First of all, I have been careless with the suggested verbs p1, p2 insofar as
(%~)i.5
0 1 1 1 1
(*)i.5
0 1 1 1 1
the "circumlocutions" (did learn this new word along the way) give
the wrong result for argument zero (as Martin cautiously pointed out).
-- begin of sidestep
Came up with a circumlocution which *does* work alright :)
(+:^:(]`(2&o.@-~))"0) -i.5
1 0.5 0.25 0.125 0.0625
-- end of sidestep
So far, I gathered these three verbs:
(+:^:(]`1:)) -i.5
1 0.5 0.25 0.125 0.0625
(+:@]^:[&1) -i.5
1 0.5 0.25 0.125 0.0625
(+:@]^:[1:) -i.5
1 0.5 0.25 0.125 0.0625
** Raul
You're right, of course; sorry for my sloppy remark - was simply
trying to express the fact that I'm still far from being safe & sure
using ([) and (]); sometimes I get it right (mostly by experimenting).
What confuses me with this is that (]) references a verb (1:) inside
the parentheses while the other ([) refers to the argument outside.
Which brings me to
** Pascal
Hooks: I've seen, written and understood basic ones, like e.g.
"scale" (%<./) or "eval cf" (+%)/ but alas failed to recognize it in this case.
I've some idea why this dyad
5 (+:@]^:[) 1
works, but am unable to follow/ describe/ argue about/ substantiate
the steps leading to the final monad
(+:@]^:[1:) 5
-M
At 2018-08-07 22:22, you wrote:
In
(+:@]^:[ 1:) 5
The ] gives the value 1 and the [ gives the value 5.
There are other contexts where they can give the same value ([@]0
being an example of that), but I don't think we're talking about that
here?
Thanks,
--
Raul
On Tue, Aug 7, 2018 at 5:54 PM Martin Kreuzer <[email protected]> wrote:
> Martin -
> Good to see you back once in a while :)
> Have been brooding over this for quite some time.
> Using (1:) seems like cheating, but ... (as you pointed out) there
> are benefits.
> I'm still puzzled by the use of (]) and ([) at the same time for the
> same argument (I remember Raul giving me some treatment on that
> subject some years ago).
> Anyway, I think this thread has produced enough information to enable
> me to put some example on the wiki ... (will hopefully get around to
> it next weekend).
> Thanks for the discussion, and all the best.
> -M
> At 2018-08-07 18:00, you wrote:
> >Hi Martin!
> >
> > > (+:^:(]`(%~))) 5
> > >32
> > > (+:^:(]`*)) 5
> > >32
> >
> >The good old Constant function 1: will give you the benefits of
> >less circumlocution, better results for non-positive arguments,
> >and coverage of array arguments:
> >
> > +:^:(]`1:) (,: -) 0 1 2 5
> > 1 2 4 32
> > 1 0.5 0.25 0.03125
> >
> >A nice puzzle, BTW. I was only capable to get to
> >
> > (+:@]^:[ 1:) 5
> > 32
> >
> >on my own, ignorant of the gerund option.
> >
> > Martin (the other one)
> >----------------------------------------------------------------------
> >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
