Thursday, August 7, 2014

Keep it simple

KISS = Keep It Simple, Stupid!

"Make things as simple as possible, but no simpler."
paraphrase of Einstein, for the real quote and its evolution, see here.

A "principal object of theoretical research ... is to find the point of view from which the subject appears in its greatest simplicity." - J. W. Gibbs

This post is a mixture of rant, observation, and exhortation.
It is not just about theoretical model building but all of life! I also value simplicity in language, communication of ideas, administrative procedures, personal finances, software, email lists, meeting schedules, BibTex (arghh!), TV remote controls, .... But it seems that when it comes to things like grant applications, course profiles, .... there is a drive for greater complexity... The bigger the better. The more parameters the better...

I sometimes wonder if the value I place on simplicity is just a personality trait or preoccupation.  Or  is there some universal value that is being lost? It seems there are several issues. Do some people not do it because they think complicated is better? Or they are just lazy, or they just don't have the necessary ability to simplify? I take comfort that I am in the good company of Martin Gutzwiller, but he said he was "fighting a losing battle" and he died earlier this year.

Simplicity in science is hard work. Sometimes it requires brilliance. But other times it just requires some effort. Giving a simple physical picture for the results of a detailed computation may not be easy. But, shouldn't one at least make an attempt? When writing a paper first give a simple summary of the essence of the argument. When plotting a graph explain what are the relevant scales, e.g., if there is a maximum in resistivity versus temperature, describe the essential energy/temperature scale involved....
Then there is the whole topic of talks and seminars. Most need drastic simplification to have any chance of being comprehensible...

I should stress that I not against highly technical calculations or big computer simulations, per se. They can be incredibly valuable, particularly when they are used in conjunction with and enhance physical and chemical insight. Furthermore, good governance and accountability may need to be complicated. But, there has to be a balance.

Am I the only one concerned about this?


  1. I have the impression the perception of people doing things too complicated is that the people "above" often are more impressed by complicated messages. Because then you're really doing something.
    Quite often (evaluations) the "people above" are managers. Sometimes scientists.

    So it's at least part perception (whether correct or incorrect) by the offenders, and I think it ties into the current climate of with the way performance is measured.

    1. pcs,

      Thanks for the insightful comment. I had not thought of this aspect of the problem. I think this is particularly a problem with big computer simulations, as recounted in this earlier post

      But, this is not the sole source the problem. Some scientists just seem to love jargon and being part of an exclusive group that knows all the technical details.

  2. Ross, you asked "Am I the only one concerned about this?"
    My answer: No.
    We all should aim to be able to communicate in clear and simple terms what the key findings of our work are. These need to be backed up by detail, but if there isn't an overall message, the detail doesn't matter much.

    1. David, glad to hear you are concerned too. I hope as a School Chair at Georgia Tech you press this issue with your colleagues, particularly younger teachable ones.

  3. Great post! Also reminds me of what Hemingway once said about his own work: that his books were easy to read because they were that much more difficult to write.