Admittedly, much of the technical stuff that I fuss about in photography relates to now-archaic technology, but some tricks are timeless

As are, it seems, some photographic errors.

Here's a fun little analysis of smart peoples' tendency to apply large amounts of brainpower and effort to solving the wrong problem. I confess I enjoy both the writing style and the fun poked at Wolfram (who, let's face it, is an easy and deserving target for vanity-deflation), but I also liked the piece for its pacing (it keeps making good points, so that the length is worthwhile) and because it suggests an intellectual basis for a gut feeling I've long had: I like tools that do what I say, not what they think I want. It's the only rationalization I can offer for my peculiar and snobbish-seeming tastes (nvi, mutt, *BSD, mechanical cameras, locally-aministered email and not facebook, etc...)

Tout l'été, les samedis matins à 8h30 (heure avancée de l'est), France Musique rediffuse des opérettes rarements entendues, cueuillies des archives de l'ancien service lyrique de la radiodiffusion française. De quoi démarrer en douceur son week-end.

Word has gone out that PRL (the Physical Review Letters) are tightening their publication standards. In a sense, this is long overdue: a journal once intended for quick turnaround of brief, legibly-written communications of broad interest is now publishing 80 papers a week, far more than anyone can follow. Unfortunately, the publishing landscape around PRL has changed, and I fear that this new strictness is being imposed largely out of a desire to compete with Science and the Nature journals, which have achieved a very high impact factor by accepting a very limited number of papers (in physics) and insisting that they be short and present major advances or discoveries. The problem with this is that you need detailed, technically sound—but not revolutionary— papers to make progress in any field, and people are becoming reluctant to write those as it becomes harder to find respectable places to publish them. If all the impressive-sounding journals act like Nature, then we'll only get Nature-style papers: lots of hype (to sell your result as revolutionary) and not much detail (to fit in the length limits).

Most of the group went out to dinner at Rangzen the other night, and so we all got to overhear someone at the next table over trying to impress his dinner companions with words like "hyperfine" and "dark state". Hint: if you want to amaze your date with atomic physics jargon, don't do it in a Central Square restaurant liable to contain ten actual atomic physicists in need of a good laugh. We were polite and didn't spoil his moment in the sun, pretending instead that the sudden outbreak of mirth in our ranks was due to our own inexhaustible wit.

The incident did spark a more serious conversation among us about the problem of explaining quantum mechanics to laypeople. I've been thinking for a while that we tend to overemphasize the wierdness of QM in speaking about it, as though the uncertainty principle, entanglement, superposition states and all the peculiar-sounding things you can do with them were miraculous revealed truths that you just have to believe in. Much better, in my view, to explain that they are perfectly sensible mathematical consequences of easily-verified experimental observations, and that you have to believe in magic and incredibly contorted chains of reasoning to avoid accepting them. It was pointed out to me, however, that this kind of teaching would make me terribly unpopular: most people want QM to be magic, and lose interest as soon as you tell them that it's sensible (if subtle and occasionally surprising) and quite well understood, and that it has nothing to do with their current feelings of indecision as to whether to have the chicken or the pasta for dinner.

Various organisations have rules and restrictions on the form of passwords (must be a certain length, contain a certain density of special characters, etc ...) Leaving aside the truly misbegotten ones like those that place an upper bound on password length, I'm beginning to wonder whether all such restrictions are a bad idea to begin with. If the restrictions are not enforced, then they are meaningless (though perhaps good advice). If they are enforced, then there is some code that manipulates the password in cleartext form, before it goes through a hash function. That seems almost as unfortunate as the websites that email my passwords back to me after I register.

Comments from people who have thought about this more than I have are welcome.

Two items in the category of feats whose attempt is baffling and whose achievement is inspiring:

For those of us not currently having fun coding on the icfp2009 contest, it's worth having a look at last year's contest. In particular, the results presentation. To be specific, the bit halfway through where they announce the Judge's Prize. That bit about implementing a soft-real-time control system in a macro-expansion-based typesetting language? Yeah, that. "Extremely cool hacker" is an understatement.

Meanwhile, someone has implemented, or rather created, Thyrd; a visual programming environment with a new Forth-like (well really Joy-like) concatenative programming notation in the string-processing language Tcl.

Which just goes to show: it's not how shiny the tools in your chest are, it's how good you are at making the most out of them.

I heard an interesting thought this morning: what if the U.S. government is financing research into quantum computation not because they think it will work but because they want to be really sure that it won't?

