Uncertain Principles

We've been watching a whole bunch of Japanese cartooooons at Chateau Steelypips recently. Most of these have been DVD's via Netflix, with a couple of things loaned to me by one of my students, though we've now moved on to episodes of Fullmetal Alchemist recorded from Adult Swim on the Cartoon Network.

The DVD sets all come with trailers for other cartoon series, which are... interesting, and I mean that in a "Go away Japanese id! You are scary!" sort of way. The number of borderline sexualized schoolgirls-with-guns stories being pitched on the Fullmetal Alchemist DVD's is really scary. I'm not what you'd call a McKinnonite feminist by any stretch, but this stuff kind of gives me the creeps (and yes, I'm aware that this is only the shiny, happy, family-friendly stuff). I pretty much gave up on Hellsing, in part because the miniskirt-and-go-go-boots uniform given to the main character (the men all get to wear pants) was just a tiny bit over the top.

So there's some creepy stuff in the Japanese id. And then, there's also the deeply goofy, such as the Japanese lyrics to the Imperial March from Star Wars (via Diane Duane).

And we won't even talk about the pop songs that run over the credits on Fullmetal Alchemist...

(Recommendations of other cartoons to check out are welcome in the comments. Bonus points for recommending stuff I can get via Netflix.)

Posted at 9:54 AM | link | follow-ups | 14 comments

We had a really interesting colloquium today, with the provocative title "Can Gravitons (even in principle) Be Detected?" The speaker, Steve Boughn of Haverford College went through a bunch of arguments to show that it's nearly impossible to detect a single graviton, and quoted Freeman Dyson as asking whether something that can't be detected can sensibly be said to exist.

Of course, for the topic to really make sense, you need to define a few terms. The experiments he was discussing were experiments to detect single quanta of gravitational radiation-- basically, the gravity-wave equivalent of looking for extremely weak pulses of light. This doesn't really have anything to do with gravity on the large scale (which unquestionably exists), or gravitons as hypothetical carriers of force. He was talking about experiments that attempt to detect gravitational waves of a particular wavelength in the limit where the total energy content of the wave is so small that it can only amount to one graviton, where the energy of a graviton is given by something like the Planck formula for light-- Planck's constant h times the frequency of the wave.

One of the especially neat things about the talk was that it was all done with back-of-the-envelope estimates. For example, he asked what you would need to do to build a detector that would have a 100% chance of detecting a single graviton hitting it (that sounds a little weird, but you can do it for photons-- you can't say where in the detector it will be picked up, but you can guarantee that it hits somewhere). Using a few basic formulas-- the interaction cross-section for gravitons interacting with resonant detectors, the formula for energy absorbed in a detector, some basic relationships involving the behavior of quantum particles-- he was able to show that there's no way to make such a detector without having it collapse into a black hole.

He did similar calculations for the case of putting a single detector in front of a large flux of gravitons and trying to only pick up one of them. When you impose the condition that the flux has to be small enough that there's only one graviton in the detector at a time (the analogous condition for photons is not sufficient to get anyone to believe that you've detected single photons, but it's the absolute minimum you have to do to get people to even consider the idea), you find that no sensible detector can hope to detect a single graviton in a reasonable amount of time. Human-scale detectors would require something like 10¹⁷ times the current age of the universe to detect a single graviton. If you stretch the definition of "sensible" to include graviton detectors the size of Jupiter, you'd still need something like 100 times the current age of the universe.

All of these conclusions were demonstrated with nothing more than back-of-the-envelope algebra, and a few fairly basic formulae. These were basically "Fermi Questions," and I'm a sucker for this stuff. It ends up beign a fairly convincing case that it's nearly impossible to detect gravitons as single particles, even when you're allowed to do things like taking half of the hydrogen in the universe and using it as a detector to look for the gravitational decay of one atom in the other half of the hydrogen in the universe.

The final example was the stuff of space opera. He imagined a sphere of silicon, honeycombed with small electrical detectors (to detect electrons knocked loose from atoms by single gravitons), with a radius roughly half of the radius of the Earth, and a mass about 1% of Earth's mass. If you were to take such a device, and put it in orbit around a white dwarf star (which produces gravitons at about the right level to be in the plausibly quantum regime), you could expect to detect something like 1000 gravitons in the age of the universe.

So, I guess it is possible to detect gravitons. I wouldn't hold my breath waiting, though...

Posted at 8:13 PM | link | follow-ups | 9 comments

Dan Drezner was denied tenure at the University of Chicago.

Now, I've never really had much interest in his blog (I think, but can't say for sure, that this is the first time I've ever linked to it), but the only junior faculty member whose head doesn't snap up at the words "denied tenure" is one who's three days dead. High-profile tenure denials are the spectacular car crashes of academia, and we can't help but look.

