Re:Mathematics: Came across the following gag. Is it true?
by xyzzyz on Friday October 19, 2012 @04:52AM
(#4672760)
This is not true. For instance, sequence 00 does not occur in sequence
010110111011110111110111111...
which can hardly be described as "repeating" (or periodic) sequence.
The point here is normality. Fix alphabet A having b letters. We say that a sequence S of elements of A is normal if, for every word w, we have
lim n->inf C(S, w, n)/n = b^(-k)
where C(S, w, n) is the number of occurrences of w in first n letters of S, and k is length of w. More intuitively, S is normal if every word occurs in S with the same asymptotic frequency.
So, if a sequence does not contain some word as a substring, it cannot be normal by above definition.
010110111011110111110111111...
which can hardly be described as "repeating" (or periodic) sequence.
The point here is normality. Fix alphabet A having b letters. We say that a sequence S of elements of A is normal if, for every word w, we have
lim n->inf C(S, w, n)/n = b^(-k)
where C(S, w, n) is the number of occurrences of w in first n letters of S, and k is length of w. More intuitively, S is normal if every word occurs in S with the same asymptotic frequency.
So, if a sequence does not contain some word as a substring, it cannot be normal by above definition.