Tuesday, June 30, 2009

(unexpected) definitions.

earlier today i was browsing the arXiv. upon glancing at the title of this preprint, i suddenly wanted to know what a "supermanifo1d" is.

so i looked up the wiki:

it wasn't what i expected [1]:
"supermanifo1d" isn't the opposite of "submanifo1d."

instead, the "super" part refers to super-symmeτry.

oh well. one more thing to remember, then.



in two separate occasions this month, i've encountered de 9iorgi classes. i may start thinking about them, a little.

[1] admittedly, during the first split-second after reading "supermanifo1d," i imagined a torus .. er, wearing a red cape with an S on it. i subsequently thought mathematically and imagined something more technical.

Friday, June 26, 2009

and so it goes.

another weekend,
another wedding,
another work stoppage.

it's just as well. the technical part of the paper writeup is almost done, but it seems like the background and references are going to be more of a headache than i thought.

just now, while browsing for a quasi-reference, i found 2-3 more papers of some relevance that i should read, just in case. unfortunately, none of them are available electronically.

as the old saying goes: one never finishes a paper; one gives up on them.

Wednesday, June 24, 2009

screen ≠ paper: the writing continues.

yesterday i was LaTexing for about 4 hours, and the day before was similar. today i looked at the printout.

for a research note, the computations look gory.

i've even split up several proofs into lemmas because, eventually, i was sick of looking at the endless eqnarray*s of estimates. maybe this will make it more readable, or at least, no less unreadable.

all of this worries me. if i wrote this and i'm sick of these technicalities, then i wonder how forgiving a referee will be ..?



the printout is less than a day old, and every margin is already half-full of edits and comments. screen by screen, it's not easy being consistent with notation.

i'm also getting forgetful.

the oldest proofs look the most like the work of a stranger:

months ago, i worked out why one proof has taken the form that it does. looking through it now, i can't remember why. something goes wrong, but i can't remember exactly what it was.

i mean, the argument is still correct, but not exactly intuitive.



there's not much exposition yet, either.
that's going to be its own headache.

Saturday, June 20, 2009

mistaken identity (concerning preprints)

a week ago, i met T.Z. in barcelona. before, he had only been a name to me, in several papers and preprints in the ana1ysis on metri¢ spaces.

he had told me about a recent work of his and kWild, but the conversation was somewhat confusing, but now i understand:

i thought he meant this preprint,
which i already knew about.

instead, he meant this one,
whose existence i've learned, just now.

i'm an idiot.
had i known, i'd have browsed it overnight and interrogated him about it!

i claim no deep understanding about this work, but this seems to me a recovery of good properties of Sobo1ev functions (in the spirit of m0rrey and resheτnyak and others).

briefly, t.z. and kWild study ah1fors(-david) Q-regu1ar metri¢ spaces which admit (weak) p0incare inequa1ities. in contrast to doubling, Q-regu1arity is more quantitative -- at least, to my dim mind -- and the parameter Q serves as the "right" critical exponent where there is a change in behavior to Sobo1ev functions.

so the work makes sense.

it doesn't escape me that ¢heeger's (measure) conjecture holds true for such spaces [Ch, Thm 13.12]. if one further assumes that such spaces can be isometrical1y embedded into euclidean spaces, then one might expect some amount of geometric rigidity [Ch, Thm 14.12]. subsequently, analysis on this subclass of metri¢ spaces would really boil down to .. well, ana1ysis on euclidean spa¢es.

then again, it's much to assume isometri¢ embedding. the work remains nontrivial. all i'm saying is that it matches with what one would (conjecturally) expect, in the euc1idean case.

Friday, June 19, 2009

thoughts while writing.

sometimes i think i write completely obvious things.

in fact, this research note that i'm writing feels like one big, obvious observation .. maybe an exercise for a graduate student [1]. most of the details boil down to basic facts about sobo1ev spa¢es and follows an outline present from a paper in 1966.

on a related note, i really like the book by εvans and gariεpy. if anything, it helps convert my convoluted proofs of basic lemmas into proof sketches consisting of a few citations.


[1] this is not far from the truth. i did work these details out when i was a 2nd year ph.d. student .. which only confirms one thing: never trust me to write up anything.

Wednesday, June 17, 2009

neither here nor there.

on one hand, whenever i am away on conference i wish i were home instead, working.

there are two reasons:
  • some talks make no sense -- i don't have the right background to appreciate them -- and i'd rather do something else.

    it might as well be work;

  • some talks are wonderful. then i realise that i should work harder, prove better things. then i wish i could use the time to work, instead.

on the other hand, when i am home i wish i were away,

visiting colleagues, possibly sharing ideas,
meeting new people and learning new things ..

