How do auto makers deal with problems where a parts on a car vibrate noisily? In some cases, they just add some weight to the part that vibrates. This is sort of a sneak preview of a post I’m working on about good hacks versus bad hacks, and how to tell them apart. And maybe another post on what a pain it is to lay out HTML forms using semantic markup in a fashion that works in all of the popular browsers.
One of the biggest challenges for software developers, I think, is figuring out which hacks are good, and which hacks are bad. In the meantime, I put adding weight to car parts to keep them from shaking in the same category as rebooting your application server every night to dampen the effects of a memory link in your code. It may be a necessary hack for the short term, but it doesn’t really cut it as a long term solution.
Update: Be sure to read the comments. What seemed to me to be a hack is actually a pretty elegant solution.
February 22, 2008 at 2:24 pm
I think I disagree on the weights thing. It’s impossible to build a car engine without any vibration, and that problem is all about movement of masses and harmonic motions and stuff. Introducing damping masses is a natural way of reducing the effects of vibration, and I doubt it’s as ad-hoc as it might look. (Well, maybe not.)
As for HTML forms from semantic markup with cross-browser compatibility, A) been there, done that, B) have you looked at XForms?, C) I think some of the Ajax toolkits that have you building forms with Javascript (e.g. ExtJS) actually suit the forms model better than even the most structured, “semantic” HTML does.
(PS: Preview comment would be nice if you ever get around to it…)
February 23, 2008 at 12:46 pm
B) have you looked at XForms?
XForms will be a sweet, sweet addition to the toolbox, just as soon as #!$@* IE supports it out of the box (yes, there are plugins, but installing such is beyond the ability of the average schmo, so it will never happen).
February 23, 2008 at 7:06 pm
Rafe, I usually like your analogies but this one is way off. When building a system of thousands of interlocking parts that will be subject to any number of periodic forces, you’re going to get resonance. The weight-the-part solution is extremely elegant given that:
a) it doesn’t significantly affect the performance or asthetics of the car
b) it is quick, inexpensive, and a well-understood solution
c) it is permanent
d) it allows the auto manufacturer to keep costs down and focus on areas that will deliver more value to the consumer
So comparing it to the reboot-to-fix is not really accurate. In fact, this isn’t even a good example of worse-is-better. It’s just prudent and efficient design for manufacturing.
-ch