[FoRK] Surviving graduate school (from a physicist's perspective)

Joe Barrera joe at barrera.org
Tue Feb 8 22:02:43 PST 2005

Surprisingly applicable to CS. e.g. "never means three months"
And, of course, "don't make it better than necessary", suprisingly close to
"worse is better".


Surviving graduate school

Critical Point: February 2005

If you want to know how to do research, you need to learn some valuable 
yet intangible lessons. Robert P Crease seeks your tips and advice for 
budding physicists

Anyone who has done a PhD knows that you go through an educational phase 
transition. You are no longer instructed by others, but instead teach 
yourself. You cease acquiring basics and begin assimilating values and 
behaviours that allow you to participate in the scientific way of life. 
It is an apprenticeship that cannot be condensed into a set of rules.

Survivors of graduate education, however, come to appreciate the fact 
that certain valuable maxims distil much wisdom. I don't mean vague 
injunctions like "persevere", "adapt" or "network"; they are good and 
true but possess all the functionality of your horoscope. Nor do I mean 
the practical advice on shelter, food and companionship ("choose a 
romantic partner who knows PowerPoint") that is generic to graduate 
life. What I mean are useful maxims for the blossoming physicist. Let me 

*Never means three months*

I once learned an important lesson from Andrew Kevey, who used to be 
chief spectrometer engineer at the High Flux Beam Reactor (HFBR) at the 
Brookhaven National Laboratory in the US. While learning the ropes at 
the lab's previous neutron source - the Graphite Research Reactor - 
Kevey's advisor had told him to design and build a small calibrated 
turntable on which to mount crystals for a spectrometer. Kevey asked how 
much load it would have to bear. A few ounces, came the reply - never 
more than a pound.

Three months later, the advisor - whose research programme now involved 
magnetic properties - handed Kevey a 50 pound magnet and asked him to 
install it on the turntable. Kevey protested, citing the advisor's 
earlier remarks about the load.

"You idiot," the advisor exploded, "don't you know that, in physics, 
never means three months?"

This important maxim, Kevey explained to me, is a condensed way of 
saying be prepared for rapidly changing demands and revised expectations.

*Build in the centre of the room*

Brookhaven physicist Laurence Passell discovered a related lesson while 
a graduate student at Berkeley. His advisor, who was conducting 
experiments in low-temperature physics, had instructed Passell to 
prepare an experiment in a new basement laboratory. When Passell first 
walked into the still-empty lab and began to set up, he ran into a more 
experienced graduate student who cast a disdainful eye on Passell's 
first efforts. "Let me give you a piece of advice," the student said. 
"Always start building your experiment in the centre of the room."

"I laughed at the time," Passell recalls, "but the advice was extremely 
useful." For doing physics, he explains, means embarking on a random 
walk. Regardless of the direction in which you set out, you rarely know 
where you will end up. And experimental divagations expand radially. So 
lesson two is: give yourself as much space as possible.

*Don't make it better than necessary*

Passell learned another important lesson at Berkeley: don't make your 
equipment better than it needs to be. "The best piece of scientific 
apparatus is one that falls apart the day after you finish using it," he 
explains. Violation of this principle cost Passell a research programme 
at the HFBR in the late 1960s. He was planning to use a new neutron 
spectrometer that was being built by a perfectionist retired navy 
officer who was aiming to build the best of its kind in the world. But 
as Passell recalls, the officer never asked himself "Is it good 
enough?". Instead, he asked "Is it as good as I can make it?".

"The answer, of course, was always no," says Passell.

At one key juncture, for instance, the officer became dissatisfied with 
the crude but effective shielding method that had to be realigned each 
time the beam direction was shifted. As the spectrometer drum turned to 
change the angle of the neutron beam, a series of wedges would lift up 
out of the way of the beam before dropping back down on the other side. 
Unhappy with this rudimentary design, the officer instead began 
installing an elegant system in which a rotating cone block inside the 
drum automatically realigned all the shielding.

Passell looked on, horrified, as the device tore up the spectrometer's 
budget and wrecked its construction schedule. His research programme was 
soon terminated.
"An experiment is perfect when the equipment is just good enough," he 

*Over-extend yourself*

A fourth key maxim that I have often heard scientists mention is: if you 
really know what you are doing, you should not be doing it. For if you 
know that much, someone else has probably already done - or is about to 
do - what you are intending. Your results, in other words, will soon be 
obsolete. As Fermilab's former director Robert Wilson liked to say, 
"Something that works right away is over-designed, and consequently will 
have taken too long to build and will have cost too much." Doing science 
effectively and efficiently requires you to over-extend yourself to the 
point where some things will not go as planned.

This principle is, of course, a secret. You won't find it in textbooks, 
nor will you hear science administrators publicly endorse it. Their job 
is to know it and cover it up.

*The critical point*

Shacheenatha Jha, a physicist from Case Western Reserve University, 
liked to tell his graduate students that "When you start telling me what 
you want to do, instead of me telling you what I want you to do, you are 
ready to graduate." Acquiring the requisite autonomy to do that cannot 
be captured in a finite set of rules.

Nevertheless, every scientist who has successfully endured the process 
learns a set of maxims for effective action. These ought to be compiled 
and passed on for the benefit of future generations. I therefore invite 
you to send me the most important one you learned in the course of your 
PhD training or other learning period, along with an example. I shall 
discuss your comments in a future column.

.... What is the most important thing you learned in graduate school? 
Send your advice to Robert P Crease at the address given below, or by 
fax to +1 631 632 7522, or by e-mail rcrease at notes.cc.sunysb.edu
About the author

Robert P Crease is chairman of the Department of Philosophy, State 
University of New York at Stony Brook, 213 Harriman Hall, Stony Brook, 
NY 11794-3750, US, e-mail rcrease at notes.cc.sunysb.edu.

Kalashnikovs but no houses
Women at the double, march
No food for the spouses
They wait for the US drop
Russians sit back and laugh

More information about the FoRK mailing list