On a related note, guess who's tackling real-world implementation of quantum key distribution networks.[1]

I saw a presentation the other day on education outreach initiatives in Zambia. The most depressing thing, to me, was not the description of the bleak conditions under which teaching must be done there, but one of the questions from the audience. We were shown a picture of a blackboard with a square root function (the model) and a bunch of data points (experimental results from the students). A well-regarded, high-calibre MIT physics graduate student had to ask how they plotted the square root function on the board. Apparently the post-graphing-calculator generation are unaware that fairly accurate sketches of simple non-linear functions were in fact possible before the advent of the Texas Instruments TI-83. In which respect those deeply underprivileged students in Lusaka have them trounced.

And finally, a brief tribute to perhaps the most surprising—and enduringly helpful—class I've ever taken. ESC399 Engineering Communications. It sounded dreadful; like a dumbed-down English class taught by a patronizing litterature professor or worse, a remedial grammar class. Nothing of the sort. It was run by a crack team whose mission was to teach us major principles (and unprincipled but powerful tricks) to writing or presenting technical information concisely and persuasively. It forever changed the way I edit, taught me to suppress my taste for deeply nested sentences, and led to a long-running joke about my slide presentations and nuclear bombs. I hope your school had something like it.

[1] Appendix for the differently-buzzworded:

Quantum communication (including quantum key distribution) is about sharing quantum states between spatially separated locations. It's good for sharing secrets protected by physical guarantees against eavesdropping. It's an immature but rapidly developing practical reality. There'll probably be a few nasty surprises as attackers learn to exploit flaws in the physical implementations of those theoretically perfect protocols, but it's on a reasonably firm technological footing.

Quantum computation is about computers that process general quantum states (including superpositions of a bit being 1 and 0). Contrary to certain claims in the popular press quantum computers don't give you a general exponential speedup in computation speed. However, it so happens that they do give you an exponential speedup on one problem in particular: factoring large numbers. Which is of great interest to the NSA, the Chinese security services, and anyone else attracted to the idea of breaking the RSA cipher that underlies most modern public-key cryptography. So the real-world applications of quantum computation include cryptanalysis of classical communications and getting lots of funding from the NSA, the Chinese security services, and anyone etc. The hitch? They don't work. There are a countably infinite number of proposals for quantum computation, many of which have taught us a great deal but none of which scale up to anything useful with currently-available technology. I would estimate at least ten years and two revolutionary ideas (and possibly much longer and more) before the error rates get low enough and the number of qubits get large enough that anything practical gets calculated on a quantum computer.

Quantum simulation is about replacing completely intractable systems with equivalent ones that are experimentally manageable. This is like sticking a model airplane in a wind tunnel: you can't reliably calculate the aerodynamics of the airplane, or even of the model airplane, but you know that the airplane and the model have same aerodynamics and you can measure the behaviour of the model. The primary goal here is to figure out what electrons are up to in high-T_c superconductors and certain exotic magnetic materials. This is a hard problem (once you adjust for density, electrons in a metal are astonishingly cold), but interesting results are only a factor of a few in temperature and perhaps a few years away.

I don't work in any of these fields (I'm into quantum metrology, which isn't sexy enough to be a buzzword), but I'm surrounded by people who do, so I claim that I'm both lightly informed and mostly unbiased.

Erik Naggum, noted denizen of USENET in general and comp.lang.lisp in particular, is reported to have passed away a few days ago. He had a fearsome reputation, mostly born of not suffering fools, ever. He was also known for being helpful to anyone who checked his/her ego at the door, for technical mastery on a range of topics and for the kind of wisdom that comes of applying impassionate logical analysis to real-life behaviour. I'll only link to one of his posts, one that made a deep impression on me and which I still occasionally quote to people: the purpose of higher education.

I think I may have been lurking on c.l.l. when he posted this; I'm not sure how else I would have run into it.

Our benevolent overlords up the administrative ladder have very carefully updated the signs beside all the doors on the hallway. The signs used to indicate the names of the occupants of each office as of late 2006. Now they have the shiny! new! CUA logo along with a list of the names of the occupants of each office as of late 2006.

Either they are giving us a hint about their priorities, or they are unaware of the fact that grad students (eventually) graduate and postdocs (eventually) get stable jobs.

Fortunately, there is usually at least one name on the list that still corresponds to an actual occupant who can tell you which continent the others went off to after their Ph.D. or postdoc contract.

