You should probably spend a few minutes studying the diagram at
https://code.jsoftware.com/wiki/Vocabulary/fcap#How_should_I_define_u_and_v.3F
The short form, however, might be summarized as:
lfold=: {{ ]F..(u~) }}
rfold=: {{ ]F.:(u~) }}
I hope this makes sense,
--
Raul
On Tue, Dec 6, 2022 at 4:06 PM Brian Schott <schott.br...@gmail.com> wrote:
>
> Raul,
>
> That is extremely helpful. The main thing I get out of both of your
> solutions, but most clearly out of the insert solution is that the argument
> to be altered should be on the right side. For some reason, I have wanted
> to reverse that. But your insert solution demonstrates that strongly.
>
> When I was trying this I had tried your Fold solution, but using *move*,
> not *move~* , and was getting a domain error, which I could not understand
> to correct. I still don't see why move~ is correct in that context because
> move's x is state and move's y is moves; the ~ reverses those, I think.
>
> Btw, I was struggling to get the email version to work when I checked it
> before sending it, too. That's why I sent the attachment, believing that
> the errors would be avoided.
>
> On Tue, Dec 6, 2022 at 3:29 PM Raul Miller <rauldmil...@gmail.com> wrote:
>
> > I've not fully digested all of the fold rules yet, so I'll first focus
> > on building an insert version.
> >
> > The key to using insert is building representing the expression which
> > would result from insert.
> >
> > In other words, instead of
> > state tomove moves
> > +-+-+----+
> > |C|M|PDNZ|
> > +-+-+----+
> >
> > We could look at moves:
> >
> > moves
> > 1 2 1
> > 3 1 3
> > 2 2 1
> > 1 1 2
> >
> > and construct a corresponding expression.
> >
> > 1 1 2 move~ 2 2 1 move~ 3 1 3 move~ 1 2 1 move~ state
> > +-+-+----+
> > |C|M|PDNZ|
> > +-+-+----+
> >
> > There's a type mismatch between the left and right arguments to move,
> > so instead of move~/ we would need to box the arguments maintain that
> > in our intermediate results.
> >
> > That gives us |.state;<"1 moves as a starting point, move~&.>/ to
> > operate on them and ; to clean up at the end.
> >
> > ;move~&.>/|.state;<"1 moves
> > +-+-+----+
> > |C|M|PDNZ|
> > +-+-+----+
> >
> > Fold would let us dispense with this boxing mechanism, and ]F..v is
> > roughly equivalent to v/ and, testing, this works:
> >
> > state ]F..(move~) moves
> > +-+-+----+
> > |C|M|PDNZ|
> > +-+-+----+
> >
> > And, I guess a rule here is that X ]F..V Y can be tested beforehand by
> > using V between items of Y with X as the rightmost such item.
> >
> > I hope this makes sense,
> >
> > --
> > Raul
> >
> >
> >
> ----------------------------------------------------------------------
> For information about J forums see http://www.jsoftware.com/forums.htm
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
