[Jprogramming] gerund creation failure

[Raul Miller]

On 5/19/07, Raul Miller <rauldmiller at gmail.com> wrote:
> I lied.
>
> This will not work.  to make this work I have to get rid of
> a set of parenthesis:
>
>   rec_case=:(fib]-1:) +(fib]-2:)
>
> The problem is that with (+ fib]-2:) I have only two verbs at
> the top level, so (fib]-1:) is used as a dyad rather than as
> a monad.

Well... ok, maybe I lied about me lying...  if the definition of
fib simply ignores its left argument (which it probably would),
then the above "issue" is not really an issue in this case.
(But could cause problems in other cases.)

Also...

On 5/19/07, Raul Miller <rauldmiller at gmail.com> wrote:
> On 5/19/07, Terrence Brannon <metaperl.j at gmail.com> wrote:
> > On 5/19/07, Raul Miller <rauldmiller at gmail.com> wrote:
> > > Why?  Well, the parenthesized expression
> > >    (] - 2)
> > > evaluates as:
> > > _2
> >
> > How did you figure out what it evaluates as?
>
> I type the expression into the J console.

I can also figure out how this expression evaluates, by hand, using
the rules at http://www.jsoftware.com/help/dictionary/dicte.htm

   queue:            rule:     stack:
   mark ( ] - 2 )    shift
   mark ( ] - 2      shift     )
   mark ( ] -        shift     2 )
   mark ( ]          shift     - 2 )
   mark (            1 Monad   ] - 2 )
   mark (            shift     ] _2 )
   mark              1 Monad   ( ] _2 )
   mark              8 Paren   ( _2 )
   mark              shift     _2
                     DONE      mark _2

Note: while I don't know where this is specified, the result of
evaluation is successful if the final stack matches
   mark CAVN
and is a syntax error if it's anything else.

FYI,

-- 
Raul