Tuesday, December 06, 2005

ahhhh: or why I became a computer scientist.

I will join in the collective sighs of relief at the completion of yet another deadline. I'd like to officially grumble about anonymous submission procedures, because one of the few pleasures of completing a paper is being able to brag about it afterwards :), and not everyone is as enlightened as the IACR.

Anyhoo, here's what I dispatched to the annual sausage convention. No pdf yet alas, because of paper release procedures and the like.

Lance had an interesting post about why he became a computer scientist: it appears that his story (and that of many in the comments) is of someone with a strong math background ending up in a math program and moving to a computer science program (one commenter even cites a complaint among math profs about losing their students to physics and CS). My story was slightly different: after reading Stephen Hawking's 'Brief History of Time' I was convinced that I wanted to be a physicist. My main interaction with computers was as a hacker, writing video games (in assembly code to boot) on my dinky ZX Spectrum.

But then I found a book in the library (actually it was the British Council Library in New Delhi, for any Delhiites among you all) that talked about Alan Turing and the Turing test. It was a short hop, skip and jump from there to Searle's Chinese Room argument, strong AI, and Godel's Theorem. AI was my hook into computer science, and remained my major interest up through graduation, although a complexity theory class by Somenath Biswas (and a thesis under his supervision) started drawing me away from the "fuzziness" of AI and into TCS itself. Through all of this, I remained untouched by the "math bug" (though I always enjoyed it), and for the last many years have been playing a desperate game of catchup !

From the point of view of training, I'd argue that a math degree is indispensable, and I fervently wish I had gone that route. My meanderings through AI did give me a more "romantic" notion of the field though, which is not necessarily a bad thing :)

Disqus for The Geomblog