Paul Kuliniewicz has an eventful day

I upheld a streak, pounded pavement, maintained a broken-electronics invariant, made NASA help me do my homework, and found a use for my numerical methods class.

The earliest class I have this semester is at 11:30. Since I typically wake up well in advance of that, I don’t normally need my alarm. In fact, all semester, I have yet to wake up to my alarm, instead regularly getting up sometime in the 9 – 9:30 zone.

However, today (Tuesday) there was a job fair at the Purdue Research Park which started at 9:00. As I have class from 12:00 to 5:45 solid on Tuesdays, I had little choice but to go at the beginning, especially since I didn’t know how long I would need to spend there. It looked as though I had to choose between upholding the not-having-my-alarm-go-off-all-semester streak and getting a job over the summer.

Luckily, I ended up getting up ridiculously early — around 6:30. It wasn’t really my intention, but instead of falling back asleep I kept thinking of my history project and a possible programming languages project. Eventually I decided that I had laid awake in bed long enough to kill my chances of falling back asleep easily, so I just decided to get up right then. The streak continues!

The job fair went pretty well. One nice thing about it was that it was held off-campus, at the Purdue Research Park. Such a location, even though it’s only a few minutes’ drive from campus, seriously cuts down on the number of students attending. Instead of lines a dozen students long to talk with a recruiter, there’d usually be one or two tops. Getting there soon after it started also helped. All of the companies there are small firms that were created to capitalize on one piece of research coming out of Purdue or another. Getting an internship at one of them would be a nice change from working at a big company over the summer. We’ll see how that all turns out.

I don’t think I’ve mentioned my roommate’s saga with getting his G5 fixed. Over the Christmas break he got a new computer, and the first week of the semester it decided to only start up maybe once out of every twenty times. He took it in to the Apple store in Indy to get it fixed. He got it back today, about six weeks later. They couldn’t figure out what the problem was, so they finally just gave him a replacement system.

Anyway, the laws of the universe state that if one piece of consumer electronics is fixed, another must break. So, the TV decided to stop working around lunchtime. It wasn’t my fault, I swear, even though I was the one who turned it on. I put it on the History channel, as is the custom, and went into the kitchen to make a sandwich. Then the screen turned green for a couple of seconds, then went blank, then the sound cut out a few seconds later, and the TV powered itself down. That’s not normal, I thought. Turning it on again restored the sound but no image. Well, that’s not true; after a good ten minutes of running in this state, an image reappeared, but the colors were rather wonky.

Fortunately, my roommate believed me when I explained that I did not play any role in the smiting down of the TV, so tonight we went to ye local Best Buy to procure a replacement, in what may go down in history as the fastest excursion to replace a TV ever. So now we have a slightly larger TV with a better (i.e., functional) picture.

Remember how I mentioned I was thinking about my history project last night? What I’m doing is comparing various historical astronomical models to modern-day planetary models to compare how well they perform by writing a computer simulation of the various models. Naturally, I need a good source of “correct” astronomical data. And so I discovered my new favorite toy of the moment: the HORIZONS System, offered by the Solar System Dynamics Group of NASA’s JPL. The tool lets you get astronomical data for just about any body in the solar system you can think of: its coordinates, illumination, distance, and a bunch of other stuff. Data for all the planets known to the ancients (i.e., up through Saturn) are available from about 3000 BC to AD 3000. Naturally, these ephemerides are based on a model of the solar system (the documentation for the system has a nice disclaimer saying you shouldn’t use it for non-lunar spacecraft), but I figure such tables are plenty good for a “modern-day” model as far as my project is concerned.

The system has a nice, easy web interface to it, and is completely free. I figure I’ll generate tables of the information I’m interested in over a few centuries and do some interpolation to fill in the gaps. I can use that as the modern-day model and compare its predictions with models like those of Ptolemy and Copernicus. Plus, instead of using the “original” parameters used by the historical models, I’ll recompute them based on the modern-day model’s predictions, using a least-squares fit. That way I can look at the model itself and not at the (possibly imperfect) application of the models that their creators may have used. Then it’s just a matter of making the models match at time t, run them for a while, and look at by how much their predictions diverge and what things they may miss (e.g., incorrect eclipse predictions out of the historical model). Gasp, I found a useful application of the stuff we’ve been talking about the past week or so in my numerical methods class!

Anyway, playing around with HORIZONS is fun. Also, I like how old-school the SSD’s web site is. It’s like the web looked back in the early 90′s: plain white backgrounds, the occasional iconic-type graphic, “e-mail webmaster@” links at the bottom, and even telnet- and e-mail-based alternative access methods to the CGI scripts! And it’s not because the pages are ten years out of date; the site does seem to be maintained pretty regularly, with a lot of pages having a last-modified timestamp within a month or two. Nifty stuff.

Comments are closed.