Wednesday, December 20, 2006

Chris Sander


Chris Sander
Chris Sander

Trained as a theoretical physicist, Chris Sander, Director of Memorial Sloan-Kettering Cancer Center's Computational Biology Center and Chairman of the Sloan-Kettering Institute's Computational Biology Program, knew early in his career he wanted to do more than compute the behavior of elementary particles. After a bold move from physics to biology, he helped to develop the field of computational biology, which aims to use mathematical algorithms and information systems to simulate the behavior of molecules, cells, and organisms, using these simulations to make useful diagnostic and therapeutic predictions.

While an undergraduate at the University of Berlin, I was intensely engaged in the study of theoretical physics and mathematics. Yet I felt that analyzing life, the living system on this planet, was a more fascinating and challenging scientific problem than studying the world of elementary particles. To explore this potential change in career direction, I sought the advice of Max Delbrück at the California Institute of Technology, a physicist originally from Berlin who was one of the founders of molecular genetics.

After visiting friends in Texas, I hopped on a Greyhound bus to Los Angeles, made my way to Caltech, found Dr. Delbrück's office, knocked on the door, and with a dash of chutzpah, asked the Nobel laureate if he had a moment to chat with an aspiring graduate student. Dr. Delbrück described the areas of theoretical physics that might be relevant to biology in the future -- information that was in the forefront of my mind as I entered the graduate physics program at the University of California, at Berkeley, in 1967.

Wanting to move from the theoretical physics of my PhD thesis to the theoretical biology I had dreamt of, I made my second pilgrimage, this time to see Manfred Eigen, a Nobel Prize-winning chemist who was studying biological evolution in Göttingen, Germany. Dr. Eigen surprised me by explaining that the field barely existed, but he did point me to three mathematical biology research problems: a theory of the immune response, neuronal mapping in the brain, and protein folding. So I packed my bags and I moved from the University of Heidelberg to the Weizmann Institute of Science, in Israel, where I began work with Shneior Lifson on the prediction of three-dimensional protein structures.

Enter the third motivator of my career: the first completely sequenced genome -- no, not in 2000, but in 1977! In that year, I saw an amazing paper in the journal Nature from Fred Sanger's group in Cambridge, United Kingdom, which included two entire pages filled with 5,375 letters, all As, Ts, Gs and Cs, representing the genetic blueprint of a small virus. I walked down the hall to ask my friend Georg Schulz, "With this kind of cryptic information coming from genomes, won't biology need computational science to decipher it?" His answer was yes, and I spent the next 23 years of my professional life preparing for the day, in the year 2001, when the 3.5 billion letters of the human genome finally became available. In the process, I helped to develop the field now known as computational biology.

The real value of computational science, when applied to any system, is to predict what's going to happen next. Weather forecasting is an example. There's an enormous amount of data collected about the weather, but the data, by themselves, are unintelligible. What's required is the application of the appropriate mathematical equations embodied in a software system, which, using the data, allows one to compute tomorrow's weather.

Applied to cancer biology, we want to be able to predict, for example, if a cancer will go from a nonaggressive to an aggressive form, or more importantly, to predict accurately the consequences of possible therapeutic interventions. The goal is to have an impact on human disease, and to do this you have to work in collaboration with physicians. In 2002, Harold Varmus presented his vision of Memorial Sloan-Kettering as the perfect environment for this -- a place with open doors between basic and clinical research, where close collaboration is encouraged.

My first action at Memorial Sloan-Kettering Cancer Center was to start the Computational Biology Center (CBC) and its Bioinformatics Core Facility. The CBC's researchers and engineers are devoted both to basic science and to the goal of developing diagnostic and therapeutic tools that help improve the lives of people affected by cancer. We often collaborate with researchers in the lab and in the clinic to translate data -- data such as the molecular profiles of cells and tissues, the billions of letters of genome sequences, and the functions and structures of key genes -- into biological insights and prediction tools. And the Bioinformatics Core, ably led by Alex E. Lash, provides internal bioinformatics training, collaboration, and infrastructure support.

One concrete example of the practical uses of computational biology is the work we have been doing with Howard I. Scher, Chief of Memorial Sloan-Kettering Cancer Center's Genitourinary Oncology Service, and Francis M. Sirotnak, Member Emeritus and Head of Sloan-Kettering Institute's Laboratory of Molecular Therapeutics. The idea is that cancer cells, like any system that recovers from a round of major damage, might be especially sensitive after a first round of therapy. With this concept in mind, we are aiming to prevent the development of aggressive prostate cancer by looking at the molecular profile of prostate cancer cells after androgen removal in mice, using DNA chips provided by our Genomics Core Laboratory. We use computer software to find needles in a haystack -- the perhaps tens of genes, out of tens of thousands, that may be a characteristic signature of how prostate cancer reacts to such therapy. We hope this will lead us to an Achilles' heel to target to avoid recurrence. It's a long-term effort but the idea is to arrive computationally at the best therapeutic intervention.

Overall, what's been most rewarding for me during my short time here is the opportunity not just to predict the behavior of biological systems, but hopefully to help improve the quality of people's lives. The dream that started with Max Delbrück's advice is now within reach.

Saturday, December 09, 2006

Super Computing

 Logo
The TOP500 project was started in 1993 to provide a reliable basis for tracking and detecting trends in high-performance computing. Twice a year, a list of the sites operating the 500 most powerful computer systems is assembled and released. The best performance on the Linpack benchmark is used as performance measure for ranking the computer systems. The list contains a variety of information including the system specifications and its major application areas.


NCSA Home
The National Center for Supercomputing Applications (NCSA), one of the five original centers in the National Science Foundation's Supercomputer Centers Program, opened its doors in January 1986. Since then, NCSA has contributed significantly to the birth and growth of the worldwide cyberinfrastructure for science and engineering, operating some of the world's most powerful supercomputers and developing the software infrastructure needed to efficiently use these systems (for example, NCSA Telnet and, in 1993, NCSA Mosaic™, the first readily available graphical Web browser). Today the center is recognized as an international leader in deploying robust high-performance computing resources and in working with research communities to develop new computing and software technologies

Blue Gene




Blue Gene is an IBM Research project dedicated to exploring the
frontiers in supercomputing: in computer architecture, in the software required to program and control massively parallel systems, and in the use of computation to advance our understanding of important biological processes such as protein folding.

The full Blue Gene/L machine was designed and built in collaboration with the Department of Energy's NNSA/Lawrence Livermore National Laboratory in California, and has a peak speed of 360 Teraflops. Blue Gene systems occupy the #1 (Blue Gene/L) and #2 (Blue Gene Watson) positions in the TOP500 supercomputer list announced in November 2005, as well as 17 more of the top 100.

IBM now offers a Blue Gene Solution. IBM and its collaborators are currently exploring a growing list of applications including hydrodynamics, quantum chemistry, molecular dynamics, climate modeling and financial modeling.

SDSC - San Diego Super Computer Center

Founded in 1985, the San Diego Supercomputer Center (SDSC) enables international science and engineering discoveries through advances in computational science and high performance computing. Continuing this legacy into the era of cyberinfrastructure, SDSC is a strategic resource to science, industry and academia, offering leadership in the areas of data management, grid computing, bioinformatics, geoinformatics, high-end computing as well as other science and engineering disciplines. The mission of SDSC is to extend the reach of scientific accomplishments by providing tools such as high-performance hardware technologies, integrative software technologies and deep inter-disciplinary expertise, to the community.

SDSC was founded with a $170 million grant from the National Science Foundation's (NSF) Supercomputer Centers program. From 1997 to 2004, SDSC extended its leadership in computational science and engineering to form the National Partnership for Advanced Computational Infrastructure (NPACI), teaming with approximately 40 university partners around the country. Today, SDSC is an organized research unit of the University of California, San Diego primarily funded by NSF with a staff of talented scientists, software developers and support personnel.





The National Resource for Biomedical Supercomputing (NRBSC) pursues leading edge research in high performance computing and the life sciences, and fosters exchange between PSC expertise in computational science and biomedical researchers nationwide.

Our focus is two-fold: computational biomedical research and outreach to the national biomedical research community through education and publications.

Research at NRBSC is centered in three areas: microphysiology; volumetric visualization and analysis; and computational structural biology.

NRBSC's education arm includes not only user training, but also software distribution, publications, and other outreach activities such as online courses and workshop webcasts.

The National Resource for Biomedical Supercomputing, formerly the Biomedical Initiative, was established at the Pittsburgh Supercomputing Center in 1987 as the first extramural biomedical supercomputing program in the country funded by the National Institutes of Health.

Tuesday, December 05, 2006

Python Resources Collection

Website



Python® is a dynamic object-oriented programming language that can be used for many kinds of software development. It offers strong support for integration with other languages and tools, comes with extensive standard libraries, and can be learned in a few days. Many Python programmers report substantial productivity gains and feel the language encourages the development of higher quality, more maintainable code.

Jython
Python is an implementation of the high-level, dynamic, object-oriented language Python written in 100% Pure Java, and seamlessly integrated with the Java platform. It thus allows you to run Python on any Java platform.




Stored in these dark caverns you may find rich veins of Python code, collected caches of Python information, and all manner of sundry Python passageways to explore. With candle or torch in hand, good hunting this night to all.Those not familiar with Python perhaps might start your quest at a brighter place.

NumPy

The fundamental package needed for scientific computing with Python is called NumPy. This package contains: a powerful N-dimensional array objectsophisticated (broadcasting) functionsbasic linear algebra functions basic Fourier transformssophisticated random number capabilitiestools for integrating Fortran code.

SciPy.org
SciPy (pronounced "Sigh Pie") is open-source software for mathematics, science, and engineering. It is also the name of a very popular conference on scientific programming with Python. The core library is NumPy which provides convenient and fast N-dimensional array manipulation. The SciPy library is built to work with NumPy arrays, and provides many user-friendly and efficient numerical routines such as routines for numerical integration and optimization. Together, they run on all popular operating systems, are quick to install, and are free of charge. NumPy and SciPy are easy to use, but powerful enough to be depended upon by some of the world's leading scientists and engineers. If you need to manipulate numbers on a computer and display or publish the results, give SciPy a try!


DISLIN Homepage
DISLIN is a high-level plotting library for displaying data as curves, polar plots, bar graphs, pie charts, 3D-color plots, surfaces, contours and maps.


wxPython is a GUI toolkit for the Python programming language. It allows Python programmers to create programs with a robust, highly functional graphical user interface, simply and easily.


VPython is a package that includes: the Python programming languagethe IDLE interactive development environment "Visual", a Python module that offers real-time 3D output, and is easily usable by novice programmers"Numeric", a Python module for fast processing of arrays

PyOpenGL Logo

PyOpenGL is the cross platform Python binding to OpenGL and related APIs. The binding is created using the SWIG wrapper generator, and is provided under an extremely liberal BSD-style Open-Source license.