.. then again, being away doesn't guarantee any of that. i've been in spain for two weeks, i haven't had any good ideas or questions, and subsequently i've been mostly silent.



i feel tired. it's not easy being a stranger in a strange land, day after day. sometimes it is easier to be frustrated at things that you know, rather than be uncertain about things of which you are ignorant.

in the end, i suppose that i am an american, after all.


this and next weekend i'll be indisposed, attending weddings. a friend is visiting town, in between. so i suppose that i can be resigned to being neither away or working.

Tuesday, June 16, 2009

somehow, over the years, i've associated conferences with finns. my advisor was finnish, and i've been to finland ..

[counts]

.. 3 times for conferences and summer schools, though each trip was not long.



when i think about it, i'm not a terribly good postdoc.
i have few ideas and few papers.

maybe there's enough time left in pittsburgh to change that.

Saturday, June 13, 2009

last vs. this weekend; an aside (with links to wikis)

the summer school is over:
two more days before the conference begins.

last weekend was a frustration of mathematical unknowns. i've learned my lesson: no more of that. in fact, i learned more.

i just don't know enough to attack the problem.

the first time -- which made my doctoral thesis -- was more a stroke of luck than anything. rather, the nature of the problem made it possible to treat the case purely as a null set. that was structure enough.

now, the problem is more subtle. the actual mass distributions (from singu1ar me@sures) are much more relevant now, and there are fine structures to build or detect.

in fact, it should really boil down to derivaτive measures of βV functi0ns. i don't know very much about these. when i learn more, then maybe i'll come back to this conundrum of a problem.

at any rate, my girlfriend is visiting, this weekend. there's a whole Old World city to explore and sample.

there is time for work, too.
it's a pleasant thing to date a fellow academic.

they understand when you have to -- er, want to? -- work all the time, because often they're in the same shoes.

Wednesday, June 10, 2009

barcelona, one week later (updated; also, post #600 ..!)

there are no afternoon lectures at the summer school today. they've set aside the time for a cultural excursion to la sagrada familia, an unfinished work of gaudí.

admittedly, i'm tempted to renege on my promise, stay in the dormitory, and instead, work on the draft of a long-promised research paper. lately i've felt like i've had no time to be productive.

then again, i remember what happened last weekend.

maybe it's time to visit the city center of barcelona.



epilogue: it was worth the trip. i learned and relearned the following facts:

  1. never count on groups of mathematicians arriving on time.

    a colleague and fellow train companion was worried that we would miss the second tour group, at 17:10. when i pointed to all the other summer school participants, equally late, on the same train car, she was not wholly convinced.

    it wasn't until reaching the site that we promptly joined the first tour group, which was 20 minutes late.

  2. an (empirical) symmetry principle. [web photo]

    from wikipedia: The Catalan architect Antoni Gaudí made extensive use of catenary shapes in most of his work. In order to find the best curvature for the arches and ribs that he desired to use in the crypt of the Church of Colònia Güell, Gaudí constructed inverted scale models made of numerous threads under tension to represent stones under compression. This technique worked well to solve angled columns, arches, and single-curvature vaults, but could not be used to solve the more complex, double-curvature vaults that he intended to use in the nave of the church of the Sagrada Familia. The idea that Gaudi used thread models to solve the nave of the Sagrada Familia is a common misconception, although it could have been used in the solution of the bell towers.




in other news, this is my 600th post on this blog. (yes: i've been complaining about mathematics and academia for that long.)

Monday, June 08, 2009

what i didn't do, this weekend.

friday: after hearing about various sights to see, i told myself (and others) that i would visit the city center of barcelona, either on saturday or sunday, but not both days. [1]

saturday: i woke up, obsessed. i told myself that i would visit barcelona on sunday. instead, i spent the day pondering and pondering ..

sunday: i told myself that i would visit barcelona next weekend, when my girlfriend will come to visit. so i tried more ideas ..



today: during the coffee break, a colleague of mine asked me how my weekend went. i tried to think of something reasonable to say, but apparently i took too long.

"you worked all weekend, didn't you?" he said, not asked.

he said he could tell from the look on my face, so there was no point in pretending. afterwards we briefly discussed the addictive, crazy nature of mathematical research ..



