Coincidentally I was trying just what you suggested using Roger Stoke's
fine EVM on 'test=: +: ifC ' and '(test=: +: ifC)' when your message came
in.
For example I executed the following (the adjustment is to narrow the wide
output).
'( test=: +: ifC )' EVM
(20&{.,100&}.)"1 hist''
' test=: +: ifC' EVM
(20&{.,100&}.)"1 hist''
The main difference in the results was that extra finishing steps were
required to process the outside parens. But I still do not understand why
the parenthesized version produces output.
On Mon, Sep 16, 2013 at 12:10 PM, Raul Miller <[email protected]> wrote:
> It's easier to answer questions like this if you quote the specific
> code you are referring to. (I could try answering, but what if I
> guessed wrong (about what confused you)? I'm something of a
> perfectionist and sometimes that means that the slightest issue
> distracts me onto something else.) And, believe me, I've plenty of
> other things to distract me! :)
>
> That said, if you want to try answering these questions for yourself
> (instead of encouraging the j community to perhaps talk about them),
> you could try running the expressions through trace, both with and
> without the parenthesis.
>
> Thanks,
>
> --
> Raul
>
>
>
>
>
> On Mon, Sep 16, 2013 at 11:27 AM, Brian Schott <[email protected]>
> wrote:
> > Twice now in this thread, parentheses have made a serious difference in
> > results. If that difference has been explained I have missed it and would
> > appreciate an explanation.
> >
> > In Pascal's message parens were need around the explicit definition of an
> > adverb.
> >
> > In Jose' s (Pepe's?) message the parens were used as follows and
> different
> > results are produced without parens.
> >
> > st=. 7!:2@:] ; 6!:2
> >
> > NB. Explicitly...
> >
> > controlA_z_ =: 1 : 0
> > flag1_z_ =: 1
> > o=. u y
> > flag1_z_ =: 0
> > o
> > )
> >
> > ifC=: 1 : 'if. flag1_z_ do. u y else. y end.'
> >
> > ( test=: +: ifC )
> >
> >
> >
> > --
>
--
(B=)
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
