There are many more different milestones and events that have defined reliability over time. And these have shaped how we deal with product and device reliability today. Learn more by listening to this podcast.
Key Points
Join Chris and Adam as they discuss how the history of humankind tends to revolve around largely ‘engineering’ events. Whether it be the iron age, bronze age, invention of the printing press or the Industrial Revolution, today’s society is largely based on huge engineering advances. And reliability engineering is a key element of each one of these leaps. You think differently about a steel axe head if it is something your son and grandson can use. You also need to make sure that you don’t design in too much margin into a system so that it stops being cost-effective – if that system will be superseded relatively quickly. In fact, you can’t talk about the history of mankind (let alone the history of engineering) without talking about reliability.
Topics include:
- How having tools and basic instruments that lasted so long that they were passed on from generation to generation meant we started to think about how to make these ‘things’ relevant over a long period of time.
- How becoming more informed on the physical and material properties of systems and devices actually allow us avoid ‘over-engineering’ … saving substantial resources when we do.
- How different cultures and normative behaviors shape our historical attitudes towards reliability.
- And finally, how all these things tangibly define what reliability engineering is today.
Enjoy an episode of Speaking of Reliability. Where you can join friends as they discuss reliability topics. Join us as we discuss topics ranging from design for reliability techniques, to field data analysis approaches.
Show Notes
Hello Chris and Adam,
Thank you for putting out another great episode to listen to. Culture has a huge influence on Reliability because of its hold on management decisions. I would like to add a point to the discussion on Toyota trying to overtake GM and then facing recalls due to “sudden over acceleration scandal”. NASA did a thorough analysis of Toyota Engine Throttle Control boards and concluded that they were not causing sudden acceleration but it was due to defective floor mat design that were pressing against the gas pedal. And in some cases a sticky gas pedal that were slow to return to normal. So, the root cause concluded from this study was floor mats and not ETC. I was surprised by this revelation provided how many Toyota-slandering articles I read. And then the famous author Malcolm Gladwell went a step further in his investigative podcast Revisionist History and exposed that the real reason these fatal accidents happened was because of people unknowingly pressing gas pedal thinking it is the brake pedal under stressful conditions. Toyota still paid billions in fines but the root cause seems to be human error. This episode made me realize the influence of psychology in root cause investigations.
https://one.nhtsa.gov/About-NHTSA/Press-Releases/NHTSA%E2%80%93NASA-Study-of-Unintended-Acceleration-in-Toyota-Vehicles
http://revisionisthistory.com/episodes/08-blame-game
Thanks,
Arun Gowtham Sampathkumar
Thanks for your response Arjun! I am glad to hear that you like the podcasts.
So let’s talk about some of the points your raise.
Malcolm Gladwell is a successful author with a journalistic background – and his podcasts are very good to listen to. They are intriguing and easy to digest. But, as with all podcasts (including ours), you need to interrogate them. And challenge the podcaster as you are challenging us.
I encourage everyone to listen to listen to Gladwell’s podcast that Arjun linked above … but only after you hear some of my thoughts below.
Gladwell is all about human behaviours. He is not an engineer. He argues that the instances of Toyota unintended acceleration are consistent with errors rooted in human psychology – not vehicle defects. And it doesn’t take long before this ceases to be a competing explanation – by the end of the podcast it become his only explanation.
Gladwell’s podcast includes the 911 call of a passenger of a Lexus being driven on a California highway. The driver was Mark Saylor – a California Highway Patrol (CHP) Officer. He couldn’t get the vehicle to slow down. And unfortunately all occupants died when the car ploughed straight through an intersection at full speed. Gladwell is convinced that Saylor panicked and simply didn’t apply the brakes after the pedal got stuck by a mat. This is plausible.
But as with everything in life, we need to talk about what is probable. The 911 call is 51 seconds long. Now think about the things you would do to try and slow your car down before getting a passenger to call 911. That list would be quite detailed. And that includes the reactions of the passengers. What would you do before you call 911? So for a period of let’s say 2 – 5 minutes (enough time for some of your senses to return after panicking), the foot never touched the brake … according to Saylor. With no physical evidence to support this.
Whenever we talk with someone, or listen to a podcast, we should understand who is talking. Is someone biased? Is someone putting forward the conclusion first, and then working out how to get there? How do you react to someone who only talks about certain aspects of a scenario? … and what if these aspects are the only ones that support that person’s argument?
To be clear, Gladwell doesn’t talk about key aspects of the Toyota unintended acceleration saga. He doesn’t even reference them. You must consider this when you weigh the merits of his (or anybody else’s) arguments.
I will list a few below (the biggest one is the last one).
