January 7, 2016 at 8:57 am
Excellent article and video about this, including a session by a hacker who reverse engineered an engine control unit.
The incredible smoking gun here was how EXACTLY the parameters matched the test conditions. Had they made the situation more rubbery, where it could go into 'clean' mode in a range of conditions, not just the test ones, the results would not be so damning.
http://www.thetruthaboutcars.com/2015/12/strangely-volkswagens-defeat-device-normal-mode-video/
Any time you have a make-or-break test, people are going to 'design to the test' (as we've heard about teachers in recent years). Optimizing for test conditions could de-optimize for other conditions, which is a kind of grey area. But this does not seem to be grey at all.
...
-- FORTRAN manual for Xerox Computers --
January 7, 2016 at 9:02 am
Hmm... the VW thing sounds like a case where IT people were being blamed for other people's problems. And while management may have been involved it likely wasn't a higher level manager I'd guess it was a low level IT manager type person who kept getting harassed because the software his team was developing kept failing tests that engineering said shouldn't be. And I've been there too when someone higher up says this result is supposed to return this value you return that value even if it's not right.
January 7, 2016 at 9:13 am
However it seems that even smaller efforts aren't worth making if there is a little more profit to be squeezed out. At least, that's what managers seem to think.
In this case, the cost cutting was not on the software side. They seem to have been trying to avoid expensive hardware (SCR etc) and did a lot of fancy stuff with the software to hide the hardware deficiencies.
We're seeing this more and more these days in the automobile biz. Things like transmission shift problems, runability problems, etc are being recall-patched by software which essentially modifies behavior to step around or hide the actual design failures. On some automotive fora I'm on, it's pretty easy to see that recalls are being done that sometimes disable or limit functionality rather than fix the mechanical issues that underlie the problem.
...
-- FORTRAN manual for Xerox Computers --
January 7, 2016 at 11:35 am
Management threw the engineers under the bus. The VW Bus. :hehe:
January 7, 2016 at 11:50 am
Wayne West (1/7/2016)
... When we have better solutions, or we can correct deficiencies, management doesn't want to spend the time or effort to do so. ...
Greed, pure and simple. The drive for quarterly profits instead of long-term viability. Warren Buffet had a wonderful suggestion that will never happen: a 100% tax on stock sales for C-level executives after they leave the company. After a year it's 80%, and goes down 20% (or whatever) subsequent years until it's down to the base rate. Executives would be strongly incentivized to ensure the company would endure rather than meeting Wall Street's demands for quarterly returns.
It's true of so many companies and organizations, and we see it on this site almost every day when people post of being allowed no time to get it right, that it must be got out the door and "we'll fix it later", meanwhile later never comes.
There's also the opposite viewpoint that software isn't released, it escapes and is dragged from the clutching developer's hands.
The thing that I find the most sad about the Volkswagen debacle is I toured their assembly plant in Dresden in June. Absolutely amazing place, it made me very sad when their fudging with emissions numbers was revealed.
Some of us remember Sybase, when they pushed a release too soon, with too many bugs.
And then Sybase was out of the database market.
Marketing usually makes the product decisions, takes credit for success, passes on the bad news.
The more you are prepared, the less you need it.
January 7, 2016 at 12:21 pm
ZZartin (1/7/2016)
Hmm... the VW thing sounds like a case where IT people were being blamed for other people's problems. And while management may have been involved it likely wasn't a higher level manager I'd guess it was a low level IT manager type person who kept getting harassed because the software his team was developing kept failing tests that engineering said shouldn't be. And I've been there too when someone higher up says this result is supposed to return this value you return that value even if it's not right.
http://dilbert.com/strip/2016-01-07
😀
January 7, 2016 at 12:28 pm
Mark A. Norris (1/7/2016)
ZZartin (1/7/2016)
Hmm... the VW thing sounds like a case where IT people were being blamed for other people's problems. And while management may have been involved it likely wasn't a higher level manager I'd guess it was a low level IT manager type person who kept getting harassed because the software his team was developing kept failing tests that engineering said shouldn't be. And I've been there too when someone higher up says this result is supposed to return this value you return that value even if it's not right.http://dilbert.com/strip/2016-01-07
😀
Great Dilbert cartoon!
