quivering through sun-drunken delight

Tuesday, February 14, 2006

Resurrection: "aufersteh'n, ja aufersteh'n wirst du"

Hi, everyone. Long time no see. Qu'est-ce qui se passe?

The big news at the beginning of the semester is and must be the new classes, but we're still easing into that (one hasn't yet begun, even), so I'll hold off for the moment. Suffice to say that this year's crop looks to surpass that from the fall, with two number theory classes, including one taught by Andrew Wiles (which hasn't begun, so who can say if it will be beyond the horizon), and a geometry-flavoured class. Indeed, all would be perfect except for a foul scheduling accident that I had to watch happen, in person, with all the dread of watching two trains ram, from a distance, slow-motion-like. The other number theory class was originally at 1630-1830 Monday evening, so it was no surprise to me when the first order of business there yesterday was to negotiate a new time. Alas! Miserere, misero me, the time picked was 1400 Tuesday, consuming, like Jormungandr, the likewise-located discrete math class of Paul Seymour that I had so anticipated. An embarrassment of riches, verily. If this is the mightiest problem we encounter --

-- but it is not; finding the damn'd local chess club is far thornier. It is far more difficult than checking their website. However, I heard today that Ed Witten (yes, that Ed Witten) plays chess at such-and-such a location Fridays at nine o'clock in the evening or so. Needless to say I was struck profoundly by this remarkable confluence, and will have to investigate this rumour.

I should say where I heard it: over dinner. The dean of the graduate school, Bill Russell, hosts monthly gatherings for a small number of random invitees. Tonight my ticket came up. Despite having several days to think about it, though, I didn't manage to ask for a suggestion on dress. This kills me every time. I end up coming with the most mongrel compromises. Today I interpolated my blue sweater between the lavender shirt (and tie) and my suit jacket, with the light pants (and the good shoes, needless to say). I think this worked quite well; in fact the assistant dean had a red sweater-vest type under his suit jacket, and I managed not to be better dressed than the host, which would presumably be unforgiveable. (That was my fear, anyway; you can set me straight on protocol, a subject which didn't quite come up in the past.)

Anyway, dinner was quite the usual story: I managed to definitely not impress before and during dinner but relaxed into the setting afterward. I had a lovely chat with variously an historian, an architect, an economist, and a neural microbiologist. I asked our economist whether he feels (as I had read in a paper recently) that his field has gotten out of touch with the real world, and he told me that he likes the elegance of his theories. I suggested he move to Fine Hall. Architects, it turns out, have a huge breadth of knowledge that they draw on professionally, and we found a mutual interest in Greek philosophy and Bach, inter alia. Finally our historian friend won my heart by enthusiastic interest in my mathematical anecdotes and asking to keep my demonstration napkin.

