Thank you Pepe,

Another option that would seem more useful/general

(a v v2 v3) captures u into the full argument u`v`v2`v3  a

(a n) would if n not gerund, u`(n ar) a, and if n is a gerund u`n a


(a n v2 v3 v4) would ar n if not gerund, and then produce u`n`v2`v3`v4 a


This would allow for easy strand notation, with any termination character. a 
(norv) can return (a gerund) to continue parsing

Any train combo/rearrangement can be done with f g h ==> f ((@.0 1 2) g h) 

(a1 v) a2 is usefully separated into a1 as the parsing guard/validator with 
initial parameters, and a2 as the final processing function.

This allows for any of the original and option 2 alternatives with much more 
flexibility.

I was asking for custom parser utilities, and not for J to change its parsing.

On Wednesday, May 6, 2020, 05:02:09 p.m. EDT, Jose Mario Quintana 
<[email protected]> wrote: 

A0 V1    adv  (x A0) V1

http://www.jsoftware.com/pipermail/programming/2016-March/044744.html


On Wed, May 6, 2020 at 4:42 PM 'Pascal Jasmin' via Programming <
[email protected]> wrote:

> I cannot find what (a v) did back then.
>
>
>
> Early J had a complete language of bidents/tridents.  It was beautiful
> and powerful.  Not five people understood it.  It was removed in J5,
> never to return.
>
> Henry Rich
>
> On 5/6/2020 3:00 PM, 'Pascal Jasmin' via Programming wrote:
> > what should (a v) do? (if it were valid)
> >
> > then what would (a v v) or (a v v v) do?
> >
> > option 1:  change v into an adverb (letting it return functions or
> return an lr that will be evaled to any form of speech), and then run v 'a'
> >
> > option 2: Compose v on (u a).  (a v) remains an adverb.
> >
> > btw, there is a missing composition operator in J.  I'd suggest O. (on)
> as a conjunction where O. allows for both dyadic u and v.
> >
> > O. =: (u@:v) : ([ u v)
> >
> > so option 2 above would be v O. (u a) .  This doesn't seem useful
> because you could just write the adverb a ( v O.) for the same result.  A
> generalization of option 2 is there is some implied conjunction partially
> bound to v that gets executed in (a v)
> >
> > Where option 2 gets interesting is in the expression
> >
> > v1 (a1 v2) (a2 v3)
> >
> > There is no reason to write such an expression if the intent were for v1
> to be an argument to the adverb (a1 v2) if there is an implied conjunction
> meaning.  Instead the above example would be a multi adverb where
> parameters are "templated in"
> >
> > u1 u2 (v1 (a1 v2) (a2 v3))  ==> v1  v2 O.(u1 a1) v3 O.(u2 a2 )
> >
> >
> > Any other options/proposals?
>
> > ----------------------------------------------------------------------
> > For information about J forums see http://www.jsoftware.com/forums.htm
>
>
> --
> This email has been checked for viruses by AVG.
> https://www.avg.com
>
> ----------------------------------------------------------------------
> 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

Reply via email to