On the left, carrots slow-cooked with onions in olive oil. In the centre, brown rice (ok, so it's a tattered old flag). To the right, a mixture of potatoes rissolées (home fries) with Italian sausage, collard greens, and snow peas. All prepared to the sound of the Spanish national Jazz broadcast.

Incidentally, I've identified a trait I like in Spanish national radio, in common with the three movies I saw this weekend (I know, ridiculous, but they were all running for a few nights only):

• RNE will play good jazz, a Stabat Mater, an Alban Berg chamber work, and movie themes from recent summer blockbusters, and have something interesting to say about each.
  
• Azur & Asmar is a simple animated fairytale without obvious moral complexity or confusing plotting, that still manages to make a strong point about religious tolerance and civilised behaviour, and where those were most likely to be found in the days of fairytales (Hint: half the dialogue is in Arabic).
  
• Pulp Fiction is one-half absurd but well-written soliloquy, and one-half allusive violence.
  
• Sita Sings the Blues retells a great Hindu epic as a music video to old Annette Hanshaw recordings, with intermittent narration, commentary and analysis from a trio of shadow puppets (their banter is the best thing about the film).
  

They're all cheerfully intelligent without being pretentious.

Free 19" CRT at the wrong end of Cambridge ⇒ One excessively good upper-body workout.

Despite my obsessions-of-the-month, and thanks in part to early sunrises, I've been reading a vaguely civilised amount. This morning I finished of Michael Chabon's The Yiddish Policemen's Union, a fascinating combination of noir mystery novel with alternative history and linguistic speculation. It posits the survival of a large population of European Jews who set up a Yiddish-speaking enclave in Alaska, with the forbearance of the American authorities. Its metaphors are endlessly inventive and frequently delightful, its world-building is thorough and intriguing, and its plot combines unabashed extravagance with a satisfying attention to detail and the tying together of clues. I originally got it as a gift for a recipient to whom I now deem it quite unsuited [1]. I'm pondering where to redirect it.

Tangentially, the same sort of surface comedy with darker underlying themes, wit, carefully-crafted foreshadowing and extravagant plotting also appear in The Brothers Bloom, the first movie I've seen in some months. It's a caper movie (sort of [2]) about a pair of besuited cruise-ship traveling brothers who run cons of great semiotic complexity. I'm quite sure I didn't catch half the in jokes and subtle references, but it was great fun anyway.

[1] There's an art to gift-booking, and the first lesson to learn is that literature is not divided into Good books and Bad books, but into books-to-send-to-Tom, books-to-send-to-Dick, books-to-send-to-Harry, and books-to-keep-for-myself.

[2] It's also a romance and a sort of delayed-bildungsroman about young adults with arrested emotional development. It doesn't take itself too seriously for easy laughs, but it has more serious things to say and less gleeful naughtiness than expected in a con-man flick.

As a playground to watch Adam Smith's idea of a free market at work, eBay suffers from two irritating flaws:

• Its operators actively encourage irrational behaviour. They don't want actors motivated by enlightened self-interest, they want victory-obsessed auction addicts. Their whole windorphins advertising campaign is a transparent example of this attitude, as are the constant email reminders to increase your maximum bid to avoid losing. If you must get emotional and competitive about these things, remember that watching your competitor overpay for something is not a defeat.
• The ratings system is sociologically broken. People are in the habit of giving perfect ratings for any satisfactory transaction, even when the feedback is secret (as with the Detailed Seller Ratings). This in turn means that perfect ratings are expected and even necessary; giving a seller reasonable but not perfect marks is now deemed offensive (I've made that mistake) and significantly penalizes the seller (or so they claim). The net result is that every seller has a rating over 99%, the differences are lost in the noise, and you can't change that without being a jerk that no one wants to deal with.

Does anyone know of a mixed choir named Viva Voce? They have a lovely, incongruously churchy, deadpan rendition of Tom Lehrer's "I Hold Your Hand in Mine". The radio announcer had to warn listeners not to listen too closely.

A novel co-written by Neil Gaiman and Terry Pratchett, dedicated to G.K. Chesterton, of which a copy reputedly exists in the Vatican library, is more or less bound to be a memorable experience.

The title, incidentally, is also the title of the novel, not just a description of its opening pages.

I am struck by the thematic similarity between this book and Sept Jours pour une éternité ..., which is also about practical earth-bound angels and demons conspiring to save the world from the abstract plans of their out-of-touch bosses (sounds like Dilbert). It's just that the latter played the story for romance and pathos and was written by a French ex-pat in San Francisco, with all that that entails, while the former is going for irony and satire, which is much more appropriate.