I'll spare us all the meditations on the nature of happiness and our relationship to the Platonic Realm and instead present a preliminary trifecta of anecdotes.
  1. Birthday Surprise. This is a fun trick to warm up with that consistently astonishes people who've never heard it. It has a good moral, too, that statistics does not come intuitively to people in general. The question is: how many people do you need in a room before it's more likely than not that two of them share a birthday? I observe, trying out some showmanship, that of course you'd need 367 to be certain of it. Now no one ever ventures a guess about number, but on first impression 23, the correct answer, is rather small!
  2. Bottle Imp Paradox. This is one of my favourite problems, but you might want to replace it with something else, because I've never yet met someone who thought the same way. The set-up is this. You are offered a chance to purchase, for any price you care to name, a bottle imp. This bottle imp grants an unlimited number of wishes for you, with the sole condition that you must in turn sell (not discard or gift) the bottle imp to someone else after some finite time span, say twenty years. The condition of sale is that you must sell the imp for a strictly smaller amount of money. (Now, this is a game theory problem, not a gedankenexperiment in the value of a thing, so what this means is that if I pay one hundred cents then I sell it for at most ninety-nine cents. There is no such thing as a half-cent or a peso or inflation or whatever.) Failure to sell (by you, or by the next person, who inherits all the conditions) results in a terrifically gruesome, unspecified penalty exacted by Mephistopheles, from whom you cannot hide. (Or maybe Samiel, the Dark Huntsman.) The question: Do you buy the bottle imp? And if so, for how much? The rational answer is that you do not buy it, for any price: for clearly you would not buy it at one cent, for then no one would be able to buy it; but then not at two cents, for then no one would be willing to buy it; or if not n cents, then also not n+1. But, and I think I'm not alone here, I definitely would buy the damn thing, for as much as I had in my piggy bank. (The bottle imp can make currency, so it's all funny-money anyway.) I'm not saying the bottle imp's wishes would make me happy, technically speaking, but it's got to be worth it; and after all if I'm being irrational in buying it, I can bet there's an irrational person to sell it to....
  3. Angle trisection. Save the best for last, assuming they're still paying attention. Many people (most, frankly, considering the audience you've got to have even to consider telling these stories) have heard that it's impossible to trisect an (arbitrary) angle using straight-edge and compass, or at least can think back to their high-school geometry and what it means. (Bonus points for me: mention that one uses Galois theory to show this, giving a beautiful application of the theory which I tell everyone I study when they ask.) Regular listeners will of course recollect that this can be done by origami, and the construction is wonderfully simple and easy to demonstrate with a pen that writes well on napkins.


So, a very pleasant evening. Moving on, more miscellany. Of course the Olympics have started. I hoped to catch highlights on cbc.ca, as they stream their daily newsprograms -- but the perverse Olympic broadcast regulations force them exactly not to stream their shows for the duration. So I'm completely blanked out, with the sole consolation that the curling scores are updated more-or-less in real time. It would have been nice to see the figure skating program (heck, to watch the curling!) although it seems my favourite guy, Alexei Yagudin, has bowed out (he was having knee problems last I heard, which was years ago... it's been a while without television, really).

Also: you may have heard there was a lot of snow here recently. We're relatively south and still got quite a bit. I'm told it's expected to warm up very shortly, and then it will all be gone. Already much of it seems to have melted. This meant only that I had to act fast, of course. So I took a quick jaunt to get the shot I needed, and picked up a few incidental ones along the way. My batteries, let it be known, did not fail me.


The view almost immediately outside my room. Continue past an archway hidden by the tree:



For those worrried I would catch my death of chill going to eat breakfast and dinner, the far building opens into the dining hall, so splice this one with the last and you'll see about how far I walk. Speaking of which, time for a little indulgence in an old passtime.



A dragon and a monkey and some other things guard the door. And look up a bit?



More. (Click through for the original.) And -- what's that? -- how can it be?



For he is the Kwisatz Haderach! Oh, yes, more tigers.


Lastly, what we were waiting for:


Click through
and compare and compare.



It was a little hard to get the shot, for this reason, but we did all right.



Which brings us about to the end, or a good enough place for one. Good night, all. Magnificent dreams.

Labels: , , , ,

20 Comments:

At 10:40 AM, Anonymous Anonymous said...

So, I need to know how to trisect an angle.

Good to see some snow, as I haven't been to the mountains in two years and nothing significant has fallen in Vancouver.

So the plans are to explore the chess possiblilities this Friday?

Sounds as though you need a friend with a video camera to attend all of the desired classes, is this an option?

Dressing for dinner is always an issue, currently dressing for work is becoming an issue. In the past it was business casual except for Friday as jeans day (not casual but jeans are permitted, a side note, I don't own any jeans). This dress code was changed recently, using your slow accident as an example, by one group after another being allowed to dress differently based on some reward. Finally the new rule, jeans allowed for everyone except Administration support and managers. Now the challenge is to get everyone to segregate the jeans from the casual and sloppy dressing.

Good luck in your classes!