The Biopython Project is an international association of developers of freely available Python tools for computational molecular biology.It is a distributed collaborative effort to develop Python libraries and applications which address the needs of current and future work in bioinformatics. The source code is made available under the Biopython License, which is extremely liberal and compatible with almost every license in the world. We work along with the Open Bioinformatics Foundation, who generously provide web and CVS space for the project

PyZine
The journal of Python Language



PyLucene is a GCJ-compiled version of Java Lucene integrated with Python. Its goal is to allow you to use Lucene's text indexing and searching capabilities from Python. It is designed to be API compatible with the latest version of Java Lucene.

Courses with an emphasis on scientific computing

Python course in Bioinformatics
Introduction to Python and Biopython with biological examples.



Monday, December 04, 2006

LaTeX Resources Collections

To start with Latex, the most convenient way is to read a short book with a strange name( sorry, I can not recall it now).
And Professor Schneider provides a page to introduce Latex for Biologist, here it is
LaTeX Style and BiBTeX Bibliography Formats for Biologists: TeX and LaTeX Resources

Monday, November 27, 2006

The International Society for Computational Biology (ISCB)


The International Society for Computational Biology (ISCB) is incorporated in the United States as a 501(c)(3) non-profit corporation, and registered in the state of California as a Charitable Trust. Now hosted at the San Diego Supercomputer Center at University of California, San Diego, the Society was officially formed in 1997 as an outgrowth of the conference on Intelligent Systems for Molecular Biology (ISMB). From humble beginnings, both ISCB's membership and ISMB's annual attendance have kept pace with the overall explosive growth experienced in the field of bioinformatics/computational biology.

For the complete story from conception to present day please visit the History link below. Additional links provide copies of documents that detail the legal structure of ISCB, which may prove interesting to our current and prospective members, as well as be of some use to regional groups around the world wanting to form national or regional societies and not knowing how to begin. And finally, ISCB's mission, vision and values are detailed in the 2003 Strategic Plan, and activities over the years are documented in the Newsletter Archives. We encourage you to peruse both of these links for a perspective on where we've been and where we may be going in the years ahead.

Ryan Songer

Sunday, November 26, 2006

ThinkWiki: Install Linux on ThinkPad

kkk recommended this site.

ThinkWiki

From ThinkWiki

Jump to: navigation, search

This is ThinkWiki, the Wiki Web for ThinkPad users. Here you find anything you need to install your favourite Linux distribution on your ThinkPad. Windows users shouldn't run away, there's a lot of useful information for them as well.

Please support us and help to extend this wiki. Thank you!



Ryan Songer

Wednesday, October 11, 2006

Zooomr and Its Free Pro Account

bird01bird01 Hosted on Zooomr

Frankly speaking, I have been a flickr fan for a long time thought it has a monthly upload limit, because I didn't expect I would use up the 50M limits(about 30+ pics) per month. However, I was frustrated when I returned from a journey with more than 50+ pics. This experience forced me to find a new photo hosting site. For a poor students like me, it should be free, free of upload limit, large enough, stable and comfortable. Under this criteria, picasa web album is poorly 250MB and still too simple. Fotkit is not free and has a messy UI. ... I have been looking for it until I came across a blog post Do We Love Bloggers? Yes We Do!
I signed up and write this post in hope a free pro account.

Thanks Kris. Yes, now I get pro account after some errors occurred though. Its seems that Zooomr is not reliably stable at this time, but I would like to try it.

Sunday, September 24, 2006

Can a Biologist Fix a Radio? -- or, What I Learned while Studying Apoptosis

Y. Lazebnik

Cold Spring Harbor Laboratory, Cold Spring Harbor

As a freshly minted Assistant Professor, I feared that everything in my field would be discovered before I even had a chance to set up my laboratory. Indeed, the field of apoptosis, which I had recently joined, was developing at a mind-boggling speed. Components of the previously mysterious process were being discovered almost weekly, frequent scientific meetings had little overlap in their contents, and it seemed that every issue of Cell, Nature, or Science had to have at least one paper on apoptosis. My fear led me to seek advice from David Papermaster (currently at the University of Connecticut), who I knew to be a person with pronounced common sense and extensive experience. David listened to my outpouring of primal fear and explained why I should not worry.

David said that every field he witnessed during his decades in biological research developed quite similarly. At the first stage, a small number of scientists would somewhat leisurely discuss a problem that would appear esoteric to others, such as whether cell cycle is controlled by an oscillator or whether cells can commit suicide. At this stage, the understanding of the problem increases slowly, and scientists are generally nice to each other, a few personal antipathies notwithstanding. Then, an unexpected observation, such as the discovery of cyclins or the finding that apoptosis failure can contribute to cancer, makes many realize that the previously mysterious process can be dissected with available tools and, importantly, that this effort may result in a miracle drug. At once, the field is converted into a Klondike gold rush with all the characteristic dynamics, mentality, and morals. A major driving force becomes the desire to find the nugget that will secure a place in textbooks, guarantee an unrelenting envy of peers, and, at last, solve all financial problems. The assumed proximity of this imaginary nugget easily attracts both financial and human resources, which results in a rapid expansion of the field. The understanding of the biological process increases accordingly and results in crystal clear models that often explain everything and point at targets for future miracle drugs. People at this stage are not necessarily nice, though, as anyone who has read about a gold rush can expect. This description fit the then current state of the apoptosis field rather well, which made me wonder why David was smiling so reassuringly. He took his time to explain.

At some point, David said, the field reaches a stage at which models, that seemed so complete, fall apart, predictions that were considered so obvious are found to be wrong, and attempts to develop wonder drugs largely fail. This stage is characterized by a sense of frustration at the complexity of the process, and by a sinking feeling that despite all that intense digging the promised cure-all may not materialize. In other words, the field hits the wall, even though the intensity of research remains unabated for a while, resulting in thousands of publications, many of which are contradictory or largely descriptive. The flood of publications is explained, in part, by the sheer amount of accumulated information (about 10,000 papers on apoptosis were published yearly over the last few years), which makes reviewers of the manuscripts as confused and overwhelmed as their authors. This stage can be summarized by the paradox that the more facts we learn the less we understand the process we study.

It becomes slowly apparent that even if the anticipated gold deposits exist, finding them is not guaranteed. At this stage, the Chinese saying that it is difficult to find a black cat in a dark room, especially if there is no cat, comes to mind too often. If you want to continue meaningful research at this time of widespread desperation, David said, learn how to make good tools and how to keep your mind clear under adverse circumstances. I am grateful to David for his advice, which gave me hope and, eventually, helped me to enjoy my research even after my field did reach the state he predicted.

At some point, I began to realize that David's paradox has a meaning that is deeper than a survival advice. Indeed, it was puzzling to me why this paradox manifested itself not only in studies of fundamental processes, such as apoptosis or cell cycle, but even in studies of individual proteins. For example, the mystery of what the tumor suppressor p53 actually does seems only to deepen as the number of publications about this protein rises above 23,000.

The notion that your work will create more confusion is not particularly stimulating, which made me look for guidance again. Joe Gall at the Carnegie Institution, who started to publish before I was born, and is an author of an excellent series of essays on history of biology [1], relieved my mental suffering by pointing out that a period of stagnation is eventually interrupted by a new development. As an example, he referred to the studies of cell death that took place in the 19th century, faded into oblivion, and re-emerged a century later with about 60,000 studies on the subject published during a single decade. Even though a prospect of a possible surge in activity in my field was relieving, I started to wonder whether anything could be done to expedite this event, which brought me to think about the nature of David's paradox. The generality of the paradox suggested some common fundamental flaw of how biologists approach problems.

To understand what this flaw is, I decided to follow the advice of my high school mathematics teacher, who recommended testing an approach by applying it to a problem that has a known solution. To abstract from peculiarities of biological experimental systems, I looked for a problem that would involve a reasonably complex but well understood system. Eventually, I thought of the old broken transistor radio that my wife brought from Russia (Fig. 1, see color insert). Conceptually, a radio functions similarly to a signal transduction pathway in that both convert a signal from one form into another (a radio converts electromagnetic waves into sound waves). My radio has about a hundred various components, such as resistors, capacitors, and transistors, which is comparable to the number of molecules in a reasonably complex signal transduction pathway. I started to contemplate how biologists would determine why my radio does not work and how they would attempt to repair it. Because a majority of biologists pay little attention to physics, I had to assume that all we would know about the radio is that it is a box that is supposed to play music.

Figure 1

Fig. 1. The radio that has been used in this study.

How would we begin? First, we would secure funds to obtain a large supply of identical functioning radios in order to dissect and compare them to the one that is broken. We would eventually find how to open the radios and will find objects of various shape, color, and size (Fig. 2, see color insert). We would describe and classify them into families according to their appearance. We would describe a family of square metal objects, a family of round brightly colored objects with two legs, round-shaped objects with three legs and so on. Because the objects would vary in color, we will investigate whether changing the colors affects the radio's performance. Although changing the colors would have only attenuating effects (the music is still playing but a trained ear of some people can discern some distortion), this approach will produce many publications and result in a lively debate.

Figure 2

Fig. 2. The insides of the radio. See text for description of the indicated components. The inset is an enlarged portion of the radio. The horizontal arrows indicate tunable components.

A more successful approach will be to remove components one at a time or to use a variation of the method, in which a radio is shot at a close range with metal particles. In the latter case, radios that malfunction (have a “phenotype”) are selected to identify the component whose damage causes the phenotype. Although removing some components will have only an attenuating effect, a lucky postdoc will accidentally find a wire whose deficiency will stop the music completely. The jubilant fellow will name the wire Serendipitously Recovered Component (SRC) and then find that SRC is required because it is the only link between a long extendable object and the rest of the radio. The object will be appropriately named the Most Important Component (MIC) of the radio. A series of studies will definitively establish that MIC should be made of metal and the longer the object is the better, which would provide an evolutionary explanation for the finding that the object is extendable.

However, a persistent graduate student from another laboratory will discover another object that is required for the radio to work. To the delight of the discoverer, and the incredulity of the flourishing MIC field, the object will be made of graphite and changing its length will not affect the quality of the sound significantly. Moreover, the graduate student would convincingly demonstrate that MIC is not required for the radio to work, and will suitably name his object the Really Important Component (RIC). The heated controversy, as to whether MIC or RIC is more important, will be fueled by the accumulating evidence that some radios require MIC while other, apparently identical ones, need RIC. The fight will continue until a smart postdoctoral fellow will discover a switch, whose state determines whether MIC or RIC is required for playing music. Naturally, the switch will become the Undoubtedly Most Important Component (U-MIC). Inspired by these findings, an army of biologists will apply the pull-it-out approach to investigate the role of each and every component. Another army will crush the radios into small pieces to identify components that are on each of the pieces, thus providing evidence for interaction between these components. The idea that one can investigate a component by cutting its connections to other components one at a time or in a combination (“alanine scan mutagenesis”) will produce a wealth of information on the role of the connections.

