[seqfan] Re: futurevalue

franktaw at netscape.net franktaw at netscape.net
Mon Aug 17 07:14:04 CEST 2009


Actually, I will modify my earlier statement: the program in A066274 is 
worth retaining.  There may be a few others.

I do think that any call futurevalue(x,c,z) with c a constant ought to 
be replaced by x*(c+1)^y, and the vast majority of these programs are 
of this form.  Many of them will then be trivial, and often duplicate 
existing programs.

And really, EVERY program from Zerinvary needs to be reviewed, and 
many, probably most, should be deleted.  Consider the following from 
A000312: n^n (attributions deleted):

   seq(mul(n, k=1..n), n=0..17);

   a:=n->mul(sum(n*(-1)^j, j=0..20), k=1..n): seq(a(n), n=0..17);

   a:=n->mul(denom (1/(n+1)), k=0..n): seq(a(n), n=-1..16);

   a:=n->mul(1+add(1, j=1..n), j=0..n):seq(a(n), n=-1..18);

   restart:a:=n->mul(sum(1, j=0..n), k=0..n): seq(a(n), n=-1..16);

First, we've got this repeated use of mul(x, k=1..n) instead of x^n; 
this is pure obfuscation.  Likewise, "sum(n*(-1)^j, j=0..20" is an 
obfuscated way of writing "n".  Also "denom (1/(n+1))" is just "n+1", 
and "sum(1, j=0..n)" is a terrible way to write "n+1".  All of these 
are just obfuscated ways of writing n^n or (n+1)^(n+1).

On the Mathematica side, we've got:

   Table[Hyperfactorial[n]/Hyperfactorial[n - 1], {n, 0, 17}]

which is marginal; this would make a better formula than a program, as 
one would never use this method to compute the sequence.

Finally, there's:

   (Other) sage: [log(e^(n^n))for n in xrange(0, 10)]

and I think that this silliness speaks for itself.

Franklin T. Adams-Watters

-----Original Message-----
From: N. J. A. Sloane <njas at research.att.com>

I do NOT agree that all the "futurevalue" Maple commands
should be deleted.

Neil




More information about the SeqFan mailing list