 
At 12:16 PM, Anonymous Anonymous said...

Dad is crazy. January Blackcomb had a record snowfall, he just doesn't have any time to go :).

Further, I wear jeans and a hoodie to work every day, except when my hoodies are all in the laundry in which case I have a number of "nicer" shirts I wear, but paired with jeans again. I don't see an issue :)

But yes, it's nice to see snow :). I think I like the current metal tiger better than autumn. But I will wait until summer-ish to give final judgement.

I liked your math ancedotes. Perhaps I'll be able to use one on Gav next I see him (hopefully this evening).

Good to hear that you are getting out and about. Is people! Is music! Is excitement!

 
At 4:37 PM, Blogger BKF said...

Forgot to throw in a link to the origami website. Updated now.

Yes, the chance to see Ed Witten, and, umm, play chess is quite enough to make me forgive the useless creatures at the Princeton chess club all the occasions where their website has incorrectly advised.

You didn't hear it here, but "quite magically" some recordings of the CBC (!) Olympics broadcasts have become available. (Nice to see Pavarotti can still bring down a stadium at 70.) Speaking of which, the cbc.ca/olympics curling page tells a stunning story about today's Canada-Switzerland game: after a steal of 3 by the latter in the third end, one point for Canada, and then four consecutive one-point steals, ending with a 7-5 victory. Now there's a game I'd like to see. So that puts Gushue's rink into shared first with a 3-1 record. A good sign, but they haven't played Trulsen's squad yet. Think good thoughts. Is people... is music....

 
At 11:36 AM, Blogger BKF said...

So: to sum up. It's warmer here the past two days and the snow is all but gone. Gushue/Howard crush Trulsen much more convincingly than the score shows, then go on to lose to Finland and Italy (!). No sign of Witten, but I did find the chess club. They seem to like their bug, a game from which we'll have to try to lead them away.

 
At 11:25 AM, Anonymous Anonymous said...

So the Canadian currling teams are doing better than the men's hockey team. Maybe we should switch national sports. Or maybe bobsled?

I spent a few hours this weekend watching the Olympics, good news and bad.

This morning it is pretending to snow. At least a dozen flakes have past my window since 8.

 
At 10:19 AM, Anonymous Anonymous said...

I have been reading I, Claudius for the last week, and have been finding it very enjoyable. Just as good as the series, I think. :)

One question though. It seems that the first book encompasses the entire series. Whats in the second book?

 
At 7:52 PM, Blogger BKF said...

The first book ends exactly where Ep. 10 "Hail Who" does, with the soldiers finding Claudius hiding after the assassination of Caligula and proclaiming in him Emperor. The second book is the source for Eps. 11-13. Maybe seems lopsided....

But. You will see for example that Herod Agrippa does not appear by deed (perhaps by mention) in the first book. His life story and relation to the Imperial family in Rome is recounted at the beginning of "Claudius the God" and then we pick up the story. Aside from this quite a lot of uncinematic material from the second book was excised. For example there is a long description of Claudius' campaign in Britain, which pretty much gets forgotten in the series while we watch Messalina go south. There's also quite a lot of narration about various other projects Claudius undertakes as Emperor, ranging from trying to institute his three new letters to draining a lake.

 
At 5:51 PM, Anonymous Anonymous said...

Ahhh, I see. Ok, that's good to know.

This is an excellent book. I think it is probably better reading after seeing the series though, as the series is complicated enough to follow (poisonings, births, relationships, etc) without having to worry about reading and understanding as well :).

I'm liking all of the extra things on Caligula. Pointing out that before the fever he was cruel, but that afterwards he was cruel and insane :P. Lots of things in here that are chopped out for the series. Although in the book there isn't anything really written about him and Drusillus except for "They had relations. She died.", whereas in the series there is that rather gruesome scene...

 
At 12:26 PM, Anonymous Anonymous said...