Firstly, Gladwell never addresses the fact that Saylor was a CHP Officer. Someone whose day job is to calmly and skillfully control high power cars at high speeds. This was his job. He was less likely to in a panicked state, thinking he is pressing the brake when he is in fact pressing the accelerator. Gladwell actually talked about a trend (which he really doesn’t reference in a way we could then go and research it) between people like the elderly or the inexperienced and sudden unintended acceleration. He called them ‘not car guys.’ Gladwell’s argument in the case of Saylor was that this was the first time he drove that Lexus. But if he really wanted to pull at that thread, he needed to talk about how the brake and accelerator pedals in the Lexus were fundamentally different from the cars Saylor routinely drove (they aren’t … vehicle standards keep them similar for a reason). Or he should have approached Hertz or Enterprise to go through their crash data – because this is where we would see it if this phenomenon was real (by definition, all hire car drivers are ‘unfamiliar’ with the car they are driving.) He did neither, and certainly did not talk about the driving skillset Mark Saylor had. Because Mark Saylor was a ‘car guy.’
Secondly, Gladwell says the 2010 NASA report concluded that the cause of the sudden unintended acceleration was not Toyota’s Electronic Throttle Control (ETC). This is wrong. NASA couldn’t identify a defect that would see the ETC result in unintended acceleration – which is different. And they said precisely that. You can read it in the link provided by Arjun above. Ask any software engineer about reliability, and one of the many issues they face is they can’t replicate faults as software is exceedingly complex. And with unintended acceleration, even its most optimistic rates of occurrence meant it was still very rare. And (by extension) very hard to replicate. Just because they couldn’t find a defect, doesn’t mean there wasn’t one. Microsoft, Apple, Google and others can attest to this fact.
Thirdly, Toyota’s own internal documents showed that they knew something was wrong with the ETC and they lied to Congress about it. This is why they were criminally charged. Not because of hysteria. Not because of a subjective interpretation of data. But because a whistleblower leaked internal memos between Toyota executives clearly showing this.
And finally, Toyota admitted they were wrong. And paid the $ 1.2 billion fine. I don’t need to say much more than this.
Gladwell can certainly create a podcast that is easy to listen to … but why (for example) would he not talk about the fact Toyota admitted to what was clearly apparent in their own documents? Why would he not advise listeners that these documents showed that Toyota engineers and executives were talking to each other about various problems with the ETC? And why would Gladwell not even start to address their lies about this?
If I follow Gladwell’s advice and look for the simplest explanation of why he structured his podcast the way he did … it’s because it is in his best interest to. He is a behaviourist who is seeking an audience. So what could be better than promoting yourself as someone whose unique behaviourist insights can trump a DOJ investigation (and the apparently Kamikaze admissions Toyota made confirming all this to be true)?
As with everything in life, we need to process information from different sources and make our own judgments about what is right and what is wrong. The balance of probabilities is key. Am I biased? I haven’t worked for Toyota, the DOJ or any of the plaintiffs involved. It is up to you (the reader) to work this out for yourself.
But one thing that is abundantly clear. There is no hysteria. Toyota admitted it. The DOJ explained it. And a whistleblower’s documents support it. And it is wrong to say that people who look at this in a holistic way must be hysterical if they disagree with Gladwell.
So is it plausible to suggest (for example) Toyota’s first and only lie was when they admitted previously lying when they were in fact telling the truth … before writing a $ 1.2 billion check? Yes. Is it probable? Not even a little bit.
And … feel free to disagree!
Chris,
Thank you for the reply. You bring up lots of good points and I like the way you’ve structured the answer based on the points in Blame Game podcast. It is true that one has to think about the causes and make a stance without any bias. The very fact that Gladwell is trying to tell us all about the quirks of our thinking might also have clouded his thinking. As you request the audience, we have to take these with a pinch of salt. In situations like these it is better to rely on the data as it is unbiased. If we could find data from Saylor’s crashed car and see that the brake pedal values are zero during the unfortunate event then it is believable that he did not press the brake pedal. This begets another question: what if the data logger itself failed and did not capture the pressing of brake? Ummm..
I second your point that it is possible that defects might be in the software even if we couldn’t produce it in testing. The real world conditions are vast and complex and it is near impossible to recreate the exact scenario in test labs. But the head of the hammer is the point that Toyota knew something was wrong, tried to hide it, and they recalled the cars and paid fine. That something is either electronics or sticky gas pedal or floor mats. This case is very interesting I would like to read more reports and share it.
Thank you for bring this case up and encouraging an unbiased thinking.
Arun
Thanks Arjun!
The way we learn is from questioning everything. Not questioning everything to the extent we have a conspiracy theory. And we should never not question what we hear.
And we do need to credit Toyota with being a very good organization when it comes to reliability. Many organizations have tried to copy (and often fail) the ‘Toyota way.’ They are still a dominant force in reliability and quality … notwithstanding the incidents above.
Happy listening!
Chris