As dozens of people have pointed out, this makes Chicago 0-for-2 in tenuring well-known bloggers. Predictably, this has led to overheated articles from the usual suspects, speculating about whether blogging played a role in the decisions. Inside Higher Ed and the New York Sun lead the way.

Sean Carroll (the other Chicago victim) has posted a very reasonable response to this whole thing:

There’s a short answer and a long answer. The short answer is “No, it’s not blogging that prevents you from getting tenure; it’s because some people in your department (or the dean, or whatever) didn’t think that your research was good enough.” The blog was not a hot topic of discussion in my case, and I’m pretty sure that many of my colleagues don’t even know what a blog is, much less have a negative opinion of mine. The longer answer must deal with the issue of why someone doesn’t think your research was good enough. {...} I think my own research was both solid and influential, and Dan’s looks pretty good from the perspective of a complete outsider; certainly neither of us had simply sat around for six years. But these are judgment calls, and a lot goes into that judgment. Like it or not, if you are very visibly spending a great deal of time doing things other than research, people might begin to wonder how devoted you are to the enterprise. To first order it doesn’t really matter whether that time is spent blogging or playing the banjo; some folks will think that you could have been spending that time doing research. {...} Of course nobody will ever say that they voted against giving tenure to someone because that person spent too much time on public outreach, or put too much effort into their teaching. But getting a reputation at being really good at that stuff could in principle make it harder to have your research accomplishments recognized — or not. It’s just impossible to tell, without access to powerful mind-reading rays that one can train on the brains of the senior faculty.

I'd be lying if I said that I didn't worry about whether this blog might have a negative effect on my tenure chances. But that right there is the problem with the current tenure system-- I worry about everything with regard to my tenure chances. Every junior faculty member does. Along with all the academic bloggers who are now fretting about whether their blogs will get them dumped, there are people out there right now wondering "Does this shirt match these pants, and how will that affect my tenure chances?" It's inherent in the system-- they're trying to pack thirty years' worth of "Is this the week I get fired?" stress into only six years before you get lifetime job security.

When it comes to the blog, I've sort of decided that you have to draw the line somewhere. There are so many other ways that I've ended up shifting my life around in an attempt to improve my tenure chances (or at least make it look like I spend all my time doing research) that it's started to make me ornery. I've managed (with help from Kate) to avoid doing anything really stupid out of sheer pig-headedness, but I need at least one stupid and insignificant act of defiance, and you're reading it.

Ultimately, I've come to the conclusion that having the chance to blow off steam on the Internet is worth the risk of the small chance that this might negatively affect my tenure case. I like having a space where I can talk informally about what I do, and kick around the occasional idea, and list pop songs that I find entertaining, and post silly pictures of the dog, and all the other nonsense that goes on in the electronic wing of Chateau Steelypips. It helps keep me grounded on the days when I feel like taking a wrench to the side of a $10,000 vacuum pump.

I don't know if I'd recommend blogging specifically to other young faculty, but I would say this: given the strain of the tenure process, and the way it can eat your life, I definitely recommend finding at least one thing you can do that doesn't have anything to do with work. Pick an activity that you like, and do that thing, and don't think about what it means for your tenure prospects. It'll be better for your sanity in the long run.

Posted at 6:43 AM | link | follow-ups | 7 comments

Via Technorati, which is wonderful when it works, I have found a new essential blog: Setshot, which offers "Basketball psychology, etiquette and lore for the aging hoopster." They provide critical advice for those of us who don't have the ups we used to (though they need to do a post on "signature moves that don't work"). And they linked one of my old posts. What more could you want?

OK, for those of you not satisfied with Setshot's brand of Supreme Court commentary, I ought to provide something more serious. Back when I was catching up on stuff I missed during my hiatus, a passing mention on Crooked Timber led me to "...My Heart's in Accra", a blog by a classmate of mine from Williams (though I think he ended up a year behind me). I didn't know him particularly well, and I'm sure he remembers me as a drunken idiot if at all, but his blog provides as much detailed discussion of serious topics relating to international development as you could possibly want.

I've said several times that the part of the blogging phenomenon that offers the most interesting potential is not the bit where random people rant about politics, but rather the bit where people from different walks of life talk about what it is that they do. In that vein, off Blogger's "Blogs of Note" list, we have Ah Yes, Medical School, a collection of, well, funny med school stories. They're sometimes a little off-color, but it's interesting to see the medical hazing training process from the inside.

Finally, no links dump would be complete without some self-promotion, so let me note that I recently updated my not-dead-yet book log, with four new books: Peeps, Accelerando, Anansi Boys, and Thud!.

Posted at 8:03 AM | link | follow-ups | 1 comment

It's another Research Day for me, so here's some filler.