I finish FF4 Advanced finally. You should play that game if you have a desire to play FF4 again, as it was very interesting.

The game is roughly the same. All new translations, but the Spoony Bard is still in. As you noticed it is the Hard Type, but I didn't think it was all that much harder. I finished Callabrena on the first try. Maybe it was luck :).

The interesting part comes after you get to the moon to kill Zeromus. You can switch party members to any person you have used, although Cecil must remain in your party. There is a new dungeon that is opened on Mt. Ordeal where you take new member of the party there (so with 5 new members you have to go at least twice) and you pick up new armour for them in the dungeon followed by a boss for each person, where you get a new and exciting weapon for them.

After that you kill Zeromus. Maybe you kill him a couple times (I killed him 3 times for reasons you will see...). After you kill him once the Lunar Ruins are opened, which is a 50-level dungeon on the moon. For each person that kills Zeromus, a new Trial opens up, which is a puzzle, a storyline, and a new piece of armour which gives them new abilities. Which also means you have to take multiple trips down the dungeon.

At each trial spot you there is a save location and a warp field so you can leave if you want, so while it's 50 levels, if you have 99 Cottages like I did, the only troubling part is that you sometimes have to leave and start again to clear your inventory :P.

The Trials were really cool, and at the end you kill Zeromus EG which was a _tough_ battle even at levels 71-75 (mostly 75, Cid was 71). No new ending though, which was lame, but the trouble they put into the rest of the new stuph was really nice.

So yeah, go grab yourself a GBA emulator and download it if you can. Or buy a GBA and borrow my copy. Good game :).

Thus ends my review. I'm now starting on FF: Dawn of Souls, which is the port of I & II for the GBA :D

 
At 11:05 PM, Blogger BKF said...

I guess I owe everyone quite a bit, like a little piece on classes I'm taking that should write itself. The short version is that I'm busy four days a week with it and otherwise at the moment having my life destroyed by chess, first the Linares tournament in Mexico/Spain (Topalov starts at -2 in the first half and then wins three straight, currently =2-4 with four games left) and second the U.S. Championship in San Diego. Catching up also with those Olympic games. Wonderful to see that curling gold medal. Breathtaking performance by Ben Heppner at the closing ceremony, a very fitting duality to Pavarotti in the opening. (Bocelli not bad either.)

I'll keep that in mind about FF4. Who knows, one day one year I'll come back to it for certain. You'll have to let me know how the newest port of I/II compares with Origins. For example, what they did with the soundtrack. I remember reading a little bit about the new port but I don't remember what I read. I actually did get a GBA emulator, to play FFTA. I thought it was largely adequate, although inferior to FFT in basically every way.

 
At 6:34 PM, Anonymous Anonymous said...

I felt the same way about FFTA. I wasn't _too_ upset when it got stolen along with my old GBA. :(

I'm presently playing FFII, which is one that I've never played all the way through before. I played the emulated NES version, but that was so long ago that I can't really remember it. I'm actually surprised by how good the story is.

FFI, when I get to it, I'll see if I can't write a comparison for you.

 
At 4:02 AM, Anonymous Anonymous said...

Though it's rare I move the conversation away from Final Fantasy, in this case I intend to. /Quelle Horreur!/

Balin, congrats on the fine dining and entertaining skills. I bet you wowed them with those numbers! I'm stealing the magic imp idea. You had already shown me the origami one, and I was suitably re-impressed. Origami is so much fun; that it's got ample math properties is bonus.

I hope your class with Wiles goes well - he is quite a bright lad, and I enjoyed reading about him immensely in the book you gave me on the Millenium problems. You must tell me if he is as charistmatic and well spoken as he seems in print.

You like curling? Did I know that? Hm. We shall have to palavar about that sometime. While I have never played the sport, I love watching it, and hope to someday try my rocks on a good piece of ice.

I hope all is well, and I look forwards to talking to you soon!

~Michael

 
At 10:32 AM, Blogger BKF said...

