If yours is a specific question about phi, then the
answer is that you deduce from your algorithm the
relationship between the number of iterations and
the number of digits.
If it is a general question, then what you can do is to
compute 0j_100":y (or however many digits are required),
and terminate the iteration when that result does not
change. You also need to know by other means that there
is a straightforward relationship between the number of
digits and the number of iterations (for example, you
need to know that the number of digits increases as the
number of iterations increases).
----- Original Message -----
From: Oleg Kobchenko <[EMAIL PROTECTED]>
Date: Monday, July 10, 2006 12:39 pm
Subject: [Jprogramming] 1000 golden digits
> Is there an established routine to get
> arbitrary decimal precision number calculations?
>
> _51]\ ({.,'.',}.)": <.(10x^100)* ([^:((=)&([: <. (10x^100)&*))
> (1+%))^:_] 1x
> 1.6180339887498948482045868343656381177203091798057
> 628621354486227052604628189024497072072041893911374
>
> http://www.research.att.com/~njas/sequences/table?a=1622&fmt=4
>
> _51]\ ({.,'.',}.)": <.(10x^1000)* ([^:((=)&([: <.
> (10x^1000)&*)) (1+%))^:_] 1x
> 1.6180339887498948482045868343656381177203091798057
> 628621354486227052604628189024497072072041893911374
> ...
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