It's been a while since I posted any new dog pictures, but she continues to be the Best Emmy Ever. Don't be fooled, though-- her cute exterior hides a Mighty Hunter...

Here she is as she appears to her humans. Awwwww....

Here she is as she appears to the bunnies in the back yard. Spooooky!

Posted at 7:10 AM | link | follow-ups | 3 comments

Last week, I talked a bit about Roy Glauber's share of the Physics Nobel, and I really ought to say something about the other two guys. The problem is, their half of the prize is really best thought of as a "lifetime achievement" sort of award. Both Jan Hall and Ted Hänsch have been around for a long time, and involved in lots of different things. Hänsch was even a co-author on one of the two proposals that led to laser cooling (Hänsch and Art Schawlow wrote one paper on the subject, and Dave Wineland and Hans Dehmelt the other. Interestingly, Wineland is now the only one of those four without a Nobel Prize, though if there's even a prize for quantum computing research, he's almost got to be one of the winners).

The one specific achievement mentioned in the citation is the development of the optical frequency comb technique, which is at least something concrete to explain, so I'll give that a shot. The problem is, it doesn't end up sounding like anything all that earth-shaking, even though it is a revolutionary development in precision spectroscopy.

The whole thing starts with the definition of time. The second is presently defined as "the duration of 9,192,631,770 periods of the radiation corresponding to the transition between the two hyperfine levels of the ground state of the cesium 133 atom." The level spacings of cesium are absolutely determined by the laws of physics, making them a perfect reference-- all cesium atoms are identical, so no matter where you make your cesium atomic clock (within reason), you'll get the same result.

To shamelessly swipe an analogy from Bill Phillips, the way an atomic clock works is pretty much like the way you would use the NIST web site to check the performance of your watch. First, you synchronize your watch with the NIST clock, then you wait a while, and check back to see if they still have the same time. If they do, then you know your watch is working; if they don't, then you correct it, and check back again in a little while.

An atomic clock works the same way: first, you set up a microwave source in your lab, operating at a frequency of 9,192,631,770 Hz, and then you send it into a sample of cesium atoms in a microwave cavity. This prepares the atoms in a state in which they oscillate rapidly, like little clocks ticking 9,192,631,770 times per second. Then you wait some time-- long enough for the atoms to either fly through the air to a second microwave cavity, or fly up into the air and drop back down through the same cavity-- and hit them again. If you've got the frequency exactly right, you'll find that all of the atoms have changed from one state to the other. If you're off by even a little bit, some of the atoms will still be in the original state, and you'll know you need to correct the frequency.

In the end, what you have is a microwave source operating at exactly 9,192,631,770 Hz, stabilized by regular comparisons to the cesium atoms. This is analogous to having a watch that you check regularly against the NIST site-- the watch itself may not work perfectly, but regular adjustments based on a better clock keep it in line.

Now, the choice of cesium is arbitrary, and largely a mater of historical inertia. You could perfectly well define the second in terms of some other transition in some other atom, and there are dozens of people working in precision measurement who will enthusiastically lobby for one non-cesium atom or another. In fact, there are some people who will tell you that cesium has a number of properties that make it just about the worst atom you could've picked for the standard.

When you sit down to start judging between candidate frequency standards, you find that the figure of merit for determining the stability of the clock goes like the frequency resolution of your measurement divided by the frequency of the transition itself (Δ ν / ν, in the usual notation). The measurement resolution is pretty similar for all the different atomic species, so that means you win big if you can use a transition with a higher transition frequency than cesium. There are lots of candidates, involving radiation with frequencies from the mid-infrared (my work in grad school was tangentially related to a proposal to make a standard out of a transition in xenon with a wavelength of 2.19 microns, or a frequency of roughly 136,000,000,000,000 Hz) into the ultraviolet (a transition in mercury ions at 282 nm, or 1,064,000,000,000,000 Hz).

The problem with these standards is not anything to do with the implementation. The mercury ion group in Boulder has done amazing things with stabilizing lasers to the Hg⁺ transition. The problem is a practical one: 1064 THz is just way, way too fast a frequency to count. Even the 9.19 GHz of the cesium standard is a bit too much for ordinary electronics, so a commercial atomic clock uses some nifty microwave electronics to divide that frequency down, and gives you an output at 10 MHz (10,000,000 Hz), which conventional electronics can easily deal with.

Those techniques are completely hopeless for frequencies up in the THz range, let alone 1,064 THz. In order to make the stabilized mercury laser into a useful clock, you need to use a whole bunch of different tricks to get that frequency down into a region you can work with-- something you can count to produce a reasonable one-second "tick." Each of those tricks introduces a little potential instability, so while you win big in stability by going to a higher frequency, the little errors that pile up in the frequency conversion process pile up, and make it less useful as a clock.