Eventually, all components will be catalogued, connections between them will be described, and the consequences of removing each component or their combinations will be documented. This will be the time when the question, previously obscured by the excitement of productive research, would have to be asked: Can the information that we accumulated help us to repair the radio? It will turn out that sometimes it can, such as if a cylindrical object that is red in a working radio is black and smells like burnt paint in the broken radio (Fig. 2, inset, a component indicated as a target). Replacing the burned object with a red object will likely repair the radio.

The success of this approach explains the pharmaceutical industry's mantra: “Give me a target!”. This mantra reflects the belief in a miracle drug and assumes that there is a miracle target whose malfunction is solely responsible for the disease that needs to be cured.

However, if the radio has tunable components, such as those found in my old radio (indicated by yellow arrows in Fig. 2, inset) and in all live cells and organisms, the outcome will not be so promising. Indeed, the radio may not work because several components are not tuned properly, which is not reflected in their appearance or their connections. What is the probability that this radio will be fixed by our biologists? I might be overly pessimistic, but a textbook example of the monkey that can, in principle, type a Burns poem comes to mind. In other words, the radio will not play music unless that lucky chance meets a prepared mind.

Yet, we know with near certainty that an engineer, or even a trained repairman could fix the radio. What makes the difference? I think the languages that these two groups use (Fig. 3, see color insert). Biologists summarize their results with the help of all-too-well recognizable diagrams, in which a favorite protein is placed in the middle and connected to everything else with two-way arrows. Even if a diagram makes overall sense (Fig. 3a), it is usually useless for a quantitative analysis, which limits its predictive or investigative value to a very narrow range. The language used by biologists for verbal communications is not better and is not unlike that used by stock market analysts. Both are vague (e.g., “a balance between pro- and anti-apoptotic bcl-2 proteins appears to control the cell viability, and seems to correlate in long-term with the ability to form tumors”) and avoid clear predictions.

Figure 3

Fig. 3. The tools used by biologists and engineers to describe processes of interest: a) the biologist view of a radio. See Fig. 2 and text for description of the indicated components; b) the engineer view of a radio (please note that the circuit diagram presented is not that of the radio used in the study; the diagram of the radio was lost, which, in part, explains why the radio remains broken).

These description and communication tools are in a glaring contrast with the language that has been used by engineers (compare Figs. 3a and 3b). Because the language (Fig. 3b) is standard (the elements and their connections are described according to invariable rules), any engineer trained in electronics would unambiguously understand a diagram describing the radio or any other electronic device. As a consequence, engineers can discuss the radio using terms that are understood unambiguously by the parties involved. Moreover, the commonality of the language allows engineers to identify familiar patterns or modules (a trigger, an amplifier, etc.) in a diagram of an unfamiliar device. Because the language is quantitative (a description of the radio includes the key parameters of each component, such as the capacity of a capacitor, and not necessarily its color, shape or size) it is suitable for a quantitative analysis, including modeling.

I would like to argue that the absence of such language is the flaw of biological research that causes David's paradox. Indeed, even though the impotence of purely experimental approaches might be a bit exaggerated in my radio metaphor, it is common knowledge that the human brain can keep track of only so many variables. It is also common experience that once the number of components in a system reaches a certain threshold, understanding the system without formal analytical tools requires geniuses, who are so rare even outside biology. In engineering, the scarcity of geniuses is compensated, at least in part, by a formal language that successfully unites the efforts of many individuals, thus achieving a desired effect, be that design of a new aircraft or of a computer program. In biology, we use several arguments to convince ourselves that problems that require calculus can be solved with arithmetic if one tries hard enough and does another series of experiments.

One of these arguments postulates that the cell is too complex to use engineering approaches. I disagree with this argument for two reasons. First, the radio analogy suggests that an approach that is inefficient in analyzing a simple system is unlikely to be more useful if the system is more complex. Second, the complexity is a term that is inversely related to the degree of understanding. Indeed, the insides of even my simple radio would overwhelm an average biologist (this notion has been proven experimentally), but would be an open book to an engineer. The engineers seem to be undeterred by the complexity of the problems they face and solve them by systematically applying formal approaches that take advantage of the ever-expanding computer power. As a result, such complex systems as an aircraft can be designed and tested completely in silico, and computer-simulated characters in movies and video games can be made so eerily life-like. Perhaps, if the effort spent on formalizing description of biological processes would be close to that spent on designing video games, the cells would appear less complex and more accessible to therapeutic intervention.

A related argument is that engineering approaches are not applicable to cells because these little wonders are fundamentally different from objects studied by engineers. What is so special about cells is not usually specified, but it is implied that real biologists feel the difference. I consider this argument as a sign of what I call the urea syndrome because of the shock that the scientific community had two hundred years ago after learning that urea can be synthesized by a chemist from inorganic materials. It was assumed that organic chemicals could only be produced by a vital force present in living organisms. Perhaps, when we describe signal transduction pathways properly, we would realize that their similarity to the radio is not superficial. In fact, engineers already see deep similarities between the systems they design and live organisms [2].

Another argument is that we know too little to analyze cells in the way engineers analyze their systems. But, the question is whether we would be able to understand what we need to learn if we do not use a formal description. The biochemists would measure rates and concentrations to understand how biochemical processes work. A discrepancy between the measured and calculated values would indicate a missing link and lead to the discovery of a new enzyme, and a better understanding of the subject of investigation. Do we know what to measure to understand a signal transduction pathway? Are we even convinced that we need to measure something? As Sydney Brenner noted, it seems that biochemistry disappeared in the same year as communism [3]. I think that a formal description would make the need to measure system's parameters obvious and would help to understand what these parameters are.

An argument that is usually raised privately is why to bother with all these formal languages if one can make a living by continuing with purely experimental research that took years to learn. There are at least two reasons. One is that formal approaches would make our research more meaningful, more productive and might indeed lead to miracle drugs. A more immediate reason is that formal approaches may become a basic part of biology sooner than we, experimental biologists, expect. This transition may be as rapid as that from slides to PowerPoint presentations, a change that forced some graphics designers to learn how to use a computer and put others out of work.

Of course, a plea for a formal approach in biology is not new. The general systems theory, developed by Ludwig von Bertalanffy because of his fascination with the complexity of live organisms, was formulated 60 years ago, as well as his concept of organisms as physical systems [4]. Bertalanffy's fundamental studies have been followed by several attempts to approach cells as systems, the latest of which, system biology, has been rapidly developing into an active field [5-11]. Available computer power and advances in analysis of complex systems raise hope that this time the system approach will change from an esoteric tool that is considered useless by many experimental biologists, to a basic and indispensable approach of biology.

The question is how to facilitate this change, which is not exactly welcomed by many experimental biologists, to put it mildly [12]. Learning computer programming was greatly facilitated by BASIC, a language that was not very useful to solve complex problems, but was very efficient in making one comfortable with using a computer language and demonstrating its analytical power. Similarly, a simple language that experimental scientists can use to introduce themselves to formal descriptions of biological processes would be very helpful in overcoming a fear of long-forgotten mathematical symbols. Several such languages have been suggested [13, 14] but they are not quantitative, which limits their value. Others are designed with modeling in mind but are too new to judge as to whether they are user-friendly [15]. However, I hope that it is only a question of time before a user-friendly and flexible formal language will be taught to biology students, as it is taught to engineers, as a basic requirement for their future studies. My advice to experimental biologists is to be prepared.

Friday, September 08, 2006

Mathematics is a game of life

source http://www.hno.harvard.edu/gazette/2001/02.01/03-mathematics.html

Ahha, another guy in my dreaming research field. However, I think his path to computational biology is at a way of probability. It takes too much time for him to find the right target of his research and he lacks a broad understanding biology in perspective of others. However, his road is promising to me, proving the era of computational biology. --by Ryan

Jun Liu uses statistics to understand genes

By William J. Cromie
Gazette Staff

Jun Liu remembers being interested in mathematics as early as age 12. It was a hard interest to pursue in the waning years of the Cultural Revolution in China. Computers were not available to him. He didn't own a calculator. Mathematics books were difficult to find.

Jun's parents, both teachers, scrounged books wherever they could. They borrowed books from older professors who had hidden them away. His father, a professor at a technical university in Beijing, copied one book entirely by hand.

"I couldn't tell high school from college texts, so I read everything," Liu recalls. "Doing math was like a game you could play with only a piece of paper and pencil. On Sundays, I rode my bike for an hour to meet friends and do problems."

Sitting in his office at the Science Center, Liu, now 35, still shows a youthful enthusiasm for math problems. He wants to find answers to fundamental questions about genes and how they control life. "Every cell in your body contains a complete set of genes; each cell could become a part of your eye, your hand, or your brain. The question that challenges many scientists is how cells decide to be part of one organ or the other."

Liu thinks he may be able to get some of the answers with statistics more quickly than biologists can with experiments.

Commenting on Liu's recent tenure appointment, fellow professor of statistics Donald Rubin called him "a great asset to the University both as a teacher and a colleague. Harvard needs his strength in computational biology. In addition, he's a wonderful warm guy, soft-spoken but with a fabulous sense of humor."

Fooling around

Liu attended Beijing University, where he admits he spent most of his time playing bridge, and "fooling around." Nevertheless, after graduation in 1985, he was among a group of top math students who came to the United States on a scholarship program supported by the Society for Industrial and Applied Mathematics.

Sent to Rutgers University in New Jersey, Liu experienced cultural shock. "It was like a movie," he recalls. "It didn't feel real."

Language turned out to be the highest barrier. "Fortunately, I could get by with understanding formulas and equations," he says. "I got straight As without knowing much of what the teachers were saying. After a year, my English was worse than when I first came."

In 1986, Liu transferred to the University of Chicago. While there, he became interested in human rights issues and spent a lot of time participating in student protests. That activism brought his adviser to ask him if he wanted to be "a politician or a mathematician."

That was a turning point for Liu. He decided he had to apply himself to a career in statistics. "I didn't want to solve problems just because they hadn't been solved by others," he says. "I wanted to connect to reality. Although I didn't know exactly what statistics is, the field appealed to me for that reason."

"Liu has very original ideas, exceptional ability, and amazing computational skills," says Wing Wong, who supervised his Ph.D research in Chicago and who is now a professor of statistics at the Harvard School of Public Health. "He's also a clear thinker who communicates his ideas well, and is an easygoing, warm, and helpful person."

Tenure struggle

Liu earned a Ph.D. in statistics in 1991, then came to Harvard. During his first year, he met Wei Zhang, a graduate student studying immunology. They fell in love and married in 1994.

The same year, Stanford offered him a position in its statistics department. "I knew the chance of getting tenure at Stanford was nonexistent," Liu says. "But the chance of getting it at Harvard was one magnitude less than that, so I went to the West Coast."

His wife earned her Ph.D. in 1995, and followed him west. She got a job as a consultant in Los Angles and they commuted between apartments in the two different cities.

