About a year ago, in a discussion on the nature of object-oriented
programming here on HP3000-L, I reposted one of my postings from
sci.bio.evolution on the nature of pleiotropy and modularity. I claimed then
that the advantages of evolving modularity are so great that its discovery is
one of the those inevitable qualities associated with the design of any form
of coding structure.

I just got back from spending the week at a very enjoyable NASA Institute of
Advanced Concepts meeting in Atlanta. One of the people I spent several hours
talking to was Hod Lipson, a young assistant professor at Cornell. His and my
interests overlap extensively and he pointed me to a recent paper of his that
I thought that some of you might also enjoy reading. He comes to the same
conclusion: the evolutionary origination of modularity in code structure is
inevitable, and he and his co-authors, Jordan Pollack and Nam Suh, published
this work in the journal Evolution this last August:

     http://www.mae.cornell.edu/lipson/papers/EvolutionPress.pdf

One moral of this work might be said to be: "No matter how poor a programmer
you are, or how dumb you might be, or how randomly your write your programs,
you will eventually discover the benefits of highly modularizing your code."

Hod was also involved during his postdoc at Brandeis with Jordan Pollack in a
bit of work on self-reproducing robots that made all of the newspapers for a
bit of time. For some his press clippings, see:

     Washington Post
          http://www.mae.cornell.edu/lipson/press/WashingtonPost.pdf

     Wall Street Journal
          http://www.mae.cornell.edu/lipson/press/Wsj_a.jpg

     Time Magazine
          http://www.mae.cornell.edu/lipson/press/TIME_a.jpg

Wirt Atmar

* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *