Toyota Corolla Usability Fail

I think you can imagine my surprise when, upon approaching my '02 Corolla the other day, I noticed that most of my left headlight assembly, including the turn signal, where totally smashed. After a bit of an investigation, including security at the firm I work at, it became apparent I wasn't going to find out who did it. While this sucks hard, it isn't the really interesting thing about this story.

As I got back into the (thankfully) still drivable car and headed home, I noticed that the left turn signal indicator was blinking a bit more rapidly than the right. I then realized that this had been happening for at least a week. Apparently, the slightly faster blink rate on the dashboard display indicates that the turn signal is out. At first glance, one might think, "How nice of Toyota to include a warning to tell me that my turn signal is broken!" If we think a bit more deeply though, it is quite apparent that the signal they chose will fail miserably in achieving its goal for many people.

Any well designed system must provide both feedback for user actions and visibility into the present state of the system. One way to accomplish this is to exploit natural mappings. For instance, the blinking turn signal on a car dashboard maps very well to the operation of the car's actual turn signal. When we see this, we can safely assume that the signal is operational. In my case, it was not at all apparent that a slightly more rapid blink mapped to the real signal being broken. If anything, I would simply assume that the real signal is blinking a bit faster.

What puzzles me about this situation is that Toyota made probably the worst design choice possible given the obvious intent. Imagine that instead of a slightly more rapid blink, they would have chosen not to have the dashboard blink at all when the light is out. This would lead a reasonable user to conclude that, since the blinking no longer occurs, either the dashboard or the signal itself is malfunctioning. Even a solid red light (as opposed to the blinking green) would probably do the trick. Either of these solutions would result in the user/driver taking a quick look at the actual turn signal.

In my case, I may have actually been able to figure out who/what trashed the front of my car.

A Value Added Approach to Teacher Performance

My wife, Holly, spent 4 years as a public high school English teacher in California.  From time to time, she would tell me about the "evaluations" she had coming up, which would consist of a senior teacher or administrator (usually her principal) sitting in on a lesson and giving her feedback about both the content of the lesson and some peripheral items, like how she managed her classroom.  While I don't doubt the necessity of these sorts of evaluations, especially for teachers in their first few years on the job, I couldn't help but wonder why there weren't any good, objective measures of teacher performance in place.

Every year, elementary, junior high, and high school students take standardized tests across a broad range of subjects.  Most importantly, they are tested against state-wide benchmarks for basic quantitative and language skills.  Students are graded on a relative basis, so a student placed in the 50th percentile is approximately average.  Changes in these percentile scores over time could provide and objective way to measure teacher performance.

For instance, let us imagine that an English teacher begins the year with a class of 20 freshman who scored, on average, in the 70th percentile on the previous year's standardized English test.  The material on the new year's test will be more advanced, but the average teacher should be able to maintain student scores in the 70th percentile.  A less talented teacher will cause those scores to drop, while a more talented teacher will cause those scores to rise.

While this idea alone could form the core of an objective, fully automated way to track teacher performance, there are still some rough edges that would have to be ironed out.  The most obvious is that is is likely much more difficult to help a student from the 98th to the 99th percentile than it is to help a student from the 50th to the 51st.  If teachers were evaluated solely on the absolute improvement they generated, teachers who inherited already stellar performers would be at a severe disadvantage in relation to those inheriting mediocre performers.

Fortunately, a basic solution to this problem is not too difficult to imagine.  All that would be necessary is to create a scale defining equivalent values for improvements from certain levels.  For instance, it may be determined that moving a student from the 70th to the 75th percentile is equivalent in value to moving a student from the 95th to the 96th.

Ultimately, it may not even be necessary to solve the problem.  It may be that some teachers work best with students in certain percentile brackets.  A performance evaluation system like the one I've described would provide the information necessary to identify the teachers and instructors that are best equipped to handle students at various aptitude levels.  I expect parents would find this information an invaluable resource as they work to provide the best education for their children.