Hi everyone,

Sorry to be without contact for so long. Haven't been in the mood to blog for a little while. No pictures, no narrative, no zazz. More soon. Only one month to go until the term is over, a very sad state.


Michael -- I'm trying to recall what book I might have lent/given/pointed out to you. I think it's "Fermat's Enigma," which didn't quite make it to the Millenium Problem list, possibly on account of being solved slightly too soon. Andrew Wiles was the (chronologically last) hero of that story.

 
At 11:09 PM, Anonymous Anonymous said...

First of all, pardon my amateurish terminology - I am just a hobbyist, not a math student. Anyway...

Given a finite set S where the elements are incremented by 2 (for instance, { 3, 5, 7 } ) and an integer N, find the consecutive subset whose sum equals N.

For instance:
S = { 1, 3, 5, 7, 9 }
N = 8
solution = { 3, 5 }

This is a specialization of the NP-Complete Subset Sum problem, and I was wondering if there are any algorithms for solving it that are faster than the one for the Generalized problem (which is 2^(N/2) N, IIRC).

 
At 1:43 PM, Blogger BKF said...

Hi Anonymous,

I think the problem you stated is quite a bit easier than the general one. I think that if N has such a representation, you should be able to find it in O(N) time. Quite possibly you can do better. I don't work in this field so I don't know what's known about these kinds of and related problems. But here is what I can do:

You are trying to write N as ra + (T_(r-1) + rb)d, where a is the smallest element of S, T_(r-1) is the (r-1)'th triangle number, b is the smallest element in S used in the representation, r is the number of terms in the representation, and d = 2 is the difference between consecutive elements of S. So at the worst you just need to guess b (there are certainly at most N/d of these to try but we can cut it down more -- see next paragraph) and then find a quick way of estimating what r's will work. But T_(r-1) = r(r-1)/2, so an r that works is a root of the quadratic polynomial r^2 +r(a + 2b - 1) - N. Solve this by the quadratic formula and check if there's an integer root. (I think that if there is any solution, then taking b = first, second, ... elements of S, the first positive integer root you find should be a solution. But say if N is larger than the sum of elements in S then we can get tripped up.)

There is a lot of fat in this solution to be trimmed -- I mean if you play around with the quadratic formula you can eliminate some nonsense. For example the discriminant of this quadratic is (a + 2b - 1)^2 - 4N. So actually if b is smaller than sqrt(N) - (a-1)there cannot be a solution! (Of course, if there is no solution at all, then you can just run this and find that out in linear time, too.)

I want to note (maybe you've already seen this, but I want to note it just because I misread your problem at first) that this isn't any more general a problem than finding _any_ representation of N as a sum of elements of S. This is because a sum of elements of S is determined by its length and its barycenter (if we have N = sum (a + k_j d, then I think of 1/2 sum (k_j) as the barycenter), -- so we can replace such a representation by one symmetric of the same length about the barycenter. Now one can solve it in the same way as above (except there is a small correction: we may or may not include the barycenter, so there is one more case to check for each b).

Hope this helps.

--


Hi everyone,

So I've been telling everyone I meet how pleased I am with the spring-like weather: the buds on the trees appeared last week and now many of them are flowering (there's a certain tree I'm keeping watch on that hasn't yet, but soon, soon) and the skies are clear and gosh-this-isn't-at-ALL-like-home.... and... then this morning it was grey overhead and snowing. When I walked out of today's class, however, it was back like yesterday.

This weather-mania is subtly disturbing to a temperate-zone man like myself (forget about the rainforest). I got a copy of "Paradise Lost" which I think I should read sooner rather than later.

 
At 3:31 PM, Blogger BKF said...

Hi again Anonymous,

I have to clear up my flubs. Sorry.