The more you are prepared, the less you need it.
January 7, 2016 at 1:39 pm
chrisn-585491 (1/7/2016)
Management threw the engineers under the bus. The VW Bus. :hehe:
*groan*
January 8, 2016 at 3:19 am
Ok,
so let's assume for a moment that the VW management aren't throwing a couple of developers under the bus because they're a bunch of feckless cheats without the morals of a backstreet crack dealer. Let us further assume that their processes are so staggeringly inept that rogue developers can promote whatever the hell they want to the control systems that are responsible for the running of their products without any even remotely effective oversight. And that their employment processes apparently favour the employment of highly dubious characters of a criminal disposition (because this is criminal activity) for developing the critical software responsible for the safety and performance of their vehicles.
They can not, therefore, ensure that someone with a newly minted swiss bank account has not promoted code that will switch off the braking systems on all their vehicles at 5:30 PM UTC 7 months from now. Because this is - basically - what they are telling us. Really?!
If you are prepared to swallow THAT, I have an iconic bridge you may be interested in purchasing ... It would certainly be a better buy than a VW
I'm a DBA.
I'm not paid to solve problems. I'm paid to prevent them.
January 8, 2016 at 6:16 am
Something that I've seen way too often: When something goes really well, management takes all of the credit and glory at most saying that it was a "team effort" yet when something does not go well then management is quick to point out hapless individuals who were not "team players" as the source of the trouble even if the true cause was poor management decisions.
I get sick every time I see a fellow team member (non-management) do well and all management does is offer general acknowledgement ("Yeah team!", "What a great team!", etc.) however whenever someone in management gets recognized it's all about the individual. I witnessed a person in management get public recognition for a successful project without acknowledgement of the individuals who created the concept and did all the heavy lifting. All he would say was "It was a team effort".
Management takes credit for innovations while others get nothing but expectations for continued "team support". When someone points out the possible consequences of poor management decisions then s/he is not a "team player". And through all this, some folks wonder why there's such a strong division between management and non-management. Go figure!!
LinkedIn: http://www.linkedin.com/in/sqlrv
Twitter: http://www.twitter.com/sqlrv
Website: http://sqlrv.com
January 8, 2016 at 11:00 am
Eric M Russell (1/7/2016)
... I saw was of a press conference where the upper management of VW seems to indicate that the software developers on the project decided to alter the software...
There are two different engineering teams; the one that works on the components of the catalytic converter system and then the one that works on the software that interfaces between the catalytic converter and the diagnostic output. If what they're saying is that the software team unilaterally decided to implement coding that fakes the performance of the exhaust system, then that's preposterous. What could possibly be the motive, unless they were somehow inventivised by executive management on the down low?
The "team" involved to me sounds like the components team, which couldn't come up with a component design which didn't affect the performance of the car. So - the converter that met the new compliance standard must have taken away a few effective horsepowers, etc... so the car would be less performant, etc.... The fact that they managed to convince the monitoring team to go long with faking the results is something of a mystery (that's a straightforward case of misconduct).
----------------------------------------------------------------------------------
Your lack of planning does not constitute an emergency on my part...unless you're my manager...or a director and above...or a really loud-spoken end-user..All right - what was my emergency again?
January 8, 2016 at 11:10 am
Matt Miller (#4) (1/8/2016)
Eric M Russell (1/7/2016)
... I saw was of a press conference where the upper management of VW seems to indicate that the software developers on the project decided to alter the software...
There are two different engineering teams; the one that works on the components of the catalytic converter system and then the one that works on the software that interfaces between the catalytic converter and the diagnostic output. If what they're saying is that the software team unilaterally decided to implement coding that fakes the performance of the exhaust system, then that's preposterous. What could possibly be the motive, unless they were somehow inventivised by executive management on the down low?
