:: rushabh ::

Blog is dead; Long live Blog

This is the last entry in my blog at redefine. I'm moving the blog to blogger. Why? Read on in the post at blogger. Thanks for reading and hopefully you'll update your bookmarks / readers to point to the new blog.

Blog is dead; Long live Blog

Beautiful Code

I just got my hands on a copy of Beautiful Code. I've been unreasonably excited about this book, where the central idea, AFAIK, is that the editors asked the best hackers in the world to show-and-tell their best work. Its got articles by lots of people, some of whom have written languages (Kernighan) or libraries (E. R. Harold) that I use every day. In fact, I got so excited by this book, that I put it on my safari bookshelf last night and started reading it, even though I knew I'd get the dead-trees version soon (and I got it today). Oh well. Furthur updates as I keep reading it.

Done with Harry Potter...

... getting on with the rest of my life.

]]>

Fireworks!

Originally uploaded by Rooosh.

I know, I know, I haven't blogged forever. Been really busy at work, etc, etc. But I did get a chance to take pictures from the office (yes, I was at work during fireworks) and they actually turned out reasonable. Take a look!

Beginning Python

Alex Martelli's python talk is focused towards people who know programming in some regular imperative / oo language such as C, C++ or Java and who are absolute or just about absolute beginners at python. (I fall into the latter category). Its a great talk, the Q/A at the end is mostly avoidable, but otherwise, its a great talk to listen to and follow along. Note that the video does not do full justice to the slides - I found it worthwhile to have the slides open on the side and follow along. Enjoy!

Mii

Thanks to Joystiq.com

Fermat

Some of the hardest problems in the world can be spelt out in a couple of sentences. Pierre de Fermat's problem was no different. He conjectured that there were no solutions to the following equation:

xⁿ + yⁿ = zⁿ where n > 2
This is seemingly wierd if you recall that Pythagoras's Theorem states:

x² + y² = z²
and has an infinite number of solutions (the first one that comes to mind is 3, 4, 5).

Fermat claimed that his conjecture was true and that he even had a proof for it. Being Fermat, however, he didn't mention the proof itself, leaving a puzzle that remained unsolved for 350 years.

I first got into Singh's books when I picked up and read The Code Book, which I thoroughly enjoyed.

Fermat's Enigma looked interesting and so I picked it up. Unfortunately for me, the bookseller had cats that were allowed to roam freely in the store and so I could really get past a couple of pages of the book without starting to sneeze like crazy and get puffy eyes and such. (Did I mention that I really, really don't like cats?) Anyway, so the book was put on hold for a while until a friend loaned me his copy (sans cat hair), and I read it over the weekend.

Singh does a really good job of writing a non-fiction book that manages to keep the reader's attention. It reads almost like a mystery novel. Singh glosses over all the hard math (a good thing, else nobody would understand anything - he repeatedly points out how there about a few dozen people in the world that understand some of the methods that Andrew Wiles used). He goes into a bunch of side-stories about various related problems and their solutions and manages to tie everything back in with Fermat's Last Theorem. This is a fantastic book and I'd highly recommend it, regardless of your mathematical inclinations.

Quoting Larry Wall: Long term lazy vs Short term lazy

I'd never imagine I'd requote something from Larry Wall (via Coding Horror by Jeff Atwood), but as Atwood points out, I don't have to respect Perl in order to respect Larry Wall. Enough said. Lets move on to the quote:

I think that the biggest mistake people make is latching onto the first idea that comes to them and trying to do that. It really comes to a thing that my folks taught me about money. Don't buy something unless you've wanted it three times. Similarly, don't throw in a feature when you first think of it. Think if there's a way to generalize it, think if it should be generalized. Sometimes you can generalize things too much. I think like the things in Scheme were generalized too much. There is a level of abstraction beyond which people don't want to go. Take a good look at what you want to do, and try to come up with the long-term lazy way, not the short-term lazy way.

This

Is

Brilliant.

All programmers (and aspiring programmers and anyone remotely interested in programming or engineering for that matter) should write this down and read it once a day till it has been burned into your brain. I wanted to go get a tattoo of that quote. But that only held for the first time.

The best city EVER!

From the BBC article:

A truck driver transporting the fueselage of a plane through Bombay got into a bit of trouble: He came across a bridge that his truck/plane wouldn't get under. Given situations like these, rational people would:

Find an alternate route

Find an alternate route?

But not our erstwhile driver. He decides to leave the truck / plane where it is and leave. Like literally. Leave.

So now Bombay has had (for the past 5 days) a plane sitting on a road, and nobody really knows what to do.

I love Bombay.

New forms of text entry

Keyboards suck. We all know it, we admit it and yet we've been stuck with them for ages. We've gone through multiple computer revolutions and yet the only mainstream thing we invented for (human) IO with the computer was the mouse.

There have been lots of very smart people thinking about lots of very interesting ideas around how to make human computer interaction easier. We type with our fingers, which are capable of phenomenally complex gestures, yet the only interface we have with the keyboard is this silly binary key-up / key-down interface.

More importantly, the way languages are structured, not all letters / words are just as likely after a given letter / word. Researchers have (post invention of the type-writer / keyboard) come up with probabilistic models of languages, which allow them to assign subword probabilities to letters and sub sentence probabilities to words and so on.

So we have a probabilistic language model and a finger (or ten) that are capable of much, much more than pressing buttons. What do we do with them? One idea: Dasher.

The talk is long, but very, very worthwhile. And once you're done, make sure you download it and give it a whirl. Cars are fun to sit in, but a whole new level of fun to drive.

From NY to London

The google maps way.

# 24: Swim across the Atlantic Ocean (3462 mi)

./demotivate

Aaron Swartz puts things in perspective
.

The universe is a bunch of random particles shooting through space following a handful of simple laws. Through completely random and unintentional properties, some of those particles bounced together to form you. But, I swear, it was a total accident.

The morning fix

Coffee!!
Originally uploaded by Rooosh.

Seattle is the best place in the country (on the planet?) for caffeine. This cuppa is from Espresso Vivace, which is one of my two favourite coffee shops in the city (and I live walking distance to both of them! Yay!)

Human Computation

This talk is about a year old, but very, very worthwhile. Luis con Ahn on using humans to solve problems that computers find very difficult (image classification) in an interesting way. One outcome of this problem solving is that you can use this data to generate a set to train better image recognition software (training computers). Crazy stuff, but what I find the most fascinating is that Luis figured out how to make this seemingly mundane task fun.