Are EEs overconfident?
Last week I quoted Holman Jenkins in the Wall Street Journal on the possibility that electronics and software problems contributed to Toyota’s current problems. Today in a WSJ letter to the editor, reader Ken Cowans weighs in on “the conflict between electronic engineers and their mechanical counterparts,” which may contribute to system-reliability problems. He writes, “We mechanical types have had an inferiority complex since sometime around the 1960s, when electronics started to dominate. The other side has developed a corresponding case of superiority, based on the abilities of solid-state circuits.” He says he’s seen overconfident electronic engineers reject mechanical backup system, only to implement one later because of “unhappy necessity.”
On the same letters page, reader Mike Cowan, who worked for McDonnell Aircraft, offers this observation: “The software program flow of the flight control computer of the F/A-18 branched once for every five instructions-an enormous number of possible paths to be tested…. We limited the F/A-18 program to 100 kilobytes.” He questions how thoroughly today’s multi-megabyte programs are tested. And reader James Tomlinson questions the appropriateness of a “pure ‘fly-by’wire’ braking* system,” suggesting that a break pedal at full travel should directly actuate a master cylinder.
But getting back to Ken Cowans’s point: are electronics engineers overconfident, and if so does that overconfidence lead to quality problems?
*Update: Commenters noted the incorrect use of “breaking” in the original post, suggesting that I’m overconfident in the spell checker (true) or that it was a Freudian slip (I’ll leave that to the psychologists).
Les commented:
I am confident in my designs, if given the time to look at all the possibilities. I am not confident in components in my designs, which could fail at a moment's notice. Braking backup should not include "stick foot out the door to stop the car."
Steven J. Ackerman commented:
When the engineers designed it, it was triply redundant. When the MBAs got through with it it was triply dependent. Death by MBA !
arclight commented:
All: Time to separate issues again.
The political questions, statements, declarations, aren't relevant to the asked question: "Are EEs overconfident?". They are relevant in other areas, but let's answer the asked question.
Are they overconfident? No more than any other discipline:
1. It varies with experience: The youngsters are more prone to be, because they haven't been around as long. That puts the burden on us "more experienced" folks to guide them, and teach them to always "ask the next question".
2. It varies with time: All of us at one time or another have assumed something we shouldn't, or done a less-than-complete job on something because we were in a hurry, or just plain been blind. It always seems to come back to haunt us. I suspect every one of us has those jobs that we really, really wish we had either never started or had conducted differently.
3. It varies with conviction: The best engineers, in my opinion, are those that can't NOT be an engineer. They aren't really in it for the $$$. They are more in it because they love what they do. That inspires them to learn more, retain more, and dig deeper. I've been doing this now for 32 years, and I still enjoy creating and seeing things come together to work for someone.
4. It varies with technology: One thing that I think is causing us difficulty is the ever-expanding depth of the technology we work with, and the increasing levels of abstraction between the engineer's mind and the actual hardware. I fear that too many young EEs really don't have a good grasp of things like interrupt / stack management, edges on digital circuits, etc., because they are too far away from the hardware. I could certainly be wrong about this, though.
Allan Campbell commented:
It is practically impossible to test every 'corner' case in a large software system. So it is wise to allow for fail-safe operation. And also to allow for error margins. In the human body a typical error margin is a factor of three, for example in the human liver. Do engineers design in a safety margin of three, in raw processing power for example, to their electronic systems? I don't think so, it is not usual. But it could be wise!
Andy T commented:
It's not the engineers, it's the LAWYERS that are to blame. My 2006 Trailblazer actually has throttle by wire, traction and stability control - it is not acknowledged as being preset ANYWHERE in the manuals or brochures and the model actually got slammed in the press for its lack of such features.
Roger K commented:
There is nothing wrong with "fly-by-wire" designs. It saves on weight and thus gas. The culprit is usually the lack of covering all the possibilities (e.g., answering all the "what if") in all designs.
Policebox commented:
As noted above, some are overconfident and others aren't. I often am annoyed by the number of engineers and programmers who have never heard of or don't remember what calculating a closure is. Especially in safety critical systems, this should be a part of the engineering analysis. You can't do it if you don't know it! For those of you who don't know what that is, it is a process of expanding every known state with every possible successor, until no new states appear. It can be very laborious if there are a lot of states and a lot of possible transitions (changing inputs, that is). But the alternative is to have transitions that lead to unknown states! That isn't safe. This, of course, is a major reason to keep safety critical systems simple. If they only have a few states and a few transitions, this analysis can be done and you can be confident in the result.
