# The Magic of Base 3

**Gordon Mohr**
gojomo@usa.net

*Thu, 25 Oct 2001 11:22:42 -0700*

>* Note that the article suggests that it should be e-valued, and that 3
*>* is simply the closest integer.
*
I am trying to imagine, but have not yet managed to imagine, how
non-integral bases would "work", while still preserving the
[symbols x places] efficiency posited by the article.
>* "Doron Zeilberger of Rutgers University, in a paper co-authored with
*>* his computer Shalosh B. Ekhad, established that..."
*>*
*>* seems perhaps more important than any particular ternary properties.
*
I missed that the first time through. Turns out that
Zeilberger likes to credit his computer as a coauthor.
His "servant" even has a homepage:
http://www.math.temple.edu/~ekhad/
- Gordon