Slashdot Mirror


VW Fiasco Puts Ethics In Engineering Under the Spotlight, CEO Steps Down

szczys writes: By now you've heard that VW has been accused of doctoring the software in their small diesel models to sidestep emissions standards. The thing that hasn't been talked about is engineering ethics. An algorithm in the code detects when the vehicle is under test conditions and causes it to perform differently. This couldn't have been accomplished by just one person. Brian Benchoff looks at the conditions leading up to the scandal and discusses the engineering ethics involved. Automotive engineers are held to a higher standard because mistakes and cut corners can kill people. This kind of suspected deceit goes well beyond concerns of environmental damage. Willing ethics violations challenge our trust of the engineering as a whole. Volkswagen‘s chief executive Martin Winterkorn has announced he is stepping down.

3 of 569 comments (clear)

  1. Whistleblowing by Dan+East · · Score: 5, Interesting

    One thing is for certain. No whistles were blown. Which is pretty impressive considering how long this has been going on and the extent of who all must have been in the know. I don't think this falls into the category of one software developer tweaking some parameters. I mean the engine was designed without a urea injection system in the first place, which is pretty much necessary to make a diesel engine conform to emissions standards that strict. So it sure leans towards the falsification pathway going way, way back.

    --
    Better known as 318230.
  2. Engineers did the deed by sjbe · · Score: 5, Interesting

    This was a financial decision, pure and simple.

    The engineers did the dirty deed. It wasn't ONLY an engineering decision but the engineers do not escape culpability here. Management cannot make this happen without the cooperation of engineering. I don't doubt for a moment that this was ordered by an executive somewhere but "just following orders" isn't a valid defense. The Germans of all people ought to know that by now.

    Someone in a suit decided it would be more profitable to hide non-compliance, rather than spend the resourcing fixing the problem *with* proper engineering.

    And someone with a pocket protector decided it was ok if they committed fraud. There will be plenty of people with their hand in this cookie jar. The real question will be how high up the food chain this reaches.

  3. Re:Speaking as an engineer... by Solandri · · Score: 5, Interesting

    It seems to me that at least some of this finger pointing should go towards the idiots who created the circumstance where the item under test was informed it was under test. That presumes an atmosphere of trust that the very idea of "testing for compliance" does not, and should not, incorporate.

    The car needs to know it's under a test. Emissions tests are usually run on a dynamometer. If you don't tell the car it's being tested, its traction control / anti-skid system will go nuts trying to compensate for the drive wheels spinning at full speed while the other wheels are stationary.

    The conditions which allowed this to happen go far, far beyond VW. A test by definition simulates the actual driving conditions. The cheating was detected by measuring emissions under actual driving conditions. That raises the question - why not just measure emissions under actual conditions? See, pulling each car off the road and testing it only makes sense when a large number of cars are not in compliance or in borderline compliance (i.e. might drift out of compliance before the next test). If a test costs $45 and 90% of cars are in compliance with emissions standards, you're paying $400 to detect each car out of compliance. And the test is worth it.

    Now what happens when 99.9% of cars are in compliance? You're now paying $40,000 to detect each car out of compliance. At that point (actually long before it) the testing isn't cost-effective anymore. California reached this threshold where the testing was no longer worth it in the early 1990s. Most cars were in compliance, and most of the air pollution was caused by about 1 in 1000 cars (mostly older models) which were spewing out hundreds or thousands of times more emissions than a compliant car.

    The companies which make the emissions testing equipment suggested a much more elegant and cost-effective solution. Stop testing each car every year. Put the emissions measuring equipment at various chokepoints on the road like free off-ramps. The equipment would then sniff the air as each car drove by, and when it detected an excessive amount of emissions it would snap a picture of the violating car's license plate. If a certain set of plates was flagged by multiple measuring stations, the State could then send the owner of that car a letter requiring its emissions be tested.

    Sounds great! It would've caught the cheating VW cars immediately. So why didn't it happen? The emissions testing itself had become a billion dollar industry. The gas stations and auto mechanics lobbied heavily to keep the mandatory testing in place. For them, a billion dollars a year were on the line. The companies making the detection equipment only stood to make a few tens of millions of dollars selling it to the state. You can guess which side won. So we ended up with testing which wastes money and isn't as effective at detecting cheating as other solutions.