Saturday, August 26, 2006


I called to make a reservation to go to SIGCOMM. The travel agent asked where I was flying to, and I said, "Pisa, Italy". I hear a lot of typing, perhaps a few "return"s, some "tab"s, several minutes pass and then she says, "That is pisa with a Z?". I say, "Uh, No. With an S." She says, "Oh, that explains it. I need to brush up on my I-talian". Type, type, type, ...

I saw the following claim: if x^3 +x is divisible by 4, then x is not odd. "Explanation" from a well known source. If x=1, that is true. Also for x=3, 5. So, true in general.

Reminds me of the book: Misteaks and How to find them before your teacher does, by Barry Cipra.

Wednesday, August 23, 2006

Wearing a wire in the subway

Residents of NY city quickly develop a sense for interacting with the subway. When doors open, you step into the car and without looking, get a measure of the people, seated or standing by the end doors or the centers, map the range of attire, shoes and bags, and get a feel for where to stand to minimize having to move for the crowd or an opening door in the stations yet to come. You also notice who has a wire peeking out from behind their clothing, presumably connecting their ears to a music player hidden in their bag; the wire used to be black and brazen in the past, these days it is white, sleek and mostly hidden.

This morning, I noticed something out of the ordinary. An unkempt middle aged man was seated, frowning, sometimes speaking to himself, front buttons of his shirt open to reveal fuzzy hair, some grey, some greying, and amidst that was a circular white patch sticking to his skin, and taped to that was a coneshaped black microphone! Someone was wearing an actual wire!

Friday, August 18, 2006

Truth in Refereeing

There is a nice article by John Chambers and Agnes Herzberg titled "A Note on the Game of Refereeing". See here. What fun. This is listed in John Chambers' page at Bell Labs with citation Appl. Statist., XVII:3 (1968), pp. 260-263.

Sunday, August 13, 2006

On the Defense

Sometimes I find myself in a situation where I have to defend algorithms-research from non-algorithmers (non-algorithmii?). This happens during work dinners, academic conferences, coffee breaks; while making a recruiting pitch; while conversating with colleagues at Universities and research labs; and with bosses or managers. Here are some standard attacks to prepare oneself:

-- [In TCS: Complexity Theorists] Algorithms researchers use a special technique for a specific problem, no general techniques for all problems of a class.
-- [In CS, Outside TCS: CS Systems] Algorithms research focuses so much on optimizing the answer! Much of the systems creativity is in formulating the key problems/issues and 80% optimal with 20% effort is good enough.
-- [Outside CS: Mathematicians, Statisticians] The theorems are stated so informally! Also, Algorithms researchers give glamorous names for simple things.


In "Smoke", a Paul Auster story, Harvey Keitel takes a photo every day of a corner in Brooklyn, for several years, rain, snow, heat wave, whatever. Still, you can not capture it all. Photography is addictive, so is New York, and NY photography NY1, NY2 even more so.

Portait of the Statue

It is difficult to paint a portrait of the Statue of Liberty because Liberty is nearly expressionless. Try painting a portrait without the quirks of a human face. Sigh.

These days I paint a little. I do esoteric paintings now, of drums, guitars, people in action, and colors. But when I was little, like 13 years old, I mainly painted portraits. Of people who were far away, but seemed to be in the psyche of everyone around me, thanks to the Time magazine that frequently made its way to my town. About 14 years after I left my high school (KV, IIT, Kharagpur), I went back there for a nostalgic stroll and happened to see some of my watercolors on display, jaded, distant, devoid in my mind of the drama when I had executed them. The way I like to tell the story, I remember when I stopped doing portraits. I saw a New Yorker depiction of stylized figures laid out in a haphazard way on the great lawn of the Central Park, figures heading off wherever, the tilt of their heads as couples listened to each other, loners intently focused away from others, all going somewhere. It was the evening, and hinted at the complexity of people's lives, taken in its entirety, and I lost my fascination with obsessing over eyebrows.

Friday, August 11, 2006

Being a Researcher on Many Fronts

Many algorithmers (algorithmus -> algorithmi?) work in different "applied" areas besides the core area of algorithm research. Besides STOC/FOCS/SODA/ICALP/.... and specialized conferences within algorithms like SPAA/SoCG/..., they also publish in SIGCOMM/SIGMOD/SIGGRAPH/.... It has its benefits, but also significant costs. For example, the deadlines for these conferences sometimes come bunched together or they are spread out through the year with narrow gaps making one lurch from one deadline to another. Also critically, one has to pick amongst equals while traveling to conferences when they overlap. I am in a quandary: I have papers in ESA and its satellite workshop (thanks to my coauthors Gagan Aggarwal and Jon Feldman, my first paper on Auctions) in Zurich Sept 11 -- 15, I am giving a talk at SIGCOMM Minenet (detailed program here) in Pisa Sept 11 -- 15, and of course I want to be at VLDB in Korea, Sept 12 -- 15. Alas.

Monday, August 07, 2006


A laconic headline some days ago was: Who spiked my sample? Awesome. I am sure you can guess the original context of the headlines, but I am an algorithms geek and the first thing that occurred to me was, what a cool title for a paper!

Friday, August 04, 2006

What is New

Graham Cormode changed jobs this week. He carries his gentle humor from the old corridors of Bell Labs, Lucent (and Alcatel) to the new, but venerable AT&T Labs. His humor once in a while bursts out in Crossword puzzles, a Lemmings walk or the paper on theory conferences.