The "team" involved to me sounds like the components team, which couldn't come up with a component design which didn't affect the performance of the car. So - the converter that met the new compliance standard must have taken away a few effective horsepowers, etc... so the car would be less performant, etc.... The fact that they managed to convince the monitoring team to go long with faking the results is something of a mystery (that's a straightforward case of misconduct).
It is a bit of a mystery when you consider how hard is normally is to get marketing, QA, and engineering to all be on the same page about something.
"Do not seek to follow in the footsteps of the wise. Instead, seek what they sought." - Matsuo Basho
January 11, 2016 at 6:16 am
Reminds me of the trouble Google got into a few years back when their street view cars were caught snooping WIFI details - wasn't that blamed on a few rogue engineers?
Given the massive benefit to the company as a whole that these 'rogue' engineers created I find it hard to believe that there wasn't at least knowledge & tacit approval from above.
I've written plenty of code 'to see if I could' but know that it's never likely to get anywhere near a prod system!
January 11, 2016 at 6:50 am
Eric M Russell (1/8/2016)
Matt Miller (#4) (1/8/2016)
Eric M Russell (1/7/2016)
... I saw was of a press conference where the upper management of VW seems to indicate that the software developers on the project decided to alter the software...
There are two different engineering teams; the one that works on the components of the catalytic converter system and then the one that works on the software that interfaces between the catalytic converter and the diagnostic output. If what they're saying is that the software team unilaterally decided to implement coding that fakes the performance of the exhaust system, then that's preposterous. What could possibly be the motive, unless they were somehow inventivised by executive management on the down low?
The "team" involved to me sounds like the components team, which couldn't come up with a component design which didn't affect the performance of the car. So - the converter that met the new compliance standard must have taken away a few effective horsepowers, etc... so the car would be less performant, etc.... The fact that they managed to convince the monitoring team to go long with faking the results is something of a mystery (that's a straightforward case of misconduct).
It is a bit of a mystery when you consider how hard is normally is to get marketing, QA, and engineering to all be on the same page about something.
It almost takes an act of congress to get them on the same page.
The whole situation reminds me of a quote I heard once:
There are two types of people in this world: Those that do the work and those that take the credit. I prefer to be in the first group - there's a lot less competition.
When things go right, top brass takes the credit. When things go wrong, it's always the fault of someone else. There's always blame to be assigned and fall guys to take the hit. It's sad, but that's the way some people work.
January 11, 2016 at 9:16 am
Ed Wagner (1/11/2016)
Eric M Russell (1/8/2016)
Matt Miller (#4) (1/8/2016)
Eric M Russell (1/7/2016)
... I saw was of a press conference where the upper management of VW seems to indicate that the software developers on the project decided to alter the software...
There are two different engineering teams; the one that works on the components of the catalytic converter system and then the one that works on the software that interfaces between the catalytic converter and the diagnostic output. If what they're saying is that the software team unilaterally decided to implement coding that fakes the performance of the exhaust system, then that's preposterous. What could possibly be the motive, unless they were somehow inventivised by executive management on the down low?
The "team" involved to me sounds like the components team, which couldn't come up with a component design which didn't affect the performance of the car. So - the converter that met the new compliance standard must have taken away a few effective horsepowers, etc... so the car would be less performant, etc.... The fact that they managed to convince the monitoring team to go long with faking the results is something of a mystery (that's a straightforward case of misconduct).
It is a bit of a mystery when you consider how hard is normally is to get marketing, QA, and engineering to all be on the same page about something.
It almost takes an act of congress to get them on the same page.
The whole situation reminds me of a quote I heard once:
There are two types of people in this world: Those that do the work and those that take the credit. I prefer to be in the first group - there's a lot less competition.
When things go right, top brass takes the credit. When things go wrong, it's always the fault of someone else. There's always blame to be assigned and fall guys to take the hit. It's sad, but that's the way some people work.
Now you know why people want to do everything via email... admissible evidence. 🙂
--Jeff Moden
Change is inevitable... Change for the better is not.
Viewing 15 posts - 16 through 30 (of 32 total)
You must be logged in to reply to this topic. Login to reply