> Good point. And a computer (human, electronic or otherwise) could
> never conclude the sum is 2 by actually adding the terms. So even if
> i. _ was valid, a sound theoretical answer would still require
> something capable of symbolic manipulation to go any further
> than just communicating the idea.

Symbolic manipulation is what f^:_1 and f&.g already do.
A system that knows that +/ r^i._  is (1-r^_)%1-r 
isn't too difficult.  (Dan Bron's original challenge question 
was +/ 2 ^ - i. _ ) 



----- Original Message -----
From: Thomas Costigliola <[email protected]>
Date: Wednesday, December 9, 2009 9:16
Subject: Re: [Jchat] Number Machines (moved from jgeneral)
To: Chat forum <[email protected]>

> On Wed, Dec 9, 2009 at 10:49 AM, Viktor Cerovski
> <[email protected]> wrote:
> >
> >
> > Thomas Costigliola-2 wrote:
> >>
> >> [...]
> >>
> >>    (-:@{. ([,+) {:)^:_ ] 1 1
> >> 0 2
> >>
> >> A little better but I still want the left arg to power to be 
> shorter>> or at least more clear.
> >> [...]
> >>
> >
> >   mp=.+/ .*
> >   (0.5 0,.1)&mp^:_ ] 1 1
> > 2 1
> >
> > Such program terminates only because of the round-off error, thus
> >
> >   (1r2 0,.1)&mp^:_ ] 1 1
> >
> > should eventually run out of memory.
> 
> Good point. And a computer (human, electronic or otherwise) could
> never conclude the sum is 2 by actually adding the terms. So 
> even if
> i. _ was valid, a sound theoretical answer would still require
> something capable of symbolic manipulation to go any further 
> than just
> communicating the idea.
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to