>(x&f)^:n y is the same as x f^:n y, by definition.

Right.  Regarding the expressions ( x f^:3 y ) and  ( (f&y)^:3 x ), they can be 
clearly different (even for an associative f) ; using symbolic verbs,

   (X u ^:3 Y)
(X u (X u (X u Y)))
   
   (u&Y)^:3 X
(((X u Y) u Y) u Y)

Ambrus already answered the original question, again,

   (X u ^:3 Y)
(X u (X u (X u Y)))

or indeed equivalently,

  (X&u)^:3 Y
(X u (X u (X u Y)))


   

-----Original Message-----
From: Roger Hui <[email protected]>
Sent: Saturday, March 20, 2010 1:52 PM
To: Programming forum <[email protected]>
Subject: Re: [Jprogramming] recursive composition

(x&f)^:n y is the same as x f^:n y, by definition.



----- Original Message -----
From: Bo Jacoby <[email protected]>
Date: Saturday, March 20, 2010 8:46
Subject: Re: [Jprogramming] recursive composition
To: Programming forum <[email protected]>

> Hi John
> but
>    (x&f)^:3 y
> _2
> - Bo
> 
> --- Den lør 20/3/10 skrev John Randall 
> <[email protected]>:
> Fra: John Randall <[email protected]>
> Emne: Re: [Jprogramming] recursive composition
> Til: "Programming forum" <[email protected]>
> Dato: lørdag 20. marts 2010 16.17
> 
> Raul Miller wrote:
> > On Fri, Mar 19, 2010 at 1:43 PM, David Ward Lambert
> > <[email protected]> wrote:
> >>   NB. j programmers MUST understand data flow.
> >>   NB. x f^:3 y  differs from  (f&y)^:3 x
> >
> > To illustrate this we should find an operation which
> > is not associative and which produces a plausible result
> > for one of those two cases.
> 
> It is easy to show by example that these are different, for example
> 
>    x=:2
>    y=:4
>    f=:-
>    x f^:3 y
> _2
>    (f&y)^:3 x
> _10
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to