Thanks a lot for the verb and also the explanation! I found that kind of
construction very useful for my work.
Kairit Sirts
>
> Raul wrote:
> > alp=: a.{~,65 97+/i.26
> > f=: (('='={.&>) # }.&.>) @ (-.@(e.&alp) <;.1 ])
>
> Alan Stebbens responded:
> > Can any of you honestly say that the above sentence is
> > "straightforward"? :^)
>
> As odd as it sounds, yes. Once you learn the language (in
> particular the tacit style), Raul's expression really isn't
> complicated at all.
>
> I see an @ gluing two processes together. Let's call these
> the primary-processor and the post-processor. Let's look at
> the primary-processor (rightmost) first.
>
> I see a train. The right tine is ] , the identity. The
> middle tine is <;.1 , cut. The left tine is longer; but
> knowing cut and knowing the relavant J patterns that surround
> it, it's almost certain it calculates a boolean mask based on
> the input (to determine where to partition that input).
> Looking deeper, this tine starts with -.@ "the negation
> of". It ends with e.&alp .
> I do not even decompose this -- it reads as a cohesive whole:
> "members of the alphabet", rendering the entire verb "mark
> non-alphabetic characters".
>
> So the primary-processor simply cuts up the input into boxes,
> putting a partition wherever there is a non-alphabetic character.
> What about the post-processor?
>
> Again, I see a train. The middle tine is # , which (most
> often) means the left tine calculates a boolean mask to filter the
> right argument. That left tine is '='={.&> , a NVV (noun,
> verb, verb) train. It compares the lead item of every partition to
> the character '=' , resulting in a boolean mask with one
> bit per partition. The right tine removes the lead item from
> every partition (because it's served its purpose). The
> result that only partitions which start with '=' remain
> (but without that character).
>
> I hope that clarifies Raul's code. But remember that once
> you become fluent in J, you won't have to go through this exercise.
> What I've done above is basically what my teachers did for me
> in Spanish class: they broke down the grammar and translated
> the vocabulary (and common idioms), resulting in a lengthy
> explanation. Whereas for a native speaker, no explanation is
> required -- he just reads the sentence "naturally". It's no
> different for J.
>
> Well, maybe one thing is different. It took me 4 paragraphs
> to explain one line of J. That's potency. And maybe a
> motivation to become fluent.
>
> -Dan
>
> ----------------------------------------------------------------------
> For information about J forums see http://www.jsoftware.com/forums.htm
>
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
