cp's OEIS Frontend

This is a front-end for the Online Encyclopedia of Integer Sequences, made by Christian Perfect. The idea is to provide OEIS entries in non-ancient HTML, and then to think about how they're presented visually. The source code is on GitHub.

A139004 Number of operations A000142 (i.e., x!) or A000196 (i.e., floor(sqrt(x))) needed to get n, starting with 4.

This page as a plain text file.
%I A139004 #56 Feb 28 2025 20:51:18
%S A139004 2,1,10,0,7,11,24,27,29,9,36,40,36,17,37,31,22,31,37,42,19,37,21,1,26,
%T A139004 13,51,41,36,6,30,41,44,33,16,33,31,64,35,50,25,43,12,18,41,18,42,55,
%U A139004 39,23,71,65,45,43,52,39,49,44,51,60,57,59,24,66,26,36,46,51,46,26,48,76
%N A139004 Number of operations A000142 (i.e., x!) or A000196 (i.e., floor(sqrt(x))) needed to get n, starting with 4.
%C A139004 Knuth conjectured that any number can be obtained in this way, starting from 4.
%C A139004 This sequence gives the minimal number of operations needed to do so.
%C A139004 To ensure the sequence is well-defined, define a(n)=-1 if it is not possible to get n in the given way.
%C A139004 If we are allowed to use tan(x) just once, then a single 2 is enough to get any positive integer, if Knuth's conjecture that one 4 is enough is true. From 2, (((-tan(2.))!)!)! = 5.592..., then floor, factorial gets 120, then sqrt, sqrt gives 3.162..., and floor gives 3, or negate, floor, negate gives 4. - _N. J. A. Sloane_, Feb 26 2025
%C A139004 The article by Bendersky is relevant because it gives an explicit formula for n using four 2's (and some logs). Good illustration of techniques. - _N. J. A. Sloane_, Feb 26 2025
%H A139004 Jon E. Schoenfield, <a href="/A139004/b139004.txt">Table of n, a(n) for n = 1..1000</a>
%H A139004 Eli Bendersky, <a href="https://eli.thegreenplace.net/2025/making-any-integer-with-four-2s/">Making any integer with four 2s</a>, Blog Post, Feb 22 2025
%H A139004 Eli Bendersky, <a href="/A139004/a139004.pdf">Making any integer with four 2s</a>, Blog Post, Feb 22 2025 [Local copy, with the author's permission]
%H A139004 D. E. Knuth, <a href="http://www.jstor.org/stable/2689238">Representing numbers using only one 4</a>, Mathematics Magazine, Vol. 37, No. 5 (Nov. 1964), pp. 308-310.
%H A139004 John E. Maxfield, <a href="http://www.jstor.org/stable/2688966">A Note on N!</a>, Mathematics Magazine, Vol. 43, No. 2 (March 1970), pp. 64-67.
%H A139004 Perlmonks.org, <a href="http://www.perlmonks.org/?node_id=443037">Chasing Knuth's Conjecture</a>.
%H A139004 Jon E. Schoenfield, <a href="/A139004/a139004.txt">Table of n, a(n), and shortest path for n = 1..1000</a>
%F A139004 a(4) = 0, a(n) = min { a(k)+1 ; n^2 <= k < (n+1)^2 or k! = n }
%e A139004 Representing the operation x -> floor(sqrt(x)) by "s" and x -> x! by "f", we have:
%e A139004 a(1) = 2 since 1 = ss4 is clearly the shortest way to obtain 1, starting with 4.
%e A139004 a(2) = 1 since 2 = s4 is clearly the shortest way to obtain 2, starting with 4.
%e A139004 a(4) = 0 since no operation is required to get 4.
%e A139004 a(3) = 10 = 3+a(5) since 3 = ssf5 and it cannot be obtained from 4 with fewer operations.
%e A139004 a(5) = 7 since 5 = sssssff4.
%e A139004 a(6) = 11 = 1+a(3) since 6 = f3. a(10) = 9 since 10 = sfsssssff4 is the shortest way to obtain 9, starting with 4.
%o A139004 (PARI) A139004( n, S=Set(4), LIM=10^4 )={ for( i=0,LIM, setsearch( S, n) & return(i); S=setunion( S, setunion( Set( vector( #S, j, sqrtint(eval(S[j])))), Set( vector( #S, j, if( LIM > j=eval(S[j]), j!))))))}
%o A139004 (PARI) { search(x,r,l=0) = local(ll,xx); ll=l; xx=x; while(ll<L, if(xx==r, L=ll; print(L); return); ll++; if(xx*(log(xx)-1)<2^(L-ll)*log(r), search(xx!,r,ll)); xx=sqrtint(xx)) } \ where L - upper bound, x - starting value, r - final value; e.g., to compute a(4), run: L=32; search(4,8) \\ _Max Alekseyev_, Nov 01 2008
%Y A139004 Cf. A139003, A055226.
%K A139004 nonn
%O A139004 1,1
%A A139004 _M. F. Hasler_, Apr 09 2008
%E A139004 a(7)-a(9) from _Max Alekseyev_, Oct 17, Nov 01 2008
%E A139004 More terms from _Jon E. Schoenfield_, Nov 10 2008