Despite his pessimism about tenure, Liu was offered the honor last year at both Stanford and Harvard.

The choice was not a difficult one for him. "I liked teaching undergraduate courses at Harvard; the students seemed more interested in the work than students at Stanford," he says.

Liu is also impressed by "the many great biologists and chemists here." Also of high interest to him is the new Bauer Center of Genomics Research where biologists, mathematicians, chemists, and others will work together to find the general principles that underlie life.

Liu now pursues the mystery of how genes are turned on and off. Using various statistical and computer techniques, he studies repetitive patterns in the DNA that lies between genes. This material contains instructions for regulating the expression of genes, and it is involved in whether the proteins produced by genes will become part of a brain or a big toe.

These on/off switches can be found by doing difficult, time-consuming experiments that require copying and mutating genes. If a region close to a gene is mutated and the gene stops producing a certain protein, that region must be part of a genetic switch. Liu believes he can locate such switches by statistical analysis of the genetic sequence patterns that occur between the actual genes.

Liu has done some of this work with collaborators at Harvard Medical School and the New York State Department of Health. For example, he has made about 2,000 predictions of where switches are located in the bacterium e-coli. In cases where these switches have actually been found by experiments, his predictions are 80 percent correct.

"What's nice (or not so nice) about this field," he says, "is that there's always a judgment day. You are making predictions of actual happenings, so you'll always know how right (or wrong) you are."

Liu is pleased with his results so far. But humans have at least 10 times more genes than bacteria, so he knows that things will get a lot more complicated.

That's OK with him. Talking with Liu in his office on a bright winter afternoon, it's easy to see that he is glad to be here. He misses playing bridge, fooling around, and traveling as a student on a few dollars a day. And his wife, now a consultant, makes more money than he does. But Liu has tenure now, and it's good to be back among attentive students and good colleagues. Best of all, he feels he can strengthen the contributions that Harvard makes to understanding what biological life is all about.

Tuesday, September 05, 2006

Quotations


There are no such things as applied sciences, only applications of science.
Louis Pasteur

Failing to plan is planning to fail.
someone
The greatest value of a picture is when it forces us to notice what we never expected to see.
J. W. Tukey

Structural genomics is a field born of the the marriage between computer science and biology. The goal is to develop new technology - specifically, computer algorithms - to enable the determination of 10,000 new protein structures in 10 years. This would have an enormous impact on our understanding of disease mechanisms, and our ability to target drugs to specific protein targets. If successful, the impact would be comparable to that of the human genome project.
Professor Bruce Donald
Success is the ability to go from failure to failure without losing your enthusiasm.
--- Winston Churchill (1874-1965)
When once you have tasted flight, you will forever walk the earth with your eyes turned skyward, for there you have been and there you will always long to return.
— Leonardo da Vinci
Not looking for the best ,looking for the best fit

most of us are the prisoners of our own experience
John Garing


Thursday, August 31, 2006

Dennis S. Hurb's Personal Homepage and Blog

I come acroos Mr. Dennis S. Hurd in orkut. He lives in Canada and work as English teacher for non-native English speakers. I find him a good friend. He maintains a daily updated blog which covers a wide range of topics, such as travel, language, etc. And now Mr.Hurd is onjoying flickr. Here is the gateway to his homepage and blog with the url: http://www.dennissylvesterhurd.com/

My personal history
and career goals



A daily blog since 09/2003
'my eJournal and images'


Sample personal images
and world ePictures

Tuesday, August 29, 2006

Graducate School Advice (Computer Science)

Brief description

A computer science graduate school survival guide, intended for prospective or novice graduate students. This guide describes what I wish I had known at the start of graduate school but had to learn the hard way instead. It focuses on mental toughness and the skills a graduate student needs. The guide also discusses finding a job after completing the Ph.D. and points to many other related web pages.


"So long, and thanks for the Ph.D.!"

a.k.a.

"Everything I wanted to know about C.S. graduate school
at the beginning but didn't learn until later."

The 4th guide in the Hitchhiker's guide trilogy
(and if that doesn't make sense, you obviously have not read Douglas Adams)

by Ronald T. Azuma

v. 1.07

Original version 1997, revised March 2000


Introduction

    "To know the road ahead, ask those coming back."
    - Chinese proverb

    In February 1995, on a beautiful sunny day with clear Carolina blue skies, I turned in the final, signed copy of my dissertation. The graduate school staff member did some last-minute checks on the document and pronounced it acceptable. After six and a half years of toil and sweat, I was finally done! While walking back to the C.S. department building, I was sorely disappointed that the heavens didn't part, with trumpet-playing angels descending to announce this monumental occasion. Upon hearing this observation, Dr. Fred Brooks (one of my committee members) commented, "And the sad fact is, you're no smarter today than you were yesterday." "That's true," I replied, "but the important thing is that I am smarter than I was six and a half years ago."

    That day was over two years ago, and since then I have had plenty of time to reflect upon my graduate student career. One thought that has repeatedly struck me is how much easier graduate school might have been if somehow, magically, some of the things I knew when I turned in my dissertation I could have known when I first entered graduate school. Instead, I had to learn those the hard way. Of course, for many topics this is impossible: the point of graduate school is to learn those by going through the experience. However, I believe other lessons can be taught ahead of time. Unfortunately, such guidance is rarely offered. While I had to learn everything the hard way, new graduate students might benefit from my experiences and what I learned. That is the purpose of this guide.

    Very little of this guide discusses technical matters. Technical skills, intelligence and creativity are certainly strong factors for success in graduate school. For example, I doubt there is a C.S. graduate student who didn't at one point wish he or she had a stronger mathematical background. However, it's beyond the scope of this guide to tell you how to be technically brilliant, as the following joke implies:

      The Feynman Problem Solving Algorithm:
      1) Write down the problem.
      2) Think very hard.
      3) Write down the solution.
    You don't have to be a genius to do well in graduate school. You must be reasonably intelligent, but after a certain point, I think other traits become more important in determining success.

    This guide covers the character traits and social skills that often separate the "star" graduate students from the ordinary ones. Who are the students who are self-motivated, take initiative, find ways around obstacles, communicate well both orally and in writing, and get along well enough with their committee and other department members to marshal resources to their cause? Which students seem to know "how the system works" and manage to get things done? These traits are hardly unique to succeeding in graduate school; they are the same ones vital to success in academic or industrial careers, which is probably why many of the best graduate students that I knew were ones who had spent some time working before they came back to school.

    This document is aimed at junior C.S. graduate students, but these observations are probably broad enough to apply to graduate education in other technical fields. My conclusions are certainly colored by my particular experiences (doing my dissertation work in interactive computer graphics in the Computer Science department of the University of North Carolina at Chapel Hill) but I think they are fairly general in application and should be of interest to readers at other schools and other C.S. specialties. Obviously, these are only my opinions and may not represent the views of other sane individuals or organizations. Some points may be controversial, but if they weren't this would not be interesting reading. Parts of this document come from two informal talks I gave at UNC about "the Ph.D. job hunt" and "observations from spending one year in industrial research." Both talks had larger audiences than any informal technical talk I gave at UNC, which told me that students are definitely interested on these subjects!


Why get a Ph.D.?

    "Being a graduate student is like becoming all of the Seven Dwarves. In the beginning you're Dopey and Bashful. In the middle, you are usually sick (Sneezy), tired (Sleepy), and irritable (Grumpy). But at the end, they call you Doc, and then you're Happy."
    - yours truly

    The most basic question every Ph.D. student must know the answer to is: "Why the hell am I doing this?"

    It's a good question. The hours are long. The pay is low, with minimal benefits. This is especially true in today's job market, where many people with computing skills are getting big salaries and stock option packages. After graduation, Ph.D. salaries are higher than B.S. and M.S. salaries, but the difference doesn't make up for the income lost by staying in school longer. The M.S. has a better "bucks for the time invested" ratio than the Ph.D. does. And in terms of social status, a graduate student doesn't rank very high on the ladder.

    If you do not have an acceptable answer to this question, then don't get a Ph.D. I repeat: if you do not have a rock-solid reason for getting the Ph.D., then it is better that you leave with a Master's.

    Why? Completing a Ph.D. is a long, hard road with many potholes and washed out bridges along the way. You may run over some land mines and have to stop and turn around and explore other routes. If the goal is important enough to you, then these obstacles will not prevent you from completing your journey. But if you don't know why you are on this road, then you will get discouraged and will probably leave without finishing, having wasted years of your life.

    I faced this situation after the first time I took the Doctoral Written Exam (which at the time was the entrance examination into the Ph.D. program). I missed passing it by just 4 percentage points. I then had to decide whether or not to try again next semester (committing myself again to spending weeks getting ready for the test) or just leave with an M.S. degree.

    I didn't come to graduate school with the Ph.D. as the primary goal. So this test result forced me to answer the basic question "Why the hell am I doing this?" After much soul searching, I found my answer and decided to take the test again, passed it, and went on to get my Ph.D.

    I got the Ph.D. because I wanted to get a research position after leaving graduate school. I wanted to work with the state of the art and extend it. I did not want to "bring yesterday's technology one step closer to tomorrow." I wanted a job that would I find interesting, challenging and stimulating. While an M.S. would give me a chance at landing a research position, the Ph.D. would give me a much better chance. And I did not want to live with regrets. If I took the Doctoral Written Exam again and failed again, then I could say that it wasn't meant to be and move on with my life. I would have no regrets because I had given it my best shot and was not able to make it. However, if I left with an M.S. without taking the test a second time, and five years later I was in a job that was boring and uninteresting, then I would have to lie awake every night for the rest of my life wondering "What if?" What if I had taken the test again and passed? Would I then be in the job that I really wanted? That was not a situation I wanted to be in. I did not want to live the rest of my life regretting what might have been.

    In hindsight, I think one of the main reasons I successfully completed the Ph.D. was the fact that I didn't pass the exam on the first try. It's ironic, but life sometimes works in strange ways. That initial failure caused me to answer the basic question, providing the mental fortitude to keep going despite the hurdles and problems I would later face.

    My answer is you should get a Ph.D. if it is required for your goals after graduate school, such as becoming a professor or a researcher in academia, government or industry. Your answer may differ from mine. As long as you have an answer that you believe in passionately, then that's enough. If you don't have an answer, then save yourself a lot of grief and don't get the Ph.D.

Academia is a business
    "Remember the Golden Rule: Those who have the gold make the rules."

    Academia is a business, and "graduate student" is a job title. This is especially true at private universities. Academia is very peculiar type of business. It is certainly not the Real World and does not work in the same way that the ordinary corporate world does. However, it is a business nonetheless and as a graduate student, you must treat it that way. Graduate school made a lot more sense and became much easier for me after I realized this. If you think of graduate school as an "Ivory Tower" free of politics, money problems, and real-world concerns, you are going to be severely disappointed. If you don't believe me, read The Idea Factory by Pepper White (listed in the references) for one account of graduate life at MIT.

