idiot drivers

There are lots of stupid drivers out there. Locally, the three main classes of stupidity seem to be snow ("what's this stuff on the ground and why am I skidding?"), tunnels ("the mountain might fall; better slow down"), and turn signals ("I have a finite number of blinks in my life and I'm not wasting any on you"). This morning I encountered a different sort; the loser behind me didn't seem to understand that a school bus with flashing red lights trumps a green light. Sheesh. We really didn't need close to a minute of uninterrupted horn blast... (So he's not only stupid, but rude.)

I was tempted to put the car in park, saunter over to him, and explain in words of few syllables about big yellow vehicles, flashing red lights, and Pennsylvania laws on the subject -- with the goal of delaying him well past the point when the bus would have -- but I decided against.

lunchtime musings

It's really sunny out today. Painfully bright, to me. (My threshold is pretty low.) I'd really like someone to invent user-adjustable sunglasses. I don't mean photogray, where you can't control how dark or light they are, or clip-on sunglasses, which only have two settings (on and off); I want to be able to turn a dial and get more or less tint in real time. Amber, as long as I'm dreaming.

There appears to be no good protocol when the person at the head of the checkout line has a problem (like a price dispute). The default seems to be to hold up the line while dealing with it; this is especially frustrating when it's the express lane and all you want to do is pay for your salad and go. It appears that the Giant Eagle registers do not support a "buffer" function that would let them ring up the next person while someone checks the price on the shelf. They could tell the person to pay for the food and then go to the customer-service desk to discuss it and maybe get a refund, but that's also unsatisfactory (and probably has him standing in line a second time when all he wanted to do was pay for *his* food and get out).

Why is the week-to-week variance in sunset times (and presumably sunrise as well) highest near equinoxes and lowest near solstices? There's one week in December (and another in June, to match) where it doesn't really budge for several days. Currently, we're getting variances of 10-12 minutes (this is in Pittsburgh; I know lattitude matters). Why isn't it consistent year-round? I assume it has something to do with the discrepancy between local noon and true noon, just as the solstice problem [*] does, but I'm having trouble wrapping my head around the geometry.

[*] The winter solstice is not the day with the latest sunrise and the earliest sunset. Here in Pittsburgh, and I presume all points north of the Tropic of Cancer, the events occur in the following order (~2 weeks apart here): earliest sunset, solstice, latest sunrise. There's a corresponding phenomenon in June, of course.

automatic design review

Another developer has joined the ranks of those who have learned that asking me to document something will expose you to a design review -- not because that's my explicit goal, but because I ask the kinds of user-oriented questions that may or may not have come up before but probably didn't get followed through. The result is generally a cleaner API that is more coherently-explained.

None of these developers have complained about this, by the way. They seem to appreciate it (and have been known to compete for my time), and I'm glad to be working with people who can appreciate it. (I, in turn, understand the concept of something that's sub-optimal but not going to get fixed. My job is mostly to ask the questions...)

Apparently the developers here have a shorthand for this phenomenon now: they call it the "Monica effect". :-)