# Course Report: A Brief History of Timekeeping 01

As mentioned a few times previously, the class I’m teaching this term is a “Scholars Research Seminar” on time and timekeeping. As this is an entirely new course, and will be consuming a lot of my mental energy, I plan to post occasional reports on what I’m doing to the blog.

Today was the first day of class, so a good chunk of the time was spent on introducing the basics of the course (my PowerPoint slides, for those who care), and going through one slightly silly example.

The stated learning goals for SRS courses (students should learn how to formulate a research question, find and evaluate evidence, present results, etc.) mean that they’re really basically “How to make an academic argument” courses, so I’m taking that as a general theme. It’s hard to assign reading for the first day of a course, though, so I went with an illustrative example that didn’t require pre-class reading: the physics of Goodnight Moon. It’s admittedly silly, but serves as an illustration of the basic process: you read it a zillion times, and you start asking “How long does this take?” After puzzling over it for a while, you realize you can use size of the Moon as a reference for its own motion, and then a little Googling gets you the size. Put that together with a little reasoning from common knowledge, and you can get an answer to the question.

I think it went over reasonably well; they were kind of quiet, but it’s the first day. It took quite a while to get them to see that the Moon can serve as its own scale reference, but they got the angular speed (360 degrees in 24 hours) pretty quickly.

I also gave them their first major assignment, which is to make a measurement of the performance of some timekeeping device. I gave them three timing resources: a digital timer like the one I tested last year, one of the plastic sand timers I tested last week, and the link to NIST’s official US time widget. I left the exact thing to be measured up to them (the due date’s the 25th, so they have time), and that gives them a reasonable number of options. We’ll see how that goes.

Friday’s class is the first of a few that will deal with kookery of various sorts, with Stick-in-the-Mud Astronomy assigned as reading, and a clip from this Ancient Aliens episode planned as a starting point for in-class discussion (the bits about Newgrange, which is also in the reading assigned from Dan Falk’s In Search of Time). These introduce the idea of solstices and equinoxes, and also provides an easy example of a slick but deeply, deeply flawed argument that we can easily take apart. Sadly, it does not feature my favorite of the kooky Ancient Alien theorists, but you can’t have everything.

1. #1 Frank Wappler
January 4, 2012

Chad Orzel wrote (January 4, 2012 1:12 PM):
> the basics of the course […] my PowerPoint slides [ http://scienceblogs.com/principles/upload/2012/01/course_report_a_brief_history/010412Notes.pdf ]

Thanks for making those publicly available.

> first major assignment, which is to make a measurement of the performance of some timekeeping device

Slide 18 of the course notes sketches the applicable definition how to measure “the performance of some timekeeping device“; essentially:

How do you evaluate the performance of a clock?
⇒ Compare to another clock

For instance, slide 12 already states an (actual, or expected) result of evaluating the performance of some particular clock(s):

[…] Use atoms as a reference […] Performance: Lose 1 s in 100,000,000 years

(i.e. apparently a “Performance” value of magnitude about 3 · 10-17; where the prescribed comparison is surely between various such/suitable atomic clocks among each other.)

Now, slide 11 apparently mentions another (also real-valued) quantity:

Quartz Oscillators […] Accurate to ~ 10 s / year

(i.e. apparently an “accuracy” value of (magnitude) about 3 · 10-7.

> students should learn how to formulate a research question, find and evaluate evidence […]

What’s the applicable definition how to measure “the accuracy of some timekeeping device“, please?

2. #2 Mary
January 4, 2012

Presumably the NIST widget is supposed to be their reference clock?

It would almost be more interesting to give them two stop watches and a sand timer, and the same problem. Because then they’d have to come to you and complain that the two watches don’t always agree, and which one are they supposed to believe? And then you could say “Yes, that is a problem, isn’t it?” and lead them down the road to developing Allan variation and understanding that you can’t ever know how good a clock is, just how well it agrees with another clock.

And then blow their minds by pointing out that we still believe our clock more than we believe our rulers, and now define the meter in terms of seconds and the speed of light…

3. #3 Tony P
January 4, 2012

You might want to reach out to Mike Umbricht at Brown University’s Ladd Observatory.

He gave an amazing talk about 19th century time synchronization involving observatory measurements transmitted as pulses over a wired network at DC401 in Providence.

I know someone recorded video of the presentation but damned if I can find it online.

4. #4 Derek in DC
January 4, 2012

Whenever I’m disconnected from the internet (yes, people can still disconnect from the internet) I’ve noticed the clock on my 6 year old Mac Mini runs slow when it’s not updating from the online Apple source. Interesting how, even today, building devices that consistently keep good time is a challenge.