[1] you see, i was motivated by one of the courses. i wanted to attack one of those conjectures that, from experience, easily becomes an obsession.

for my own good, i think i'll have to set aside this problem again. it leads to nothing but trouble.

Sunday, June 07, 2009

not that kind of summer school.

lately i haven't had too many opinions or observations. when not attending lectures, i've been puzzled by some of my own research inquiries or encountered dilemmas in writing.

since i became interested in mathematics, i've attended three summer schools [0]. this is the fourth.

the term 'summer school' is almost a misnomer. a better name would be a 'lecture series' or a 'seminar.' when i say 'summer school,' it resembles too closely undergraduate coursework of a dim nature; there are even lacklustre films about this latter kind of thing.

what comes next is probably redundant, for those of you who've attended these kinds of things. maybe i'll post something more interesting later.



by the second day of the summer school, there were bound copies of lecture notes for the courses [1]. shortly after their appearance, proofs during the courses became sketches of proofs.

that sounds condemning at first, but i don't mean it that way.

these courses are not in your standard mathematics curriculum. the invited lecturers are presenting advanced topics in QC, PDE, and GMT.

the audience does not consist of wide-eyed, first-year students, either. we all know what a proof is, how hard rigorous details can be. we're not here to be instructed. instead, we're here for the ideas. if the ideas are good -- and certainly they are -- and if we are sufficiently interested, then we can roll up our sleeves and work out the details on our own.

intuition and insights are rare commodities in this business. that's why we're here.



that's the amazing thing about a summer school. like a conference, many researchers are gathered in a single place and share new insights, either by giving talks or discussing ideas between talks. the collected potential, within a single building, is staggering.

the difference is the pace: at a summer school, one isn't bombarded by 20- or 30-minute research summaries, glutted into a short span of days. over a series of lectures at a summer school, one sees several ideas as developed and discussed by an expert, who willingly shares a wealth of intuition from his/her own months of careful work and thought.

so yes, it's like a seminar in the sense of pace, but like a conference in its setting.

..
put that way, i'm probably not taking this opportunity to the fullest extent that i can. even after all these years, i still find it difficult to meet mathematicians and share my ideas.



[0] maybe 4, going on 5. in order, they are --

2003: park city, ut,
2004: pittsburgh, pa & jyväskylä, finland,
2006*: ann arbor, mi,
2009: barcelona, spain.


[1] well, 3 of 4. for the remaining course, the notes are printed separately and will probably go online.

Wednesday, June 03, 2009

short post: barcelona, day 1. [0]

  1. i still haven't fully read the henc1-ma1y preprint. at best, i printed it out this morning, browsed through it, and became confused about (genera1ised) 1inking num6ers.

    then again, i had reached that section of the paper. five minutes later, it was time to register for the summer school at the ¢rm.

  2. summer schools, like conferences, are like reunions: i've reached a point in my life where i see familiar faces more often than i meet new people.

  3. it is hard to judge a mini-course based on the first lecture. for two of the courses, some of the background i know, but i have no sense of how they will proceed.

    the third course today was like an old acquaintance. it's a subject i've learned before, briefly -- i think many in the audience can say the same -- but enough time has passed that i am curious again and it does me good to see it again.

    this seems to me a course which is more difficult for the lecturer than the other two courses. for the latter, they are discussing their own recent research and are masters of their domain. as for the former, the subject is established (regularity theory of de 9ior9i-na$h-m0ser), so the lecturer has to address both extremes:

    * the newcomers who know nothing at all, even the basic objects such as sobo1ev spaces;

    * the seasoned full professors who know the theory well and, given some days, could prepare lectures on the same content and range.

    maybe i'm too young to master this dichotomy, but it seems difficult to give lectures which are both clear and interesting.

[0] technically, i arrived yesterday, but the summer school started today.

Monday, June 01, 2009

a non-arXiv preprint.

interesting.

Ja¢obians of Soboleν homeom0rphisms
S. Henc1, J. Ma1y - KMA Preprint 2009/307 [PDF]

Abstract. Let Ω be a domain in Rn. We show that each homeomorphism f in the Sobolev space W1,1loc(Ω;Rn) satisfies either Jf ≥ 0 a.e. or Jf ≤ 0 a.e. if n = 2 or 3. For n > 3 we prove the same conclusion under the stronger assumption that f is in W1,sloc(Ω;Rn) for some s > [n/2] (or in the setting of Lorentz spaces).

commentary, forthcoming;
i haven´t read the preprint yet.