By the way, I take offense at the jerks who insert irrelavancies, like political statements into these discussions. They are just wasting our time by displaying their inability to reason properly.
rf_austin commented:
Speaking of "fly by wire", even mechanical systems fail in spectacular ways. Think Flight 232 in 1989, where the no. 2 engine failed in such a way to take out a triply redundant hydraulic flight control system. No one thought that was possible, or if possible, so remote as to not require safeguards. To my knowledge, no one has ever given a satisfactory answer as to why a rear left tire blowout by a Firestone radial will cause a Ford Explorer to flip over. Not every failure mechanism can be accounted for in complex systems... and these are purely mechanical failures. Now add code. I don't think it is over confidence. I think it is just the nature of complex engineered systems.
an EE commented:
I have driven in several cars that have had braking failures due to faults in the hydromechanical systems (master cylinder failure). No failures yet with the electronics. The fact is that Prius uses the by-wire system because it is the only practical way to implement the functionality they wanted. I never recommend using sofware or electronics unless absolutely necessary - good engineers are lazy. Certainly it is much harder to ensure that these complex systems are safe. Still wait to see the root cause, it may not be megabytes of code - probably a mechanical failure and failure of he system to recognize and react in a safe way. All safety reviews must involve the "over confident" EEs, MEs and others.
Off-Earth commented:
Right you are, Bear. I'll add an unpopular comment (not the first). Many engineers today, are here - not because they should be, but because it was (and is still touted as being) "where the money is". The reward for doing something for money is money. The reward for doing it for love is pride.
Yawn commented:
Not overconfident - but lacking awareness of all possible operational regimes for a new design which results in the oft-claimed "software bug" - - a copout for "bad design" - - - which most likely results from a push to market by marketeers who accepted full recognition for a "successful" deployment but no responsibility for ensuing product liability claims.
bear commented:
I am sure some are overconfident, as they are in any profession. However, in my experience it is the management which rushes things through and won't make the funds and other resources available to do a thorough testing job. I have been an electronics engineer for over 45 years and would not trust my life to a "drive-by-wire" car. I will accept power assist, but not a purely electronic/electrical solution. There has to be a mechanical override. I know all modern airplanes have "fly-by-wire", but their redundancy far exceeds anything you'll ever find in cars.
Off-Earth commented:
The products on the market were put there by a committee of businessmen. Putting a mechanical computer in a system where an electronic computer isn't fast enough won't happen because marketing will argue that the facts are backward. I believe that it is the world outside of engineering that brings the wrong solution to market in most cases
Widar commented:
Some electrical engineers are overconfident and others are not but it is eventually the responsibility of management to decide which ones will implement a safety critical system. Hopefully the correct decision will be made.
MikeT commented:
No, if anything goes wrong it's obviously Obama/unions/communism/socialism to blame, not engineers or the market-driven business people who drive us.
HalfSlave commented:
I dont see any confiddence. Maybe they are less paranoid and beaten down. I'm sure Obama has a plan to "FIX" it. So much for freedom.
Dan Quixoté commented:
Yes. Some are justified in that, some aren't. Most fall into the latter category. The ones in the former category are the ones who are always pushing themselves to know more about more topics, and generate ever-higher quality work output with every new task, and making sure to take full advantage of the knowledge and experience of others. Luckily, we do have a few of those type where I work.
Spelling commented:
Break or Brake - Hmmmm! Is Rick being overconfident in his Spell Checker written by a Software Engineer for a Computer designed by an Electronic Engineer instead of using a dictionary like a Mechanical Engineer!
kd commented:
Breaking system? Is that a Freudian slip?
EE1 commented:
Good question for generating web site traffic. Overly general for a succinct answer.
When safety if concerned there should always be backup systems and quality should be very high. Unfortunately, Obama and his comrades on the left in the unions are rich money lovers who care nothing about safety. If you remember only one thing, remember this - with the democrats, everything is political. With the democrats everything is political, from the Haitian earthquake to brakes on a car.
The recent focus on brake by wire or steer by wire are motivated by the communists in the unions and their proxys in government to secure more money to win political campaigns. It is misdirection as well, to take the focus off the failure of communism/socialism here or anywhere else to increase the quality of life.
Irony commented:
Absolutely
