Sunday, July 31, 2005

The Secrets of Success

What does it take to be a successful in our profession?
  • Intelligence. You need an innate talent in different forms to succeed as a scientist.
    • Problem Solving. Using well-established techniques in the appropriate way to find solutions.
    • Creativity. Original research means one needs to look beyond the current set of tools and develop new approaches to problems.
    • Vision. Discovering new problems and directions of research.
  • Hard work. Enough said.
  • Luck. Working on the right problem at the right time. If you work long enough` the law of averages will catch up with you (for good or for bad).
  • Discipline. The discipline to focus on research for a period of time without getting distracted from other responsibilities or by the internet or other activities. Some people find it best to schedule time for research and hole themselves up somewhere to think about a problem.
  • Commitment. Be willing to spend a considerable amount of time on a problem even if you keep running into dead ends.
  • Training. Taking and working hard in classes. Having and taking advantage of a good advisor. Reading papers and textbooks. When you see a theorem in a paper try to prove it yourself first. Only then can you truly appreciate a proof and learn from it.
  • Colleagues. Having co-authors, especially those that complement your talents, can help you do more than you could on your own. But just having good people to talk to, to bounce off proof ideas and discuss research directions can greatly help you find the right approach to a problem.

9 comments:

  1. Is there something that you wish you had realized when you started your career?

    ReplyDelete
  2. There's a very well written essay by Paul
    Graham about this topic with the same conclusion (or points). What You'll Wish You'd Known.

    ReplyDelete
  3. -- Is there something that you wish you had realized when you started your career?

    Innate talent is overrated and hard work is underrated.

    ReplyDelete
  4. Lance,

    Can you please post (or point to an earlier post), on what a "good advisor" is? Is it one who sits with you closely and monitors your progress on solving a problem, perhaps even giving his own suggestions to solve the problem? Or is it one who supervises you from a high-level, not getting into the details of the specific problem you are solving, but only giving high-level advice?

    I can see advantages to either kind of advisor: the former helps you solve the problem, while the latter helps you develop your own individuality as a researcher.

    Can you please comment on this?

    ReplyDelete
  5. I did have a post on choosing an advisor. As to your question, it depends on the particular student and their personality and skill level.

    ReplyDelete
  6. Another great text on this is Richard Hamming's "You and Your Research".

    - Wim

    ReplyDelete
  7. ... from the "do as I say, not as I have done" file.... some things that I wish that I had known or put into practice earlier:

    Just because you have had some success does not mean you should work only on big problems from here on out. This is a recipe for frustration and limited output. Small problems are good for getting into new areas, meeting new colleagues, keeping your spirits up, and keeping your publication list presentable.


    That being said, you need to work on some important problems. Just don't confuse "difficult" for "important". Some obscure areas are very difficult, and get a lot of respect from people "in the know" for being hard. But, to the uninitiated, these same topics can seem obscure to most theoretical computer scientists, and irrelevant to the vast majority of computer scientists. This can bite you hard when it comes time to look for jobs.

    ReplyDelete
  8. Add Malleability. While a good researcher needs to have certain level of self confidence, it shouldn't be so much that it gets on the way for learning from one's supervisor, mentor, colleagues and in due time one's own students.

    ReplyDelete
  9. If, by "successful", you mean "able to get a tenure-track job", then I would say, work in a hot area. First, it'll be much easier to publish, since, if you find even a minor result, the many people currently working in that area will be interested and want to hear about it. Second, universities are more likely to want to hire someone from that area.

    That is, for the same amount of ability and hard work, working in a hot area vastly improves your odds of success.

    ReplyDelete