The frequency comb technique is a way to get around this problem. What Hänsch and Hall (and some other people) did was to make a device that lets you connect the many-THz optical frequencies of the ion standards directly to the GHz-level frequencies that are practical for real applications. It turns out you can make a laser that produces light at multiple frequencies (hundreds of thousands of them) separated by something like 1 GHz. If you set it up right, one of those laser modes will be close to the transition frequency you want to use as a standard; if you lock that frequency by comparing it to an atomic sample, you can (relatively) easily measure the separation between laser modes, and end up with a 1 GHz signal with the stability of your 1064 THz frequency standard. You go from optical frequencies to microwave frequencies in a single step, and the uncertainty introduced is minimal.

If you're not normally concerned with the detailed workings of atomic clocks, this may not seem like a very big deal, but it is. It's a remarkably elegant solution to a deeply vexing problem, and the idea has really taken the precision measurement community by storm (and precision measurement people are a pretty cautious bunch-- for this to catch on as quickly as it did is saying something). It's also taken a great deal of work by a number of exceptionally smart people to make it a practical device. That alone is worthy of recognition; put together with all the many other things Hänsch and Hall have done for the field of precision laser spectroscopy, and, well, it's richly deserving of a Nobel prize.

Posted at 7:56 AM | link | follow-ups | 7 comments

I don't regularly read Slate any more, so I'm a little late noticing Set Stevenson's article on why he should stick to tv. (I could've sworn the pointer to it came from a Making Light comment thread, but now I can't find it...) The argument really centers around the death of a character in Serenity, and the relative lack of effect this has on people who aren't familiar with the original show:

Take that character who dies in Serenity. Had Firefly lived on as a TV series, Whedon would have invested the character with foibles and hidden strengths. Our bond with the character would have had ample time to develop as we watched countless informal, telling moments. Then the character might have been killed in Season 3—only after this loss would be certain to stomp the heart of any die-hard viewer.

That's true, as far as it goes-- as Stevenson notes, the character in question doesn't play much of a role in the movie, so the death scene relies on knowledge of the backstory for its impact. But then, I'm not convinced that this necessarily requires three seasons worth of exposition-- it seems to me that, done right, it ought to be possible to get some serious impact out of the death of a supporting character within the context of a two-hour movie.

Consider, for example, the other great geek saga of recent years, Peter Jackson's The Lord of the Rings movies. Specifically, Boromir's death in The Fellowship of the Ring. That had a lot more punch than the death scene in Serenity, because Peter Jackson and Sean Bean did a superb job of making you care about what happened to that character.

"Yeah, but that's a three-hour movie," you say. Sure, but you're an hour into it before Boromir shows up. "But Serenity is an ensemble cast, and can't spend that much time on one character." Yeah, and Fellowship isn't an ensemble piece? Boromir is one of nine major characters, the same number as the principals in Serenity. (You can bump them both up to ten by including the villains.)

(It's also worth noting that there are actually two deaths of pre-existing characters in Serenity. Stevenson only mentions one, because the other is practically an afterthought. To an uninitiated viewer, I'm not sure either would have much more impact than a third death, that of a minor character who was first introduced in the movie. If anything, I suspect the new guy probably comes off better.)

Now, it's true that Boromir gets more to do in Fellowship than the dead character in Serenity does, but that's a choice on the part of the director. Jackson sets Boromir's death up as an essential part of the climax of the story (even more than Tolkien does), while Whedon makes the death of his character merely an event that happens near the end of the movie.

Jackson also manages to give Boromir a compelling character over the course of several relatively minor scenes-- sparring with the hobbits, being tempted by the Ring, asking mercy for the hobbits outside of Moria, nearly losing his nerve in Lorien. He gets his due without dominating the action, thanks to shrewd directorial choices, and a stellar performance by Sean Bean. Whedon's character is barely in the movie, and as a result, the scene is heavily dependent on knowledge of the tv series for its emotional impact.

This can't be put down entirely to the limitations of film as a medium. At least part of the blame for the relative weakness of the death scene in Serenity has to fall on Whedon as a writer and director. Yeah, it's difficult to do justice to a large cast in a two-hour movie, but it's by no means impossible. Other people manage to do it all the time.

(The strange thing about this is that it's not like Whedon is incapable of doing better-- the two-hour Firefly pilot does a better job of introducing all of the characters than the movie does. The movie gives pretty much everybody other than Mal and River short shrift-- to continue the Peter Jackson comparison, I'd say that five of his nine (Frodo, Sam, Gandalf, Aragorn, Boromir) get better treatment, in character terms, than anyone in Serenity other than those two.)

This is why I'm not sold on Whedon as one of the great storytelling geniuses of our age. I shouldn't need to watch a whole season worth of Firefly or Buffy the Vampire Slayer to make me care about the characters.

Posted at 3:25 PM | link | follow-ups | 13 comments