Tuesday, March 27, 2007
Metrics can be useful; with a good set of metrics you know where you stand and how far you have to go. Metrics give you something to aim for, which is sort of the problem. It's too easy to just aim for the metric and ignore the overall purpose. Thus it is important to make sure that your goals and the metrics that support them are completely aligned. Metrics that are poorly defined can be met without actually bringing you any closer to your goal. In extreme cases, poorly defined metrics can act counter to the original goal.
For example, consider my friend Dave. Dave designs software that is intended to help get species off the "Endangered Species List." Now, as it turns out, there are two ways in which a species can be removed from that list. The first is for the animal to increase in numbers until it is no longer considered endangered. This process is what Dave's software is intended to track.
Dave, who is a very helpful person, claims that he is so dedicated to this cause that he uses his spare time to help remove species from the list. He's doing what is within his power as a private citizen, so instead of a computer he claims to use a rifle (or in the case of baby seals which may or may not even be on the list, a club). He figures it is easier to get animals off the list by pushing them out the bottom than by attempting to carry them up to the top.
Now before you start sending Dave hate mail, it is worth pointing out (for those of you who are not familiar with my friend) that Dave's stories have roughly the same relationship to truth as Penrock's. So odds are good he's not actually spending his weekends trying to tip endangered species into extinction. However, that's technically a valid response when the goal is to reduce a metric solely defined as "number of species currently considered endangered."
So, metrics are important, but it is more important to have the right metrics. Faulty metrics can be ruinous. Dave's example might be a little extreme, but there are plenty of actual cases of such dangerous metrics (the No Child Left Behind Act springs immediately to mind). This is true for personal projects and it is certainly true for public policy . Unfortunately, since the ultimate metric in public policy is the vote, we may be in trouble.
Saturday, March 17, 2007
We have a number of fascinating and useful machines in our shop. Today I wish to talk about a personal favorite: the project-ruining machine. This particular machine performs a variety of tasks but its chief purpose is to completely ruin your project at exactly the wrong time. Actually, it’s not a very difficult job. Sometimes you can even get it taken care of without a machine, but the machine adds a certain finesse. There’s something elegant and efficient about the way the project-ruining machine can turn a simple twitch of the hand into complete devastation.
When I was first introduced to the project-ruining machine, I thought “That looks remarkably like a belt sander.” I was, however, assured by my professor that while it could function as a belt sander from time to time, it was in fact a project-ruining machine. It has a remarkably intuitive interface: simply apply your project to the sander at the wrong angle (easy to do since there are so many wrong angles compared to correct angles), or fidget on your approach, or press a little too hard on one side. The only way to cause the machine to malfunction such that it fails to ruin your project is to approach it at just the right angle, without jitters, and with even pressure. Fortunately, even if you do mess things up and fail to ruin your project the first time, the project-ruining machine generally gives you plenty of chances and really, it only needs to succeed once.
If the belt sander version isn’t what you’re looking for, the project-ruining machine actually comes in a wide array of shapes. For instance, our shop has project-ruining machines that look like table saws, band saws, chop saws, orbital sanders, jig saws, milling machines, drill presses, lathes, and of course several varieties of belt sander. Once I knew what to look for I discovered that the project-ruining machine is pretty much the only machine our shop offers – one machine, nearly infinite varieties.
I have not quite got the knack of it yet. So far I have (mostly) failed to ruin my models, but I look forward to the day when I, too, get to hear that minor change in pitch, sense that brief twitch, and experience that tremendous sinking feeling in the pit of my stomach as I realize I have successfully used one of the project-ruining machines to nullify more than thirty hours of work in a single instant. What an impressive machine.
Saturday, March 10, 2007
Sunday, March 04, 2007
I drew the outline by hand, smoothed it out in Adobe Illustrator, and then rendered it in Photoshop. If any of my dear readers have something similar they need or would like to see, let me know. I need all the practice I can get.
Thursday, March 01, 2007
Also Mike has taken the liberty of evaluating our updating habits. At the moment of this writing, my ranking is "disappointing." This is not acceptable to me.
So you can thank Penrock for my renewed determination to update this blog on a regular (read: not disappointing) basis.
While my semester continues at its current pace I will not be able to produce anything too extensive, but I am determined to produce something (on the assumption that it's better than nothing).
I recognize that acquiescing to Mike's demands may send him the wrong message, but in this case his desires happen to coincide with my own (and presumably yours) so I'm going to risk it.