    A few graduate students are independently wealthy or have fellowship and scholarship money that cover all their expenses for their total stay in graduate school. Such students are rare, however. Most of us needed financial support, in the form of Teaching Assistantships or Research Assistantships (RA's). In general, RA's are more desirable to students since those can directly fund the research you need to finish.

    Where does the money come from to fund RA's? Your professors have to raise funds from external organizations. These include government agencies such as the National Science Foundation (NSF), Defense Advanced Research Projects Agency (DARPA), the Office of Naval Research (ONR), and others. Private companies also fund some university research, although this tends to be less common, in smaller amounts, and in the form of equipment donations. These organizations don't just give money away as charity. They expect their money to accomplish something. Increasingly these days, this takes the form of a contract for a working demonstration that must be shown at the end. That means once the money is delivered, your professors must come through with the working demonstration. It is rare that they do this by themselves. Instead, they find some very capable, young, self-motivated people who are willing to work long hours for small amounts of pay. In other words, they fund RA's.

    The RA job is crucial to the academic business. If the RA's cannot successfully conduct the research, then the demonstration will not work in the end and the funding agencies may not be happy. They may choose not to fund your professor in the future, which will bring his or her research program to a halt. And there are many professors and other researchers chasing too few research dollars these days; it is a competitive market. Thus, each professor wants the best students available. These students are the most capable ones who can get the research done required to fulfill the funding contracts.

    That means you must treat an RA like a job. You must prove to your professors that you are capable of getting the work done, being a team player, communicating your results, and most of the other characteristics needed to do well in regular jobs. That's why many of the upcoming sections in this guide sound like ones written for the regular workplace.

    What do you get out of this? At the start, you may have to do tasks specifically related to the funding contracts. But eventually your professor must be flexible enough to fund your own specific research program that leads to the completion of your dissertation. Your stipend and tuition waiver should be enough to live on frugally without going into debt. You will learn the state of the art in your chosen speciality and conduct cutting-edge research on a subject that you find interesting and enjoyable. If you don't find this compensation sufficient, then you shouldn't be in graduate school in the first place.

    The bottom line: realize that academia is a peculiar kind of business and the role you play in this enterprise. If you do your job well (and have good negotiation and interpersonal skills, as discussed in future sections), both your needs and your professors' needs will be met. But don't enter an RA position thinking that the computers, research equipment, staff members and other resources that you are provided with are your birthright. Don't take them for granted! Most of those exist only because your professors have been able to raise the money to provide those to you. In turn, you must fulfill your end of the deal by doing great research with those resources. If you don't do your job well, don't be surprised if your professors choose not to fund you in the future. They do not have to provide you with an RA job or let you use the computing equipment they acquired. And the student who has no funding, no tuition reimbursement and no access to required computing resources is the student who leaves the university that semester.

    How do you make sure you are one of those best, highly desired RA's? Read on!

Graduate school is a different ballgame
    "Don't let school get in the way of your education."
    - Mark Twain

    "The IQ test was invented to predict academic performance, nothing else. If we wanted something that would predict life success, we'd have to invent another test completely."
    - Robert Zajonc

    If you go through a Ph.D. program, you will find graduate school a very different world from undergraduate school. If you just get an M.S., then graduate school may not be much different from undergrad (depending on where you get your degree), except that the courses are deeper and more advanced. But for a Ph.D. student, graduate school is a whole new ballgame. The students who do well are the ones who learn this earlier rather than later and make the necessary adjustments.

    Graduate school is not primarily about taking courses. You will take classes in the beginning but in your later years you probably won't have any classes. People judge a recently graduated Ph.D. by his or her research, not by his or her class grades. And, without any offense to my professors, most of what you learn in a Ph.D. program comes outside of classes: from doing research on your own, attending conferences, and talking to your fellow students. Success in graduate school does not come from completing a set number of course units but rather from successfully completing a research program.

    Graduate school is more like an apprenticeship where each student has his or her own project, and the masters may or may not be particularly helpful. It's like teaching swimming by tossing students into the deep end of the pool and seeing who makes it to the other end alive and who drowns. It's like training clock designers by locking students inside a clock factory with some working clocks and lots of clock parts and machines for building clocks. However, the instructions are at best incomplete and even the masters themselves don't know exactly how to build next year's models.

    Excelling in a Ph.D. program requires different skills than doing well in undergrad. Undergraduate education tests you through class projects (that do not last more than a semester), essays, midterms and finals. For the most part, you work alone. Your professor may not know your name. Every other student in your class takes the same tests or does similar projects. But in a Ph.D. program, you must select and complete a unique long-term research program. For most of us, this means you have to learn how to do research and all that entails: working closely with your professors, staff and fellow students, communicating results, finding your way around obstacles, dealing with politics, etc.

    I'm not saying that tests and grades are completely unimportant in graduate school. One of the two biggest hurdles in completing a Ph.D. is passing the qualifying exam. (The other is finding an acceptable dissertation topic.) But because graduate school is not nearly as exam-based as undergraduate education and requires different skills, the GRE and undergraduate grades are not as good an indicator of who will excel and who will drop out as admission committees seem to think. Those tests do not measure creativity, tenacity, interpersonal skills, oral presentation skills, and many other important traits.

    The next several sections discuss these traits.

Initiative
    "The difference between people who exercise initiative and those who don't is literally the difference between night and day. I'm not talking about a 25 to 50 percent difference in effectiveness; I'm talking about a 5000-plus percent difference, particularly if they are smart, aware, and sensitive to others."
    - Stephen R. Covey, The 7 Habits of Highly Effective People

    The dissertation represents a focused, personal research effort where you take the lead on your own, unique project. If you expect that your adviser is going to hold your hands and tell you what to do every step of the way, you are missing the point of the dissertation. Ph.D. students must show initiative to successfully complete the dissertation. This does not mean that guidance from professors is unimportant, just that this guidance should be at a reasonably high level, not at a micromanaging level. If you never do any tasks except those that your professor specifically tells you to do, then you need to work on initiative.

    At UNC, there is a famous anecdote about a former UNC graduate student named Joe Capowski. Many years ago, UNC got a force-feedback mechanical arm to use with molecular visualization and docking experiments. The problem was how to move it to UNC. This mechanical arm is a large, heavy beast, and it was in Argonne National Labs in Chicago, IL. Unfortunately, there was a trucker's strike going on at the time. Joe Capowski, on his own initiative (and without telling anyone), flew out to Argonne, rented a car, drove the mechanical arm all the way back to North Carolina, and then handed the computer science department the bill! Many years later, Joe Capowski ran for the Chapel Hill city council and won a seat. Prof. Fred Brooks gave him an endorsement. I still remember the words Dr. Brooks said: "I may not agree with his politics, but I know he'll get things done."

    While the Joe Capowski anecdote is perhaps a bit extreme, it does show that it is often better to ask forgiveness than permission, provided you are not becoming a "loose cannon." Certain universities (e.g. MIT) are good at fostering a "can do" attitude among their graduate students, and therefore they become more assertive and productive. One of the hallmarks of a senior graduate student is that he or she knows the types of tasks that require permission and those that don't. That knowledge will come with experience. Generally, it's the senior graduate students who have the most freedom to take initiative on projects. This privilege has to be earned. The more that you have proven that you can work independently and initiate and complete appropriate tasks, the more your professors will leave you alone to do what you want to do.

Tenacity
    "Let me tell you the secret that has led me to my goal. My strength lies solely in my tenacity."
    - Louis Pasteur

    You don't need to be a genius to earn a Ph.D. (although it doesn't hurt). But nobody finishes a dissertation without being tenacious. A dissertation usually takes a few years to complete. This can be a culture shock to former undergraduates who have never worked on a project that lasted longer than one quarter or semester (at the end of which, whatever the state of the project, one declares victory and then goes home). No one can tell you in advance exactly how long the dissertation will take, so it's hard to see where the "end of the road" lies. You will encounter unexpected problems and obstacles that can add months or years to the project. It's very easy to become depressed and unmotivated about going on. If you are not tenacious about working on the dissertation, you won't finish.

    Tenacity means sticking with things even when you get depressed or when things aren't going well. For example, I did not enjoy my first year of graduate school. I didn't tell anyone this until after leaving UNC. I was not on a project and was focused on taking classes, some of which I didn't do all that well in. I didn't feel a part of the Department, and really wondered whether or not I fit in. Still, I stuck with it and when summer rolled around and I got a job in the Department, I became much more involved in research and enjoyed graduate school much more. Part of earning a Ph.D. is building a "thick skin" so you are not so fragile that you will give up at the first sign on any difficulties.

    One lesson I learned as a graduate student is the best way to finish the dissertation is to do something every day that gets you closer to being done. If all you have left is writing, then write part of the dissertation every day. If you still have research to do, then do part of it every day. Don't just do it when you are "in the mood" or feeling productive. This level of discipline will keep you going through the good times and the bad and will ensure that you finish.

Flexibility
    "Back in graduate school, I'd learned how to survive without funding, power, or even office space. Grad students are lowest in the academic hierarchy, and so they have to squeeze resources from between the cracks. When you're last on the list for telescope time, you make your observations by hanging around the mountaintop, waiting for a slice of time between other observers. When you need an electronic gizmo in the lab, you borrow it in the evening, use it all night, and return it before anyone notices. I didn't learn much about planetary physics, but weaseling came naturally."
    - Clifford Stoll, The Cuckoo's Egg

    "The Chinese call luck opportunity and they say it knocks every day on your door. Some people hear it; some do not. It's not enough to hear opportunity knock. You must let him in, greet him, make friends and work together."
    - Bernard Gittelson

    Flexibility means taking advantage of opportunities and synergies, working around problems, and being willing to change plans as required. As a graduate student, you are on the bottom of the academic totem pole. Even undergraduates can rank higher, especially at private universities (because they actually pay tuition!) You cannot order anybody to do anything. In general, you will be in the position of reacting to big events rather than controlling them. Therefore, you must be flexible in your approach and research program.

    For example, you may not have as much access to a piece of laboratory equipment as you would like, or maybe access is suddenly cut off due to events beyond your control. What do you do? Can you find a replacement? Or reduce the time needed on that equipment? Or come in at odd hours when no normal person uses that equipment? Or redefine the direction of your project so that equipment is no longer required?

    Events can be good as well as bad. The difference between the highly effective graduate student and the average one is that the former recognizes those opportunities and takes advantage of them. I had nothing to do with bringing Gary Bishop to UNC. But after he arrived I realized my research would progress much faster if he became my adviser so I made the switch and that was a big help to my graduate student career. Opportunities for synergy and serendipity do occur, but one has to be flexible enough to recognize them and take advantage of them.

