As far as I can see the verb,

  (<(<,':'),<(<(,'0');1),<(,'0');1 0$'u') (1 : 'u u`:6`:6 y')

is, or resembles, a train of the form (noun adverb) but this kind of train
is not documented.  Is it?

(The verbs produced by my version of the Y combinator were

On Fri, Dec 7, 2018 at 11:12 AM Henry Rich <[email protected]> wrote:

> It turns out, though, that the bug exists only for certain adverbs on
> the right.  So there are some details involved.
>
> Henry Rich
>
> On 12/7/2018 11:09 AM, Raul Miller wrote:
> > That can't be right - in that case the parenthesis are redundant.
> >
> > It's only when non-redundant parentheses are not reproduced that we
> > have a bug in linear representation.
> >
> > That said, if I am understanding this thread, the top-level
> > parenthesis around any sub-expression to the left of an unknown adverb
> > or conjunction (or to the left of any parenthesized expression which
> > contains an unknown adverb or conjunction) when building a linear
> > representation must be considered non-redundant because you don't know
> > what grammar the expression will be used in.
> >
> > This, in turn, suggests that those parenthesis are not the
> > responsibility of the code representing the sub-expression itself
> > (because they are not redundant there), but in the code which
> > assembles that representation into the larger expression.
> >
> > I haven't looked at the implementation though - so it's possible that
> > actually implementing this concept would require a major restructuring
> > or rewrite of some sort.
> >
> > Thanks,
> >
>
>
> ---
> 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

Reply via email to