Monday, March 14, 2005

What Makes a Good Collaborator?

This week I visit CWI in Amsterdam where I spent my sabbatical year eight years ago and continue working relationships with many people here especially Harry Burhman. Harry is my strongest collaborator in the sense that I have written far more papers with him than any person. So what makes for a good collaborator?

  • Strength—A good collaborator should of course be a strong researcher in my area of interest and Harry certainly fits that bill. But there are many great complexity theorists I have hardly or never worked with.
  • Compatibility of Strengths—The strengths should complement each other nicely. Good collaborators know their areas well and can quickly focus the inherently difficult parts of a problem and have different tools and approaches they can bring to the table.
  • Respect—Good collaborators need to trust and respect each others ability and judgment.
  • Philosophy—Long-Term collaborators need to share beliefs on what problems are important and worth working on.
  • Personality—You need to have a friendly relationship outside of work. It helps immensely if your respective families get along.
  • Luck—Finding the right problems to work on together at the right time. You need a good first collaboration before you start making time for further collaborations.
  • Distance—This seems counterintuitive but two people in the same geographical area rarely have a long history of collaboration. It's hard to make time for working together when you are in close proximity. Also two people who see each other constantly get tired of working with each other no matter how compatible they are. Better to keep in email contact and have several short and long visits where one can allocate time for the other.
There is something else that I can't perfectly describe where something just "clicks" when you have someone you can work with well.


  1. I've always been sort of puzzled by the distance aspect. It seems pretty rare for people at the same university to work together. The only exception is when the collaborators are from different departments (could it be working in a forgein area forces closer contact in order to be effective?).

    Are there any other reasons this would be the case?

  2. Well, how often do you find two researchers at the same university working in very similar areas?

  3. An off-topic comment about your posting:
    It was made at 3:14 AM on 3/14 :-)

  4. Given that it was probably 10:14am at CWI there cannot have been many chances for collaboration at that moment indeed:-)

  5. Not sure if the distance aspect is essential to good collaboration, but rather an accident of the university system. I mean, what about Hardy and Littlewood? Hardy and Ramanujan? Maybe distances were longer back then, or these were isolated exceptions. But I do feel that good collaborations require personal contact.