Interpersonal skills
    "For humans, honesty is a matter of degree. Engineers are always honest in matters of technology and human relationships. That's why it's a good idea to keep engineers away from customers, romantic interests, and other people who can't handle the truth."
    - Scott Adams, The Dilbert Principle

    "I can calculate the motions of the heavenly bodies, but not the madness of people."
    - Isaac Newton

    Computer Science majors are not, in general, known for their interpersonal skills. Some of us got into this field because it is easier to understand machines than people. As frustrating as computers can be, they at least behave in a logical manner, while human beings often do not. However, your success in graduate school and beyond depends a great deal upon your ability to build and maintain interpersonal relationships with your adviser, your committee, your research and support staff and your fellow students. This does not mean you must become the "life of the party." I am not and never will be a gregarious, extroverted person. But I did make a serious effort to learn and practice interpersonal skills, and those were crucial to my graduate student career and my current industrial research position.

    Why should this matter, you may ask? If one is technically brilliant, shouldn't that be all that counts? The answer is no, because the situation is different from your undergraduate days. In both graduate school and in business, you must depend upon and work with other people to achieve your goals To put this in perspective, I have excerpted the following from an article called "Organizations: The Soft and Gushy Side" by Kerry J. Patterson, published in Fall 1991 issue of The Bent:

      I first learned of the capricious, human side of organizations some 15 years ago while studying the careers of engineers and scientists. The research design required that I spend eight hours a day in one-on-one interviews. For two hours I'd ask "career" questions of an engineer, chemist, physicist, or applied mathematician -- all of whom worked for a Fortune 500 firm. During these 120 minutes, the subjects talked about the perils of the organizations. Two hours was scarcely enough time to share their stories. All energetically discussed their personal careers. Most had been frustrated with the "soft and gushy" side of organizations. Some had figured out the system and learned to master it. Others had not.

      As part of the research design, we asked to talk to low, medium, and high performers. This in itself was an interesting exercise. To determine performance rankings, we would place in front of a senior manager the names of the 10-50 people within his or her organization. Each name would be typed neatly in the middle of a three-by-five card. After asking the manager to rank the employees from top to bottom, the managers would then go through a card sort. Typically the executive would sort the names into three or four piles and then resort each pile again. Whatever the strategy, the exercise usually took only minutes. Just like that, the individual in charge of the professionals in question was able to rank, from top to bottom, as many as 50 people. It rarely took more than three minutes and a couple of head scratches and grunts. Three minutes. Although politics may appear ambiguous to those on the receiving end, those at the top were able to judge performance with crystal clarity.

      This performance ranking (conducted by individuals not involved in the interviews) was then used as a dependent measure. Those of us conducting the interviews attempted to surface information (independent measures) that would predict the ranking. What about a scientist's career would lead to a top ranking? What trashed a perfectly good career? Surely scientific prowess would have an impact. And it did.

      But technological prowess wasn't as predictive as another factor. We discovered that we could tell what performance group the interviewees belonged to within a minute or two by their attitudes toward people and politics. Individuals who were ranked low by their managers spoke of organizational politics as if it were poison. They were exceptionally annoyed by the people side of the business. They frequently stated they would rather be left alone to conduct their research untrammeled by human emotions. They characterized the social side of organizations as "soft and gushy." They sounded like Spock turned bitter.

      Top performers, in contrast, found a way to work within the political system. They hadn't exactly embraced politics. They didn't appear like that toothy kid you knew back in college who lived to fight political battles. They didn't come off as glad-handling sales folks. These were professional scientists who were often top ranked in their field. They looked and talked liked scientists. The difference between them and those ranked at the bottom of the totem pole was clear. They had found a way to make peace with organizations, people, and politics. They climbed to the top of their field by mastering both hard things and soft and gushy people.

      Engineers and scientists aren't the only ones who find the human side of the organizations to be annoying. As we expanded our research to include professors, accountants, and other professionals, the findings were remarkably similar. All found political machinations to be distasteful. It's just that some had found a way to master the social aspects -- the top performers.

    Students usually look down on politics, but politics in its most basic, positive form is simply the art of getting things done. Politics is mostly about who is allowed to do what and who gets the resources (money, people, equipment, etc.) To succeed in your research, you will need resources, both capital and personnel. Interpersonal skills are mandatory for acquiring those resources. If you are incapable of working with certain people or make them mad at you, you will not get those resources and will not complete your research.

    For example, which group of people did I try my best to avoid offending? Was it my committee? No, because healthy disagreements and negotiations with your adviser and committee are crucial to graduating within a reasonable amount of time. Nor was it my fellow students, because I did not need help from most of them, and most of them did not need me. The critical group was the research and support staff. These include the research faculty and all the various support positions (the system administrators, network administrators, audio-visual experts, electronic services, optical and mechanical engineers, and especially the secretaries). I needed their help to get my research done, but they did not directly need me. Consequently, I made it a priority to establish and maintain good working relationships with them.

    Cultivating interpersonal relationships is mostly about treating people with respect and determining their different working styles. Give credit where credit is due. Acknowledge and thank them for their help. Return favors. Respect their expertise, advice and time. Apologize if you are at fault. Realize that different people work in different ways and are motivated by different things -- the more you understand this diversity, the better you will be able to interact and motivate them to help you. For certain people, offering to buy them dinner or giving them free basketball tickets can work wonders.

    A true example: at one point in my research, I needed to make significant modifications to some low-level code in the graphics computer called "Pixel Planes 5." Doing this required expertise that I did not have, but another graduate student named Marc Olano did. How should I tap into Marc's expertise and get my necessary changes done?

    The wrong way is to go up to Marc, explain the problem, and get him to make the changes. Marc doesn't need the changes done; I do. Therefore, I should do most of the work. Expecting him to do the work shows disrespect of his time.

    What I actually did was to explain the problem to Marc and he sketched out a possible solution. Then I ran off and worked on my own for a few days, trying to implement the solution. I got part of it working, but ended up getting stuck on another part. Only at that point did I go back to Marc and ask him for help. By doing this, I showed that I respected his time and wanted to minimize his burden, thus making him more willing to help me. Months later, when he and Jon Cohen needed my help in setting up a system to demonstrate some of their software, I was more than happy to return the favor.

    Interpersonal interaction is a huge subject and goes far beyond my description here. All I can really do in this section is (hopefully) convince you that these skills are vital to your graduate student career and encourage you to learn more if you need to improve these skills. I still have a lot to learn myself. I recommend reading The 7 Habits of Highly Effective People and Type Talk (both listed in the References section) as starting points. The magazine article "How to be a star engineer" (listed in the References) also touches on this subject.

Organizational skills
    "Failing to plan is planning to fail."

    Since academia is a type of business, you will have responsibilities that you must uphold. You will be asked to greet and talk with visitors, give demos, show up to meetings, get projects done on time, etc. If you are not well organized, you will have a difficult time meeting those obligations. A technically brilliant student will be greatly hampered if he or she exhibits an "absent minded" personality and develops a reputation for being disorganized.

    There are many different time management and organization skills, and you can find many books on those at your local bookstore. This guide is not going to describe them. Find one that works for you and use it. I can highly recommend Stephen Covey's book, listed in the references. But whatever system you pick, just make sure it works for you. I have never found anyone else who uses my filing scheme, but it is effective for me (by minimizing the combined time of putting away and locating a piece of information). All that really matters is whether or not it works.

    One metaphor I found useful is the following: Organize your tasks as if you were juggling them. Juggling several balls requires planning and skill. You must grab and toss each ball before it hits the ground. You can only toss one ball at a time, just as you can only work on one task at a time. The order in which you toss the balls is crucial, much as the order of working on tasks often determines whether or not you meet all your deadlines. Finally, once you start a task (grab a ball) you want to get enough done so you can ignore it for a while (throw it high enough in the air so it won't come down for a while). Otherwise you waste too much time in context switches between tasks. Do you see jugglers try to keep each ball at the same height above the ground, frantically touching every ball every second?

    Randy Pausch (a professor at CMU) has a set of notes on time management. Three words in his guide summarize the most vital step: Kill your television. He asks you to keep your priorities straight. What is the most important thing to a Ph.D. student? It should be finishing the dissertation, not watching every episode of Friends. That doesn't mean dropping everything else in life, but it does mean knowing what takes priority and allocating time accordingly.

Communications skills
    "What is written without effort is, in general, read without pleasure."
    - Samuel Johnson

    "Present to inform, not to impress; if you inform, you will impress. "
    - Fred Brooks

    I am always amazed that articles written about businesses consistently put good communication skills at or near the top of list of skills that employers want to see in people but rarely find. But you know what? It's true!

    Communication skills, both written and oral, are vital for making a good impression as a Ph.D. student and as a researcher. At a minimum, you have to defend your dissertation with an oral presentation. But you should also expect to write technical papers and reports, give presentations at conferences, and give demonstrations to groups of visitors. If you can write and speak well, you will earn recognition and distinguish yourself from the other graduate students. This is especially true when giving presentations in front of important visitors or at major conferences.

    Conversely, if you cannot communicate well, then your career options after graduation will be limited. Professors spend most of their time communicating: teaching, fundraising, guiding graduate students, and documenting their results (through papers, videos, viewgraphs, etc.) In industry, we need people who can communicate well so they can work in teams, learn what businesses and customers need, present their results, raise funds, and transition to leadership roles in projects and personnel management. If you are technically brilliant but are incapable of communicating, then your results will be limited to what you can accomplish alone and your career growth will be limited, both in industry and academia.

    Unfortunately, not all graduate students receive training in giving presentations or writing technical documents (which are different from English essays). These are skills that can be learned! Don't worry if giving presentations and writing papers are not something that comes naturally to you. I was not very comfortable giving oral presentations when I started graduate school, so I made a concerted effort to learn how to do so, by taking classes, reading about the subject, and practicing. It's not easy, but it's well worth the investment. If you need practice, try giving informal talks at research luncheons, joining Toastmasters, and studying good speakers to see what they do.

    Covering everything about this subject would fill a guide by itself (check out the SIGGRAPH page on preparing and giving presentations), and would probably better done through a videotape than a written document. But here are a few basic points:

  • Organization counts. Within the first few paragraphs or first few minutes, tell me why I should read your paper or listen to your talk. Make it clear where we are going and what we have already covered.
  • Make the text in your slides large enough so that people sitting in the back can read them. For large presentation halls, this usually means no more than 6-7 lines per slide and 28 point type minimum. You'd be surprised how many experts on visualization (especially tenured professors!) give presentations with unreadable slides.
  • Variety retains interest. Vary your pace, tone, and volume. Emphasize the important points. Look around the room. Throw in some video, pictures, or live examples.
  • Don't stand in front of the screen and block everyone's view. You'd be surprised how often people do this without realizing it.
  • Point out the limitations of your work. That helps your credibility. Similarly, give credit where credit is due.
  • Make friends with the A/V crew! Running A/V is a thankless, negative reinforcement job. If everything runs smoothly, well, that's what was supposed to happen so nobody says anything. But if anything goes wrong, the entire audience looks back at the control room. Help the A/V people help you. Always check in early and test the equipment. Tell them what you are going to do in your presentation (e.g. I'm running 3 video segments). Make sure you know how everything works long before you come up to the podium. And thank the A/V crew for their help after you are done!