Spotted during a Debian update: Canorus, a GUI music editor project aiming for Lilypond interoperability. This might be of interest to some choir-masters I know. Or, given that they're also TeX-masters, not.

There is something subtly wonderful about the flavour of a well-boiled Yukon potato.

Though it took us longer than our advisor remembers to produce squeezed spin states at all, and though we've spent the winter in unfruitful attempts to officially convince the world (by publication) that we have done so, we had squeezed states within a day of starting up the machine again on Wednesday. We're now talking about producing these strange quantum states as a necessary first step to investigating other things, rather than as an end in itself. Thus, inch by inch, the revolutionary becomes routine.

This was brought home to me by having to skim through some old lab books yesterday, seeing the first embryonic sketches of what now seem obvious and inevitable ideas, as well as the obvious and inevitable assumptions that came back to haunt us six months later when they proved unfounded.

As a student and experimental scientist the taking of notes is one of my main activities. Another is calculating stuff, and while I can type and edit English much faster on screen than on paper—these days I tend to write text from the middle outwards, working on whichever paragraph I see most clearly at the moment—I've yet to see a computer interface that would let me sling around math, physics or spatial information with anything like the ease of a little stick that makes marks on a surface. Thus, while my demands on my pens and pencils are more modest than those of a draughtsman, artist, or writer, I do care about having writing implements that don't annoy me while I'm listening to a lecturer, watching an experiment or trying to think. And since one never knows when an idea or TODO list will need to be jotted (if only to get it out of my head so I can stop worrying about it), I want a writing implement and scrap paper to hand wherever I go.

For a long time my solution to this problem has been to disperse good, cheap pens widely: I have two or three Pilot V7s in my pocket, a handful more in my backpack, several on my desk, in lab, or any other place I work regularly. While engineering school was nearly all-ink for me, in grad school I started using graphite for calculations and problem sets: a few inexpensive draughting pencils that I get along with and one good ProTouchII, bought on a whim a couple of years ago, that I quite like.

Since I found a lovely ruler on Monday, I've been experimenting with changes to this arrangement. Partly, this is about my eternal search for verdant pastures beyond the barbed wire, and because Spring is a fun time to mess with your ingrained habits. Partly, too, because my embryonic ecological conscience and my inherited distaste for waste and nondurable engineering both give me a minor pang every time I empty out a pen and toss it into the nearest bin. Finally, my writing and note-taking habits are going to have to change soon, as the neat division of my life into lectures and lab gives way to random talks, informal discussion and exploratory calculation. Of all the changes I'm facing this is the easiest to accept and do something about, so of course I'm obsessing about pens, pencils and notebooks.

I'm trying out a clutch pencil, and I've provisionally adopted as my main working pen a good ballpoint that a friend gave me. I'll probably try a few more options in search of a non-disposable pen that has the same precision and smoothness as the Pilot needlepoints and a steel-gripped pencil with the right balance and solidity. Hopefully, I'll find them and convert myself to using a few good long-lasting scribbling tools rather than an army of cheap ones.

I'm still keeping a box of V7s in reserve, just in case.

Lime-lemon-chocolate-cranberry cheesecake. It wasn't quite marbled, perhaps swamp-mossed would be a better description, but it survived a bicycle ride and mostly failed to survive the ensuing party.

Getting a couple of Roy Orbison CDs. I hadn't realised how long he'd been active as a musician; his career is basically the history of mainstream rock. I also hadn't realized (or remembered) that the man had a real singing voice.

Ken Macleod's Star Fraction is a mix of Trotskyism, Libertarianism and William Gibson. I enjoyed the political speculation more than the hard sci-fi, which probably means I'm growing up.

The quiet little Agnus Dei (soprano aria) in Mozart's Coronation Mass is a little pearl surrounded by sparklier and more attention-grabbing jewels.

Listening to Radio New Zealand in the early afternoon. They start their morning's programming with a recording of a bird call; a different bird each morning.

Cleaning out an office for two new postdocs who just landed. Besides the old computers and forest of catalogs that have been donated to foraging undergrads, I claimed for myself an enormous water cup, a set of small screwdrivers, some rain gear, and a gorgeous hardened steel Starrett ruler with 1/64" marks. Now to find a pencil worthy of its company.

A rebuttal to all the Chuck Norris jokes: Johnny Cash vs. Chuck Norris. I liked Chuck Norris jokes as much as the next guy when they first appeared, but that was nearly a decade ago.