[seqfan] Re: Typo in B-file for A063539

[Hugo Pfoertner]

I made a fit with Giovanni's and my data, taking into account the error
term from the Ramaswami article from 1949, which is now consistent with the
asymptotic factor of (1-log (2)) or 1/(1-log(2)) ~=3.26... If anyone likes
to take a look at it, that's in the draft version of A063539,
https://oeis.org/draft/A063539. Perhaps there is also an explanation for
the value of the factor -0.594... for the "x/log(x)" error term. If anyone
can put it better, please do so.

Hugo

On Sun, Apr 12, 2020 at 11:50 AM Giovanni Resta <g.resta at iit.cnr.it> wrote:

> I got a(10^9) = 3571942311, a(10^10) = 35410325861, and a(10^11) =
> 351498917129.
>
> Giovanni
>
> Il 11/04/2020 22:37, Hugo Pfoertner ha scritto:
> > The convergence to an asymptotic factor of 3.26.. is actually very slow.
> So
> > far I have calculated the following values:
> > A063539(10^6)=3697669, A063539(10^7)=36519633, A063539(10^8)=360856296
> > It will take some time to get the next value A063539(10^9).
> >
> > Hugo Pfoertner
> >
> > On Sat, Apr 11, 2020 at 3:51 PM Neil Sloane <njasloane at gmail.com> wrote:
> >
> >> Dear Jean-Paul, Allan, et al.
> >>
> >> I'm glad this mystery has been cleared up!  Could one of you add some
> >> comments (and references) to the appropriate sequence?
> >>
> >> Best regards
> >> Neil
>