    Confidence is the key to giving a good presentation. And the way to gain confidence is to give good presentations. When you're just starting out, this is a Catch-22. However, once you become good enough, this turns into a positive feedback cycle that can make giving talks a pleasure.

    Writing papers and getting them published is vital for Ph.D. students who want to get jobs in research after graduation. Your ability to write well significantly improves the chances that your paper will be accepted. When I was a young graduate student and read a paper that I didn't understand, I thought "Gee, I must be dumb." Today I will read the same paper and think "Boy, this is a lousy paper. The authors did not do a good job explaining and presenting their work." If I am reviewing that paper, such a reaction is enough for me to reject the paper.

    Where do you submit your papers? Your professors will help you with this choice, but in general I would suggest shooting for the best conferences or journals where you think it has a reasonable chance of being accepted. It's not much more work to write, submit and present a paper in a highly respected venue than in less respected venues. And if you don't shoot for the top you'll never know if it would have made it. The field of computer graphics is a bit unusual in that the most desirable place to publish is a conference (SIGGRAPH), rather than a journal. Be aware that journals can take years to publish submitted papers; the turn-around time is much faster in a conference.

    Finally, don't forget to communicate with your professors and your teammates. Keep your committee appraised of your progress. One thing I do (which few others do) is write short (1 screenfull) status reports, which I religiously e-mailed to my professors and team members on a weekly basis. These serve as an efficient way of keeping everyone up to date on what I'm doing. They are also a good way for me to record my progress. If I need to remember what I got done during a six month period, I have plenty of old status reports that I can read. You'd be amazed how appreciative professors and managers are of this simple practice. I also throw in a different humorous quote at the end of each week's report to reward people for reading it.

    When you are working in the lab and you reach a milestone or achieve a result, let people know about it! Bring in your professors and fellow students and show it off! That's a win-win situation. It lets others know that you are making progress and achieving results, and you get valuable feedback and advice.

Choosing an adviser and a committee
    "Some students in the lab are only nominally supervised by a thesis advisor. This can work out well for people who are independent self-starters. It has the advantage that you have only your own neuroses to deal with, not your advisor's as well."
    - from "How to do research at the MIT AI Lab"

    The choice of an appropriate adviser is crucial to successfully completing the Ph.D. Your adviser must be someone who can cover your area of specialization and someone you can get along with. When I started graduate school, I thought the adviser - student relationship was supposed to be very close, both professionally and socially. In reality, the relationship is whatever the professor and the student choose to make of it. It can be close, with invited dinners at the professor's home, or it can be distant, e.g. meeting once per semester just to remind the professor that the student is still alive.

    One basic question in choosing an adviser is whether to pick a junior (non-tenured) or a senior (tenured) professor. Non-tenured professors tend to travel less and are generally more available. It is difficult to get help from an adviser who is never in town. Non-tenured faculty have fewer advisees that you have to compete with to get time with the professor. They are more likely to be personally involved with your research -- writing code, spending time in the lab at midnight, etc. Non-tenured faculty must be energetic and hard working if they want to be awarded tenure, and this work habit can rub off on their students. However, tenured faculty have several advantages as well. They are usually the ones with most of the money and resources to support you. They do not have to compete with their students for publications and recognition. The advisee does not run the risk of having his or her adviser not getting tenure and leaving the university. Tenured faculty are more experienced with "how the game works" and thus may be better sources of guidance, personal contacts, jobs after graduation, etc.

    I ended up with a non-tenured professor (actually, he was not even on the tenure track at the time) as my adviser, but also put several tenured professors on my committee, including some of the most senior ones in my specialty. In that way, I got the best of both worlds: the day-to-day attention from the primary adviser, combined with the resources and experience of the committee.

    Professors develop reputations amongst graduate students. Some are known to graduate their Ph.D. students rapidly. Others are impossible to get hold of, so their students take forever to finish or leave without graduating. Some dictate what their advisees have to do, while others are accommodating of student interests. Ask around. What you learn may be revealing. And if circumstances change to make another professor a more appropriate match to your needs, don't be afraid to switch if that is an overall win.

    When picking a committee, you want to make sure they can cover all the areas of your thesis. You also want to make sure that it is likely that all the committee members will be available for meetings! Including too many professors who travel often will make it difficult to get all five or six together in one room for a three hour oral exam or proposal meeting. When scheduling such meetings, start by finding times when the difficult-to-reach professors are in town, and then add in the other committee members.

Balance and Perspective
    "Life goes by so fast, that if you don't stop and look around, you might miss it."
    - from the film Ferris Bueller's Day Off

    "Generally speaking, people provide better maintenance for their cars than for their own bodies."
    - Scott Adams, The Dilbert Future

    When I was in graduate school, my top priority was crystal clear to me: getting out with a Ph.D. Other people described me as "focused like a laser beam" on that goal. In retrospect, I may have been too focused. There is more to life than graduate work. Keeping your health and your sanity intact are both vital to achieving the primary goal of getting out.

    Repetitive Strain Injury (RSI) is a major occupational hazard in our industry. Carpal Tunnel Syndrome is just one type of RSI. If you do not know how to set up your workspace for good ergonomics, learn now! The Pascarelli reference at the end of this guide is a good book on this subject. Over a dozen of my friends and coworkers have been inflicted with this problem. In severe cases, RSI can be a career-ending injury. If you can't type, it's rather difficult to write papers, computer programs, presentations, etc. Don't let this happen to you! Prevention is the way to go. Recently I have been working with weights to strengthen my shoulders and wrists as an additional preventative step.

    Earning a Ph.D. is like running a marathon. You have to learn to pace yourself and take care of your body if you want to reach the finish line. Unfortunately, students often act like sprinters running a marathon. They are highly productive for a while, but then fall by the wayside because they aren't eating correctly, exercising, taking time out to recharge their batteries, etc. You maximize your long-term productivity by not ignoring those other aspects. While I was in graduate school, I took time out to travel up and down the East Coast, from Boston down to Orlando. That was an important part of keeping my stress down and recharging my batteries. I also did some running and circuit training for exercise. For shorter breaks, I shot nerf basketballs at a tiny hoop mounted in the graphics lab and kept a guitar in my office. Figure out what works for you.

    It's easy to lose perspective while in graduate school. You are surrounded by so many other smart, hard working people that it is easy to feel inferior and lose self-esteem and confidence. But without an underlying confidence that you do have what it takes to complete a dissertation, it's too easy to drop out when the going gets tough instead of sticking it through. I found it useful to keep in touch with the "real world," to remind myself that the graduate student population is not representative of humanity in general and to keep my perspective. You got into graduate school because you have already shown to your professors that you have potential and skills that are not typical among most college students, let alone most people -- don't forget that.

The Ph.D. job hunt
    On r‰sum‰s: "The closest to perfection a person ever comes is when he fills out a job application form."
    - Stanley Randall

    Real World, The (n.): Where a computer science student goes after graduation; used pejoratively ("Poor slob, he got his degree and had to go out into the REAL WORLD."). Among programmers, discussing someone in residence there is not unlike talking about a deceased person."
    - the fortune program

    Ideally, the job hunt begins years before you graduate. Networking is very important: while you are in the middle to late phases of your graduate studies, try to get yourself noticed by professors and industry people at other sites. One way to do this is to offer to give a talk about your work at another site. This is not that difficult to do, since most research places love to host seminars and bring in fresh ideas. Attending conferences and working elsewhere during the summer are other ways to get exposure. Make friends with graduate students and personnel at other schools. Make and carry your own business cards. Schmooze with important visitors during major site visits. For about two years, I ran the informal "Graphics Lunch" symposia at UNC. That means I was the point of contact for many speakers who visited UNC and that helped me make contacts. There is also a "star" system that exists. Certain outstanding graduate students can get labeled as "stars" by their professors and that can be an enormous help in getting an interview at CMU or other prestigious locations. It's nice if you can get on that track but one shouldn't rely upon it!

    Networking is important because many jobs are found and filled that way. I got my position at HRL partially because I visited there, at my own expense, two years before I even started my job hunt. That meant that when I circulated my r‰sum‰, I was more than just a piece of paper to them. You are not going to be looking for job ads in the newspaper. Instead, you'll look for announcements in major journals, at conferences, on the Net, and through your contacts. For industrial positions, it is crucial to get past the Human Resources department and find the individual with the ability to hire and deal with that person directly.

    When do you start asking for interviews? You can start when you are able to give a talk about your dissertation work. Don't be too early or too late, because you only get one chance per site. Academic positions generally have a particular "season" (much like getting admitted to school) that starts in the Fall and ends around April; industrial positions generally don't follow that. The job hunt and interviewing process can take months; factor that into your time allocation.

    The job supply and demand situation can vary dramatically in a few years. For example, during the time I was job hunting (end of '94 to early '95), good positions were not easy to find. If I had a dollar for every site that told me "We don't have a permanent position, but would you take a postdoc?" I could buy a lot of lunches. However, since 1997 the graphics job market has been very strong, with many individuals getting multiple offers with high salaries. 1998 was an excellent year for people looking for tenure-track graphics faculty positions. I know many friends who found good tenure-track positions that year. And unless you haven't been paying attention, you know that the demand for C.S. majors is quite high in industry right now.

    Before starting the job hunt, determine your goals and parameters in advance and the "angle" you will take to sell yourself. For example, my strength was in systems, so I chose to emphasize that in my cover letters. Customize your approach to each site, if time permits. What you do for your thesis determines who will and who won't take a look at you. Try to get at least one reference from outside your university.

    This guide is not going to cover the basics of interviewing; you can get that from many books (e.g. the Martin Yate and Bob Weinstein books listed in the references). However, I will mention some tips. Don't interview on the day of arrival, and try to avoid Mondays and Fridays. Be prepared for hard or illegal questions, but you probably won't get them. Do your homework on each site before interviewing! It continually amazes me that people show up for interviews without knowing anything about the institution they want to join. If the target is a research lab for a major company, you can easily find Wall St. Journal articles, annual reports and stockbroker reports in your library. If your goal is an academic position, check out the Tomorrow's Professor site for guidance. If you interview at a university, get their course catalog and use their numbering scheme to describe the courses you can teach. Interview to find out more about them, not just to sell yourself. Your 45-60 minute research presentation is crucial; make sure you practice it thoroughly. Interviews create interviews. That is, if you've already gone on many interviews at other places, then that makes you appear more desirable since others want you, and that makes it easier for you to get more interviews. Broadcast this fact by keeping your interview schedule on your web page. There is an anecdote about one student who received offers to interview at many different places, but only after Stanford interviewed him! Keep logs on who you talk to, what you talked about, and when. That makes it easier to keep things straight when juggling several contenders. The major conferences in your field are a good place to schedule preliminary interviews to get your foot in the door, because it is cheap for the company or university. The people you need to meet are already there, so that saves them the expense of having to fly you out and house you at their site.

