My fight against the check engine light still goes on, though, like an uncle with an extensive collection of pornography, I'm sure many of you were hoping I'd just keep it to myself. Well, like your pervy uncle, I won't. I can't. And, I'm not alone.
Among the many emails I received, most suggesting I speak with my clergyperson, I received an email from Richard Stallman, the founder of the free software movement and the man who started the GNU/Linux operating system. Here's what he had to say.
REMINDER: Go sign our White House petition to ban the check engine light! Do it now — for your mom!
Dr.Stallman's ideas are interesting, though I don't necessarily think they'll change anyone's mind, and in some ways, they represent an engaging extreme; where I was thinking of providing more information for laypeople to be better informed about their vehicles, Stallman opens the possibility of an even deeper sort of automotive interaction and customization, where examining and altering a car's entire engine management computer system is possible.
I'm not certain most people will be interested in that level of involvement, but for niche groups and hardcore enthusiasts it would prove very appealing.
Here's what he had to say, in italics, and some questions from me, in bold:
I agree with you about the "Check engine" light, but that is a symptom of a broader and deeper problem: the owners of cars do not control, what the car's computers do. These computers are running _proprietary_ software, software that controls its users.
If ABS brakes are controlled by the car's computer, and throttle, etc., then how can people be certain a software bug won't end up causing a real-world crash or loss of car control?
I suspect people are exaggerating a spectacular-sounding tiny danger. Most of the software in cars is not crucial to being roadworthy, just as a lot of the physical structure is not crucial to being roadworthy.
But if someone changes software that has some relevance to safety, and if it causes a problem, that problem will probably show up in a not-quite-grave situation and he will stop running it. Thus, a real problem is possible but unlikely.
Many car owners modify car hardware. Some modifications might not be roadworthy, and some might make a small danger more likely. This is not considered a reason to physically prevent all modifications to cars, or forbid non-professionals from altering their cars. Why should software be treated differently?
It is impossible to be certain of that [lack of bugs]. Experience says: where there are programs, there are bugs. If you want to be sure a software bug won't cause trouble, don't use a computer.
What is the major argument for free software in an embedded system such as an engine management computer? For the general public, I suspect it will remain mostly inaccessible, though enthusiasts may be able to alter/replace the software. So why should general drivers care?
1. So anyone can maintain the engine — including your choice of mechanic.
2. To protect against abusive features (see http://www.bostonreview.net/BR33.2/stallman.php). In this case,hiding information from you. With free software in that computer, people would make programs to give them more info than the Check Engine light gives. And you could use them even if you don't modify your car in any way.
Maybe there are no actively abusive features in engine management computers, but there are actively abusive computer features in cars. GM cars have cell phones that tell Big Brother where the car is at all times.
I'm not exactly sure if I agree with all of what Stallman suggests , and I'm not certain of his experience with cars, but it's absolutely interesting to hear these opinions from the perspective of the man behind the open source movement.