> 4 : '(g^:x)y' > > > I wonder why f is not translated into tacit code.
Bo, although several alternatives were offered, I did not see an answer to your tacit question. You can find the reason by looking at the "Proof of Completeness" in http://www.jsoftware.com/jwiki/Essays/Trains . http://www.jsoftware.com/jwiki/Essays/Trains The short answer is because x is used as an argument to a conjunction. You might also be interested to see the results of searching (Google/Bing) using the query: Collatz site:www.jsoftware.com. Incidentally, the sequence A008908, a relative of cn ( cn=. _1 , A008908 @: <: ) can be fixed in extended J as, A008908=. 1:`(1 + $:@:(%&2`(1 + 3&*)@.(2&|)))@.(1&<)"0 M. (103!:0)@:>:@:I. A008908 71 1 2 8 3 6 9 17 4 20 7 15 10 10 18 18 5 13 21 21 8 8 16 16 11 24 11 112 19 19 19 107 6 27 14 14 22 22 22 35 9 110 9 30 17 17 17 105 12 25 25 25 12 12 113 113 20 33 20 33 20 20 108 108 7 28 28 28 15 15 15 103 On Thu, Sep 19, 2013 at 5:16 PM, Bo Jacoby <[email protected]> wrote: > g=.-:&(+ (2&|*>:&+:)) NB. (g y)= 2%~1+3*y if y is odd, (g y)=2%~y if y > is even. > f=.13 : '(g^:x)y' > f > > 4 : '(g^:x)y' > > > I wonder why f is not translated into tacit code. > > It works nevertheless: > 45 f 27 > 4616 > ---------------------------------------------------------------------- > For information about J forums see http://www.jsoftware.com/forums.htm > ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