    Offers are a waiting game. Be prepared for lots of frustration. You need a written offer or nothing is official; you should also accept or reject in writing. Negotiate, but be aware of the strength or weakness of your position. Starting salary may not be as important as the type of work, benefits, and growth potential. Drug tests and other factors are becoming more common; you will have to decide how you want to respond to those.

    Ah yes, salaries. Everybody wants to know about those. For academic (tenure track) salaries, you can get typical numbers from the Taulbee surveys, printed in the Computing Research News newsletter and the Communications of the ACM. Realize that these are 9-month salaries. Whether or not you can procure funding that covers 2 or 3 months of summer salary makes a big difference to your bottom line. Also, professors can make money by consulting at rates of $1000-2500 per day, although this is more common among established professors. Figures for industrial salaries are harder to come by. The Maisel and Gaddy references are the only ones I have found that specifically surveys young Ph.D.'s in industry (also see the chart a few paragraphs down). Salaries depend heavily on geography. Silicon Valley is in a league of its own, with salaries far above any other region. But before you decide to move to Palo Alto, remember that the cost of living there is also in the stratosphere. In Sept. 1997, a $60k salary in Indianapolis bought the same standard of living as a $101k salary in San Jose! The cost of living difference is larger today. Decent houses in the Silicon Valley can easily cost half a million or more. More general computer science salary surveys are run by the IEEE and EE Times, available at the JobSmart salary survey site.

    Acquire salary information on your own by making use of your network. Don't ask for someone's salary directly, unless it's someone you know very well and even then be very careful. Instead, bounce figures off people and see how they respond. Do they think the figure you mention is high, low, or about right? By seeing how people respond you can get an idea of what the market range is.

    Factor in benefits and the expected workload into your compensation evaluation. In particular, stock options make up a large fraction of the compensation package for Silicon Valley companies and startups. But that $100k offer may seem less attractive if you have to work 80 hour weeks in that position.

    The type of work and compensation varies dramatically with the types of positions. Academic positions are tenure-track, research staff (non tenured) and postdocs. Tenure-track positions at major universities are fairly hard to come by; you need to be both good and lucky. Read the Feibelman and Ralston references for more details. The tenure-track also requires a lot of hours and dedication. As Randy Pausch put it, tenure is a competitive process where you get compared with the other assistant professors and the already-tenured professors. If they worked 70 hour weeks for six years to get tenure, don't expect to get away with working 40 hour weeks. Postdocs are low paying but good for padding your C.V. if you think you need it to get a tenure track position. Just be sure to read the Feibelman reference, which tells you exactly what you need to do to survive a postdoc. In general, academic positions don't pay as well as industrial positions, but universities offer more freedom, prestige, a richer intellectual environment and the possibility of long-term stability (with tenure). There's a big difference between startups, regular industrial jobs, and industrial research positions. Startups can be the most lucrative financially, although that's a big gamble. Read the Kawasaki and Bell references if you want to work at a startup. Expect to put in long hours while losing contact with the research community. Industrial research lies in an uncomfortable middle ground between production jobs and academic research, and blends the advantages and disadvantages of industry vs. academia.

    Salary curves for Ph.D.'s

    The figure above comes from the November 1999 issue of Computing Research News. It charts academic and industrial Ph.D. salaries vs. experience. The academic figures are the hollow icons, while the industrial figures are the filled-in icons. The diamond, square, and triangle represent minimum, mean, and maximum figures for industry. The circle, square, and triangle represent minimum, mean, and maximum figures for academia. The industrial salaries come from a survey of eight industrial research laboratories, representing 644 individuals. The industrial salaries include the estimated value of bonuses and stock options (but not benefits), and they are 12-month salaries. The academic figures are 9-month salaries, and these come from the Taulbee Survey for all US C.S. academic departments. The 3 year, 9 year and 18 year academic figures are tied to the assistant, associate and full professor levels. The survey was conducted in 1998. While this chart seems to show that academic salaries are far behind industrial, remember that the academic figures are 9-month salaries, so getting two or three months of supplemental pay from research grants makes up much of the difference.

    Compared to last year's chart, the industrial salary numbers are much higher. Also interesting is how flat the chart is for the first few years, which tends to indicate that the competition for hiring has pushed up the salaries of new employees.

    If you are in C.S., I'm sure you know of people who have gone to startups and may have done very well financially, especially if they are in internet or web related fields. I know a few people like that, and at least one who (once his options vest) will be able to retire and never work another day in his life. It is certainly tempting, and some people are leaving school early (without a degree) to join "in the game" before the Internet mania ends. If you have an entreprenurial spirit and you are itching to go to a startup, then maybe you should reconsider whether the Ph.D. is the right route for you. The Ph.D. does tend to maximize your career options after you leave, but you do pay a price of missed opportunities during the years you spend in school. It all depends on what you want and where your priorities are.

    No matter where you go after you graduate, maintain your contacts with your alma mater. You may change jobs and move from place to place, but you will always have your degree from your university. If you keep good relations with your university and your fellow former students, that will serve as an excellent base for your personal network.

Conclusion
    "Dissertations are not finished; they are abandoned."
    - Fred Brooks

    The following story, called "The Parable of the Black Belt," is excerpted from Built to Last: Successful Habits of Visionary Companies, by James C. Collins and Jerry I. Porras.

      Picture a martial artist kneeling before the master sensei in a ceremony to receive a hard-earned black belt. After years of relentless training, the student has finally reached a pinnacle of achievement in the discipline.

      "Before granting the belt, you must pass one more test," says the sensei.

      "I am ready," responds the student, expecting perhaps one final round of sparring.

      "You must answer the essential question: What is the true meaning of the black belt?"

      "The end of my journey," says the student. "A well-deserved reward for all my hard work."

      The sensei waits for more. Clearly, he is not satisfied. Finally, the sensei speaks. "You are not yet ready for the black belt. Return in one year."

      A year later, the student kneels again in front of the sensei.

      "What is the true meaning of the black belt?" asks the sensei.

      "A symbol of distinction and the highest achievement in our art," says the student.

      The sensei says nothing for many minutes, waiting. Clearly, he is not satisfied. Finally, he speaks. "You are still not ready for the black belt. Return in one year."

      A year later, the student kneels once again in front of the sensei. And again the sensei asks: "What is the true meaning of the black belt?"

      "The black belt represents the beginning -- the start of a never-ending journey of discipline, work, and the pursuit of an ever-higher standard," says the student.

      "Yes. You are now ready to receive the black belt and begin your work."

    To me, there are two lessons in this story.

    First, the Ph.D. is the beginning, not the culmination, of your career. Don't worry about making it your magnum opus. Get out sooner, rather than later.

    Second, if you bother to talk to and learn from the people who have already gone through this process, you might graduate two years earlier.

    Good luck.

Other Related Guides Recommended Reading
    Bell, C. Gordon and John McNamara. High-Tech Ventures: The Guide for Entrepreneurial Success. Addison-Wesley, 1991. ISBN 0-201-56321-5.
    A must read if you want to work for a startup.

    Bronson, Po. The Nudist on the Late Shift. Random House, 1999. ISBN 0375502777.
    A fun read, giving the flavor of what working in the Silicon Valley is like. Many of the chapters previously appeared as articles in Wired.

    Covey, Stephen R. The 7 Habits of Highly Effective People. Fireside Simon and Schuster, 1989. ISBN 0-671-70863-5.
    Excellent overall, with sections on time management, guiding principles and interpersonal skills.

    EE Times 1998 Salary Survey Issue. Issue #1023, Monday August 31, 1998.
    A good survey and commentary about industrial salaries. This survey is conducted annually.

    Feibelman, Peter J. A Ph.D. is Not Enough! A Guide to Survival in Science. Addison-Wesley, 1993. ISBN 0-201-62663-2.
    Good discussion of research career paths. A must read if you choose to take a postdoc.

    Kawasaki, Guy. The Macintosh Way: The Art of Guerrilla Management. Harper Perennial, 1990. ISBN 0-06-097338-2.
    Despite problems that occurred at Apple, this book shows the energy and chutzpah required to survive in a startup.

    Kelley, Robert E. How to be a star engineer. IEEE Spectrum (October 1999), 51-58.
    Good description of the skills that are needed to excel at work, which go beyond sheer technical skills.

    Kroeger, Otto and Janet M. Thuesen. Type Talk: The 16 Personality Types that Determine How We Live, Love and Work. Tilden Press, 1988. ISBN 0-385-29828-5.
    Introduction to the Myers-Briggs type indicators, useful for interpersonal relations.

    Maisel, Herbert and Catherine Gaddy. Employment and Salaries of Recent Doctorates in Computer Science. Communications of the ACM 40, 9 (September 1997), 90-93.

    Maisel, Herbert and Catherine Gaddy. Employment and Salaries of Recent Doctorates. Communications of the ACM 41, 11 (November 1998), 99-101.
    One of the few surveys I have seen for recent C.S. Ph.D.s that includes both industry and academic numbers. The low sample size is a problem, however.

    Pascarelli, Emil and Deborah Quilter. Repetitive Strain Injury: A Complete User's Guide. John Wiley and Sons, 1994. ISBN 0-471-59532-2.
    A good introduction to RSI injuries and avoiding them.

    Pastore, Robert R. Stock Options: An Authoritative Guide to Incentive and Nonqualified Stock Options, 2nd edition. (printed Dec. 1999). ISBN 0966889924. PCM Capital Publishing.
    I haven't read this but I have been told this is an excellent reference for those of you fortunate enough to have a bundle of stock options. Give me a few options as a tip for finding this book, ok? :-) The book covers tax and legal issues and gives advice on when to keep or exercise your options.

    Ralston, Anthony. The Demographics of Candidates for Faculty Positions in Computer Science. Communications of the ACM 39, 3 (March 1996), 78-84.
    A must read if you are looking for tenure track positions. The author is a former CS professor who led a faculty search, so if you don't believe what I say, then listen to him.

    Weinstein, Bob. R‰sum‰s Don't Get Jobs: The Realities and Myths of Job Hunting. McGraw-Hill, 1993. ISBN 0-07-069144-4.
    Gritty, realistic job hunting guide for today's market.

    White, Pepper. The Idea Factory: Learning to Think at MIT. Plume (Penguin Books), 1992. ISBN 0-452-26841-9.
    While this is not about C.S., it does dispel the notion of graduate school as an ivory tower environment.

    Yate, Martin. Knock 'Em Dead: The Ultimate Job Seeker's Handbook. Bob Adams, Inc.
    Good generic guide to job hunting and interviews, including a long section on interview questions.


Last updated: Fri March 10, 2000

Questions? Mail to azuma@HRL.com

Ron Azuma's page of guides on CS graduate school

Ron Azuma's home page

Copyright 1997-2000, Ronald T. Azuma, except for portions excerpted from elsewhere