Wheeling the Gurney Over a Cliff

The recent passage of health care reform by the federal government has been a topic of great contention in every corner of this nation, and it will continue to be so as dozens of states line up in opposition to the bill on constitutional and other grounds.  However, there seems to be only a feeble attempt in all of this noisy discussion to answer the fundamental question that underlies the debate: "Is access to medical care a fundamental right?"  Reason, a tool foreign to and inconsistently applied by most of our leaders, suggests that it is not.

This case suggests a proof by contradiction.  First, we will assume that no fundamental right may infringe upon any other fundamental right.  (This is necessary for the set of fundamental rights to be intelligible and consistent.)  Let us then assume that access to medical care is a fundamental right.  Also, we acknowledge that one of the most fundamental human rights is the right to freely dispose of one's abilities and resources.  If it can be shown that the right to medical care conflicts with this right, we must conclude that either one or both of them are not, in fact, rights at all.

If every human being is entitled to medical care, then it will be necessary for some entity to provide it.  In a free society, health care is provided in a marketplace where all transactions occur with the consent of all relevant parties.  In the case where no voluntary arrangement can be reached, the only way for the transaction to occur is through coercion by some outside authority.  It is in this way that the right to medical care obliterates the right of some individual to his or her own abilities or resources.  Whether we consider the surgeon who is forced to take less than market price for his services or the taxpayer who is forced to pay for the medical care of others, the right to medical care is entirely incompatible with individual liberty.

We are then forced to chose which of our supposed rights are valid.  Unless we wish to sacrifice liberty for the promise of care from a bankrupt bureaucracy and essential freedom for secondary want, we must dismiss the notion that medical care even remotely resembles something we would call a fundamental human right.  This moral reality makes possible a principled opposition to all forms of socialized medicine.

A friend of mine recently reminded me of Thomas Jefferson's advice on the subject.  He said, "I predict future happiness for Americans if they can prevent the government from wasting the labors of the people under the pretense of taking care of them."  Free men and women everywhere would do well to heed our third president's wisdom.

Census Time

The official 2010 Census form arrived today.  Here are some excerpts:

"Census results are used to decide the number of representative each state has in the the U.S. Congress.  The amount of government money your neighborhood receives also depends on these answers.  That money is used for services for children and the elderly, roads, and many other local needs."

"8. Is Person 1 of Hispanic, Latino, or Spanish origin?"

"9. What is person 1's race?"

It is entirely possible that our government thought it would be a good idea to waste space and ink on these two questions and their response fields, but it isn't probable. So, we are left to deduce that the questions have some bearing on the items mentioned in the first excerpt.

One possibility is that race has some role to play in the creation of congressional districts.  While I cannot say in certain terms whether or not this is a good thing, districting along such lines seems arbitrary at best.  The strategy may, in fact, be divisive and inflammatory.

The other possibility is that race somehow affects the amount of "government money" received by certain geographic regions.  This quite clearly crosses the line between arbitrary and discriminatory.  (It bears mentioning here that there is no such thing as "government money".  Governments do not create wealth - they merely confiscate it from citizens and redistribute it to whomever they wish.)

My essential problem with the census is that it seems like a thinly veiled attempt by politicians to gather information about manipulable voting constituencies.  (If I wanted to buy votes, I would surely attempt to have the voters pay for a census.)

If governments are to build and maintain roads, then they might need to collect information on vehicle usage patterns.  If they are to maintain parks and utilities, they might need to collect information on population densities.  I can think of no legitimate need for governments to collect information pertaining to race.

A Delightful Drinking Game

Last weekend, after finishing an invigorating game of Bananagrams with my wife and a couple of our friends at my early birthday party, we struggled to think of the perfect drinking game.  Clinging to the crosswordy awesomeness of my newly discovered favorite game, I suggested that we place a shot in the middle of the table and each take 10 tiles from the bag.  After someone started the game with the familiar "split", the last of us to make a word five letters or longer with our tiles would have to take a drink.

As you would expect, we all agreed that this was the greatest drinking game ever in the history of drinking games.  Although there is no way I was the first person to combine drinking and Bananagrams, I'm writing this post in the hope that someone, someday, will find this post and go on to drinking game glory.