nico.rikken at fsfe.org
Fri Oct 9 06:57:40 UTC 2015
Hi Vitaly, Daniel and other subscribers,
I'd like to bring some nuance to this issue. I was told in college that
such practices were already going on, that was years ago, so for me it
didn't come as a surprise. I was however surprised by the strong
reaction, which in hindsight makes sense.
Car emissions requirements are very strict, and it is a real challenge
pushing technology to meet these ever stricter demands. For the EU these
emissions are verified based on a predefined tests the car has to go
through. If there is plenty of headroom in the technological solution
you can just pass any test flawlessly. If this headroom is lacking and
you want to pass the test the first step is to optimize the system for
the test-curves. The next step, we've now observed, is to detect the
chain of events in the test and handle the situation differently.
Would not any modern car optimize its performance for different sets of
driving habits and environmental circumstances? Were the tests or
regulations too naive not to set (I assume) requirements for
optimizations or related software auditing?
Making the point that an open source development model would have
prevented this issue seems a bit overblown to me. Yes it would be easier
to spot, but would it? As I've read this specific issue was discussed on
a technical forum between regulators and engineers months earlier which
did not directly trigger any greater discussion. Furthermore as I've
described earlier the industry has been aware of such practices for many
years, so I guess this is just the first real case made against such
Spotting such specific code in a software project as complex as a modern
car would not be simple. But of course having the code at hand combined
with real outside-in tests would be surely help.
Don't get me wrong, this is an environmental crime which is serious and
a through investigations and regulatory adjustments are needed. But to
me this issue has more to do with regulations, requirements, company
ethics, and less to do with the licensing or development model of the
I'm more concerned about the security implications of hidden source code
(VW cars apparently had an unresolved security flaw in their lock system
) and the limited possibility of executing maintenance or
modifications due to software restrictions (John Deere). I'm not sure
how this should be regulated, but having decent laws in place to provide
freedom for car drivers and car owners would be a great step forward.
Glad to hear others' thoughts on this.
On vr, 2015-10-09 at 08:02 +0200, Daniel Pocock wrote:
> On 09/10/15 07:55, Vitaly Repin wrote:
> > Hello,
> > http://arstechnica.com/tech-policy/2015/10/volkswagen-pulls-2016-diesel-lineup-from-us-market/
> > states:
> > The VW scandal came to light when researchers from the International
> > Council on Clean Transportation and West Virginia University performed
> > all kinds of tests on VW vehicles, discovering that when the vehicles
> > were on the road, they polluted substantially more than when they were
> > being tested for pollution emissions.
> > If this software was open-source (not even free!) the process to find
> > the problem woudl be much easier and less costly. And most probably the
> > issue would not occur at the very first place - VW engineers woudl not
> > have temptation to do what they did.
> > Do you think VW case can be used as an example of the benefits which
> > free software might provide to the society?
> Did you see my blog about this?
> The key point here is that it is a conspiracy: (a) it involves
> /deliberate/ dishonesty, (b) lots of people must have known, (c)
> closed-source, non-free obfuscation was used to hide it and (d) it was
> used to deceive government(s) and now they all look stupid.
> In the law, people convicted for conspiracies are usually given bigger
> punishments than people convicted of doing something bad on their own.
> I just feel this is an opportunity too good to miss for free software
> There is a revolving door between industry and the regulatory agencies
> and they won't ever fix themselves without outside pressure.
> With consumer organizations making lawsuits, it could also be a good
> opportunity for FSFE to forge alliances with other groups and share the
> free software message more widely.
> Discussion mailing list
> Discussion at fsfeurope.org
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 213 bytes
Desc: This is a digitally signed message part
More information about the Discussion