trough@ihuxa.UUCP (Chris Scussel) (03/11/84)
OK, now we finally have a real answer to the chain puzzle (from Dave Seaman).
I was sorry to see that it stirred up such a fuss about infinite/indeterminant
distributions. I still maintain that the range of the relevant distribution
is not infinite, since "required lengths" larger than the original length
of the chain can't be obtained regardless of where the chain is cut. I was
also sorry to see the assumption of a particular distribution in order to obain
an answer. Shouldn't there still be a best answer even with no knowledge of the
distribution (possibly based on the assumption that all distributions are
equally likely; that should be amusing)? When I "guessed" a uniform distribution
I came up with same answer as Seaman: L/3. Interesting. Any takers on the n-piece
problem, or even the 3-piece one? Seems to be quite a bit more complicated.
Chris Scussel
Bell Labs
AT&T Bell Laboratories (sorry!)
Naperville, Illinois
{AT&T BL}!ihnp4!ihuxa!trough