July 22, 2004

When black boxes break

I experienced a black box failure today. The IC Igniter on my 1985 Kawaski GPz 750 is going flaky, and there's nothing my mechanic can do about it. My motorcycle has been misbehaving for a while, generating enough buildup on the #1 sparkplug (it's a 4 cylinder engine) that a new plug gets as much carbon buildup in 5 minutes as it would in 5000 miles on a normally-working cylinder. This causes it to backfire and, eventually, makes the plug--and the cylinder--go dead, so I ride around on a bike that's weak, shakes and sounds like a truck.

At first my mechanic thought it was a problem with the plugs, so he had me put in new plugs. Then he thought maybe it was dirt in the #1 carb (the bike also has four of these, one per cylinder), so I first flushed it out with a solvent. Then we thought it was a short, but there seemed to be more than enough current going to the plug.

It turns out that it's the only piece of computer equipment on the bike. My bike is from the very first generation of engines that had any kind of computer equipment in them. In the early 80s, car and motorcycle companies realized that they could get much better control over spark timing by using electronics, rather than the old mechanical distributor. So the distributor went away and was replaced by the unsexily-named "ignition module," a cluster of relatively simple electronics that monitors some sensors in the engine and processes that information to adjust the spark timing. The distributor was soon followed into extinction by the carburetor (whose job it is to mix air and fuel), which was replaced by electronic fuel injection.

The benefits of this way of doing things made cars in the last 20 years much more efficient, reliable, predictable and durable. All great stuff, except when something goes wrong. Now, when there's a problem, there's no way fudge, tweak or adjust around it. Digital technology is binary: it works, or it doesn't (this isn't quite true: you can hack any digital technology if you know what you're doing, but you have to know a whole host of new information to hack the electronics embedded in solid blocks of epoxy, which is how ignition modules are made).

One of the complaints that people have when they criticize the inclusion of information processing into everyday technology is this opacity of operation. Well, I experienced this first hand today, and, as someone who evangelizes for the introduction of information processing into everyday objects, I had to deal with an everyday object whose information processing was faulty. I tried to evaluate the effects the opacity had on my life: would it have been better if the bike, otherwise completely mechanical, did not have this electronic part to fail? My opinion: the module is worth it. My bike ran for almost 20 years with no ignition problems. Having all 4 carburetors adjusted cost me $400 earlier this year, and that's something that has to be done every 5 years. A new ignition module will cost me under $100 and the installation process will consist of popping my bike seat off, unplugging the old black box and plugging in the new one. I say it's a win for the black box.

Posted by mikek at July 22, 2004 04:03 PM | TrackBack

Granted, this philosophy doesn't do anything for the concept of repairable, rather than replaceable components, which is an issue.

Today's Wall Street Journal has an article that talks about how cars, with all of their expensive safety features, are becoming totaled more and more often. Here's a quote:

"In May, for instance, a 2002 Saturn sedan arrived at one of the franchises in Mainville, Ohio, after a low-speed crash n which the car was driven off the road. The exterior looked fine, but the bump set off the airbags, damaging the dashboard, steering wheel, cruise control and even the sunroof. What would have been a $5000 suspension repair was pushed up to $9200 and the car was totaled."

...and even $5000 for a suspension repair seems high to me, but the economics of car prices and repair are a different discussion. What's salient is that the more the parts of a thing are disposable, the more economics will weigh on the whole thing being disposable, whether that's a computer, a car, an appliance or, even, a building.

Posted by: Mike at July 22, 2004 05:15 PM

hi mike,

two questions and a thought:

question one: imagine the first cars and bikes. were they transparent to their users?
pre-question: isn't transparency of technology something that is achieved through time and cultural assimilation of it?

question two: let's go before the car and the bike. is a horse transparent technology? don't i need a vet to "hack" it if something malfunctions?
comment: i do take into account that horses and animals in general are very smart systems that can, up to some extent, self-maintain themselves in good conditions.

thought: maybe in about 1000 years digital technology will be transparent to us as it is mechanical tecnology nowadays, after so many years of learning it at school, and dealing with it in a daily basis and ubiquitous way.
comment: not to be ridiculous, but to build upon an image, think of "reading the matrix", hehe.

Posted by: victor zambrano at July 23, 2004 12:55 PM

Question 1: I believe that relatively simple mechanical devices, as bicycles certainly are transparent to their users. As the hotrod movement showed, teenagers and people with no specialized education COULD figure out how they worked at a detailed level. This is much harder now with black-box electronics, so I think a line has been crossed and the AVERAGE level of understanding today is much lower than during the car era.

Question 2: People had a profoundly different relationship to the natural world at the point that horses were "technology" than to today's manmade technology. Our understanding of the natural and animal world is more based on psychology than our understanding of the manmade world, which is based on our gut-level understanding of physics. A point that I've discussed before is that as our physical understanding of the manmade world becomes insufficient, we begin to use psychological models to explain how things work. This is the basis of animism.

And, in terms of your final thought, I actually think that the complexity of technology has passed the capabilities of our minds to comprehend it on a literal level. I think that our relationship with our technology will become more and more metaphorical, psychological. Our comfort level may be much higher, but I don't see a much greater understanding of how things actually work, merely more elaborate mental models.

Posted by: Mike at July 23, 2004 01:13 PM

Hello folks nice blog youre running

Posted by: lolita at January 19, 2005 04:37 PM
Post a comment

Remember personal info?