At 10:37 AM -0500 2/26/03, Perry E. Metzger wrote:
>Ignatios Souvatzis <is@netbsd.org> writes:
>>  > Not necessarily - that's why the +1.
>>  >
>>  > |---high bit 0---||---highi bit 1---||---high bit 0---|
>>  >
>>  > If the last value of the first "high bit 0" equals the first value of
>>  > the second "high bit 0", it's just generated a duplicate, 2^15+1 from
>>  > the first copy to the second.
>>
>>  err, right. Stupid me.
>
>I thought we were generating an ergodic sequence here, in which case
>that isn't possible...

I think this qualifies as a sequence of ergodic sequences.  The 
behavior of the high bit doesn't qualify, and the individual 
fixed-high-bit sequences have no guarantees w.r.t each other.  The 
2^15+1 repetition case is of course unlikely, just not guaranteed 
impossible.

Did that rather abstract response actually help?
-- 
The opinions expressed in this message are mine,
not those of Caltech, JPL, NASA, or the US Government.
Henry.B.Hotz@jpl.nasa.gov, or hbhotz@oxy.edu