Panic
©Sourendu Gupta

Contact me

This section

How science works

I'm often asked by my mystified family members "What do you really do? How does one actually work at science?" I can try to give you a view from the inside, but if I do it is also going to be one person's view. I work in theoretical physics. It has been decades since I handled electronic hardware. I've never handled animals in the lab after leaving high school. The last time I worked with chemicals and glassware in a chemistry lab was halfway through my bachelor's degree. I extract small signals from gigabytes of data, not petabytes, as high energy physics experiments will soon do. My tools of the trade are computers which compute at tens of teraflops, and hopefully, soon at petaflops. It would be useless to enumerate my experiences unless my methods have validity in a circle beyond my immediate colleagues. I believe they do. This is why I'm trying to set it down.

Bytes and flops A byte is roughly the amount of data represented by a single letter of the alphabet, or a single digit of a number. A gigabyte is a billion bytes. A petabyte is a billion gigabytes. A single flops is equivalent to adding two large number (usually 8 bytes long) in one second. A teraflops amounts to doing 1000 billion (trillion) such computations per second. A petaflops amounts to the ability to do a billion billion such computations every second.

The most important thing I do is to talk with others, read what they write, write down my thoughts as clearly as I can for them. Large parts of my day are spent communicating. Some of my friends and family are bemused by the amount of time I spend in a cafetaria drinking coffee and chatting. Others, the engineers who have morphed into managers, have an idea how vital this part of my day could be.

The main method of science is to criticize ideas and test them to failure. This is why reading, writing and talking is important. When I have an idea I first check whether it is obviously false. If it is not then I invest more time on it and explore it a bit. Often at this second stage I find something wrong, and I have to modify the original idea. After rounds of checking and modification, I find that the thing I was working on actually leads to roughly where I thought I wanted to go. Somewhere around this time I may begin to talk to some of my colleagues.

Usually the first colleagues I talk to are those who are in the best position to tell me that I am wrong. Talking is not always face to face: I use email, and, more and more often, skype (the importance of cheap communications to science is beyond measure). I do this, because I don't want to spend a lot of my life working on things which lead nowhere. If the work passes that test, I often give a semi-formal seminar to a larger set of colleagues. At this point, sometimes, new connections come up that require a closer look.

Criticism One of the first things that I had to learn when the next generation of my family became able to think for themselves was to couch criticism in encouraging terms. The years of training as a scientist are essentially years of training in how to react sceptically to everything you hear. When you carry this over to the rest of the world, it can give rise to hilarious fun or extreme hostility. Scientists are human too; and there are anecdotes about how they take constant specticism. Boltzmann was unable to take it, went into depression and committed suicide. Feynman famously submitted an article written on toilet paper to a journal. Some scientists famously write spoof papers in other subjects.

A large part of my lonely personal and unsocial work consists of following through a mathematical computation and transcribing parts of this to a computer program which would run for several months to years. The checking of computer programs is an academic speciality in itself.

The next step in the development of the work is to write up the paper and make it available to the world through the arxiv. This has two uses. One is that it established priority: the arxiv keeps track of when your paper was submitted. The other is to open up the paper to every person in the world who is interested in my subject. At this point I usually wait for a few days for comments to trickle in.

After a few days I submit my paper to a journal for peer review. This is often the tensest part of the lifecycle of every paper. Peer reviewing is anonymous. Reviews may reject a paper as being wrong, or worse, inconsequential. When a paper has passed that stage, it becomes part of canon. Its importance is after that judged by how useful it is: how many other ideas it has sparked.


Copyright: Sourendu Gupta ; Last modified on 28 Mar, 2024.