> This is closer to what I am trying to do:
>
> (+1:)`*`%`(+2:)@.(0 , (1 + 2&|) , 3:)"0 [ 5 6 NB. fictional usage
> 0.857143 56
This explicit verb seems to produce what you want,
u=. 3 : '(+1:)`*`%`(+2:)@.((0,(1+2&|),3:) y) y'"0
u 5 6
0.857142857 56
Writing a tacit counterpart of u is a conundrum... But, breaking rules (at
your own risk) helps,
an=. <@:((":0) ,&< ])
agenda=. (<'@.')(0:`)(,^:)
v=. (((+1:)`*`%`(+2:) agenda (0,(1+2&|),3:)) (0 1 agenda~ ;) an )"0 f.
v 5 6
0.857142857 56
(Untested using the latest and greatest interpreters.)
On Sat, Dec 7, 2019 at 8:08 PM ethiejiesa via Programming <
programm...@jsoftware.com> wrote:
> > is the following what you want ?
> >
> > 1 2 3 (+ ; * ; -) 6 7 8
>
> Not quite. The example in my original email was probably overly minimal to
> be
> completely unhelpful.
>
> This is closer to what I am trying to do:
>
> (+1:)`*`%`(+2:)@.(0 , (1 + 2&|) , 3:)"0 [ 5 6 NB. fictional usage
> 0.857143 56
>
> The salient feature being that the choice of * or % depends on the
> argument. I
> certainly could just compute both * and %, selecting the desired result
> post
> facto, but I would like to avoid the unnecessary computation.
>
> That said, perhaps I am just going about the solution in an entirely non-J
> way.
> The problem is Advent of Code 2019, day 2, so I essentially have an array
> of
> integers where each row is an "opcode". I am attempting to decode and
> execute
> these opcodes. This is why I want to compute the arguments as well as the
> operation from the input.
>
> I certainly want to solve this on my own, but if I would love any pointers
> on
> how to *think* about the problem in a J-ish way.
>
> Jimmy Gauvin <jimmy.gau...@gmail.com> wrote:
>
> > Hi,
> >
> > is the following what you want ?
> >
> > 1 2 3 (+ ; * ; -) 6 7 8
> > ┌──────┬───────┬────────┐
> > │7 9 11│6 14 24│_5 _5 _5│
> > └──────┴───────┴────────┘
> >
> >
> >
> > On Fri, Dec 6, 2019 at 9:54 PM ethiejiesa via Programming <
> > programm...@jsoftware.com> wrote:
> >
> > > Am I just doing something silly? Or does @. really not support building
> > > trains
> > > when the right operand is a verb? Here is an overly minimal example of
> > > what I
> > > want:
> > >
> > > (1:)`+`(1:)@.(0 1 2) 0
> > > 2
> > > (1:)`+`(1:)@.(0 1 2"_) 0
> > > |rank error
> > > | (1:)`+`(1:)@.(0 1 2"_)0
> > >
> > > This is an obvious enough feature, that I feel I must be missing some
> > > obvious
> > > construction.
> > >
> > > For context, I was futzing around with this year's Advent of Code[0],
> day
> > > 2,
> > > and found myself wanting a fork-like that behaves as
> > >
> > > x (u v w) y <-> (x u y) (x v y) (x w y).
> > >
> > > After sufficient floundering, I decided to just read NuVoc for all
> > > modifiers
> > > and discovered that I had somehow not noticed the usefulness of @.
> > > However, the
> > > above restriction really surprised me.
> > >
> > > Anyway, happy holidays, J birds.
> > >
> > > [0]:https://adventofcode.com/2019/
> > > ----------------------------------------------------------------------
> > > For information about J forums see http://www.jsoftware.com/forums.htm
> > >
> > ----------------------------------------------------------------------
> > For information about J forums see http://www.jsoftware.com/forums.htm
> ----------------------------------------------------------------------
> For information about J forums see http://www.jsoftware.com/forums.htm
>
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