First, the meaning of b in the first equation and what I said after is not the same. Let me stick with what I said after: b is the first element of S in our representation. Then we should have N = rb + T_(r-1)d, (because b + ... + b + (r-1)d has r b's and d times 1 + ... + r-1). So the quadratic you want to solve is r^2 + (b-1)r - N.

Second, the discriminant of this is (b-1)^2 + 4N. Here I have fixed my sign error that led me to erroneously conclude that there was a lower bound on b. But there certainly cannot be one: for example if N is just the sum of elements of S. I think in general we could really be looking for any b in [1, N/d].

There are some things it would be nice to know. For example, how many b's should work? Obviously we need r dividing N, and N >= r^2 (solve for b above); these are the only conditions, so certainly not more b's than divisors of N. For another, how often is (b-1)^2 + 4N a square? This is the same as asking how often 4N is the difference of two squares.

Sorry for the confusion.

 
At 4:02 PM, Blogger BKF said...

You know, it's maybe better to pick r than b. Because we know r < sqrt(N) and then b is determined. So that would be O(sqrt(N)) time. I had a hunch we should do better than linear. Good, good.

I'm done, now.

 
At 1:48 AM, Blogger oakleyses said...

oakley sunglasses, kate spade outlet, jordan shoes, prada handbags, oakley sunglasses, longchamp outlet, christian louboutin shoes, longchamp outlet, gucci handbags, michael kors outlet, replica watches, louis vuitton outlet, michael kors outlet online, louis vuitton, tiffany jewelry, burberry handbags, christian louboutin outlet, nike air max, uggs outlet, oakley sunglasses wholesale, polo ralph lauren outlet online, nike air max, michael kors outlet, nike outlet, uggs on sale, longchamp outlet, ugg boots, louis vuitton outlet, oakley sunglasses, prada outlet, tiffany and co, ray ban sunglasses, nike free, louis vuitton, ugg boots, michael kors outlet online, michael kors outlet online, cheap oakley sunglasses, uggs outlet, chanel handbags, michael kors outlet online, ray ban sunglasses, christian louboutin, ray ban sunglasses, louis vuitton outlet, burberry outlet, polo outlet

 
At 1:50 AM, Blogger oakleyses said...

ray ban pas cher, sac vanessa bruno, coach outlet store online, ralph lauren uk, michael kors, coach purses, hollister pas cher, converse pas cher, nike air max, nike roshe run uk, kate spade, hogan outlet, nike free run, true religion outlet, ray ban uk, north face, replica handbags, burberry pas cher, coach outlet, new balance, true religion outlet, nike air max uk, hollister uk, polo lacoste, nike air max uk, nike free uk, oakley pas cher, air max, true religion outlet, michael kors pas cher, nike air force, michael kors outlet, nike roshe, jordan pas cher, michael kors, vans pas cher, lululemon canada, nike tn, timberland pas cher, polo ralph lauren, nike blazer pas cher, abercrombie and fitch uk, louboutin pas cher, longchamp pas cher, north face uk, sac longchamp pas cher, sac hermes, guess pas cher, mulberry uk, true religion jeans

 
At 1:53 AM, Blogger oakleyses said...

converse, canada goose, louis vuitton, converse outlet, vans, moncler outlet, moncler, pandora jewelry, toms shoes, replica watches, lancel, hollister, montre pas cher, juicy couture outlet, swarovski, karen millen uk, moncler, ugg,ugg australia,ugg italia, moncler, links of london, pandora jewelry, canada goose jackets, louis vuitton, ugg pas cher, ugg,uggs,uggs canada, supra shoes, coach outlet, ugg, moncler outlet, hollister, pandora charms, canada goose, ugg uk, nike air max, louis vuitton, swarovski crystal, ray ban, wedding dresses, barbour uk, canada goose outlet, pandora uk, canada goose outlet, canada goose, barbour, marc jacobs, doudoune moncler, moncler, gucci, thomas sabo, juicy couture outlet, canada goose uk, moncler uk, canada goose outlet

 

Post a Comment

<< Home