# [seqfan] Re: A051501

Joerg Arndt arndt at jjj.de
Fri Jul 31 06:00:28 CEST 2009

```? b=1.2516475977905
1.25164759779050
? floor(2^b)
2
? floor(2^2^b)
5
? floor(2^2^2^b)
37
? floor(2^2^2^2^b)
137438953481
? floor(2^2^2^2^2^b)
*** floor: precision too low in truncr (precision loss in truncation).
? (2^2^2^2^2^b)
4.07049553396324 E41373247570

What about "The next term has more than 41*10^9 decimal digits."
I'd simply remove the mention of the largest known prime.

* franktaw at netscape.net <franktaw at netscape.net> [Jul 31. 2009 13:45]:
> I have a couple of problems with
> http://www.research.att.com/~njas/sequences/A051501
> I'm not sure how to fix them.
>
> First, the comment from T. D. Noe; specifically the statement that the
> largest known prime ... is only 2^32582657-1.  This statement is out of
> date; as far as I can tell, the largest known prime is currently
> 2^43112609-1.  This could obviously be corrected; but, it will likely
> become out of date again.  I guess what is needed is a reference to a
> web site with the largest known primes.
>
> Second, the "Extension": "The next term is too large to display and in
> any case b is not known sufficiently accurately to compute it."  This
> suggests that one would compute more terms of the sequence by getting a
> sufficiently accurate value of b, and plugging it into the formula.  In
> fact, just the opposite is the case: one would get a more accurate
> value of b by determining the next term of the sequence, and working
> backwards to determine what value of b that corresponds to.  (Not that
> anyone is likely to do that anytime soon.)  I'm not sure how to reword
> this so that it is less misleading.
>