https://www.jsoftware.com/help/dictionary/d331.htm says:
u;.3 and u;._3 yield (possibly overlapping) tesselations. x u;._3 y
applies u to each complete rectangle of size |1{x beginning at integer
multiples of (each item of) the movement vector 0{x , with an infinite
size being replaced by the signed length of the corresponding axis.
In the example of
(3,: _) ];._3 ] 3 0 7 2 9 1 5 8 4 6
this means that the result should match
(3,: 10) ];._3 ] 3 0 7 2 9 1 5 8 4 6
3 0 7 2 9 1 5 8 4 6
This doesn't contradict your "all the cells have the same size"
constraint, so shouldn't require NuVoc to change.
Thanks,
--
Raul
On Fri, Nov 1, 2019 at 3:27 PM Henry Rich <[email protected]> wrote:
>
> I will change NuVoc to indicate that this applies only to ;.3, not ;._3.
>
> An important feature of ;._3, especially to the implementation, is that
> all the cells have the same size, viz the size you asked for. Fragments
> are discarded. It is possible to have 0 cells.
>
> Henry Rich
>
> On 11/1/2019 1:08 PM, Gilles Kirouac wrote:
> > The recent thread on odds and nths made me review the Cut conjunction
> > doing tessellation, i.e. case ;._3 .
> >
> > About the tile size, both the Vocabulary and the NuVoc page specify
> > that an infinite size can be used:
> >
> > Voc : an infinite size being replaced by the signed length of the
> > corresponding axis.
> >
> > NuVoc : A value in the second row of x larger in magnitude than the
> > length of the axis causes that axis to be taken in full. Values of _ and
> > __ may be used.
> >
> > In J807 and beta-n
> >
> > (3,: 4) ];._3 ] 3 0 7 2 9 1 5 8 4 6 NB. overlap possible
> > 3 0 7 2
> > 2 9 1 5
> > 5 8 4 6
> > (3,: 8) ];._3 ] 3 0 7 2 9 1 5 8 4 6
> > 3 0 7 2 9 1 5 8
> > (3,: 10) ];._3 ] 3 0 7 2 9 1 5 8 4 6
> > 3 0 7 2 9 1 5 8 4 6
> > (3,: 11) ];._3 ] 3 0 7 2 9 1 5 8 4 6 NB. oversize, empty result!
> > #(3,: 11) ];._3 ] 3 0 7 2 9 1 5 8 4 6 NB. I would expect 10
> > 0
> > #(3,: _) ];._3 ] 3 0 7 2 9 1 5 8 4 6
> > 0
> >
> > ~ Gilles
> >
> > Le 2019-11-01 à 06:14, R.E. Boss a écrit :
> >> 3 (-@[{:\ -@(|#)}.]) 3 0 7 2 9 1 5 8 4 6
> >> 7 1 4
> >>
> >> A simpler version (at least for me) of Rich's solution is
> >> 3 (,:~@[ {:;._3 ]) 3 0 7 2 9 1 5 8 4 6
> >> 7 1 4
> >>
> >>
> >> R.E. Boss
> >>
> >>
> >>> -----Oorspronkelijk bericht-----
> >>> Van: Programming <[email protected]>
> >>> Namens 'Mike Day' via Programming
> >>> Verzonden: donderdag 31 oktober 2019 20:02
> >>> Aan: [email protected]
> >>> Onderwerp: Re: [Jprogramming] odds and nths
> >>>
> >>> Unfortunately, Lorenz Koehl wants every item with an odd index, or with an
> >>> index equal to -1 modulo n.
> >>>
> >>> So if
> >>> [q =. 10?10. NB. 3rd call of 10?10, I think
> >>> 3 0 7 2 9 1 5 8 4 6
> >>> odd q NB. Koehl
> >>> 0 2 1 8 6
> >>> 3 nth q NB. Koehl
> >>> 7 1 4
> >>> 3 ({:;._3~ ,:~)~ q NB. Henry Rich
> >>> 7 1 4
> >>> 3(]#~<:@[=|)q NB. R E Boss
> >>> 2 5 8
> >>>
> >>> Henry’s method looks pretty minimal!
> >>> This doesn’t quite work in the general case, but fwiw, anyway:
> >>> _3 {: \ q. NB. !
> >>> 7 1 4 6
> >>>
> >>> Mike
> >>>
> >>> PS I see you’ve just spotted the difference!
> >>>
> >>> Sent from my iPad
> >>>
> >>>> On 31 Oct 2019, at 18:14, R.E. Boss <[email protected]> wrote:
> >>>>
> >>>> And minus another one:
> >>>>
> >>>> 3(]#~<:@[=|)i.8
> >>>> 2 5
> >>>>
> >>>>
> >>>> R.E. Boss
> >>>>
> >>>>
> >>>>> -----Oorspronkelijk bericht-----
> >>>>> Van: Programming <[email protected]>
> >>>>> Namens Raul Miller
> >>>>> Verzonden: donderdag 31 oktober 2019 18:27
> >>>>> Aan: Programming forum <[email protected]>
> >>>>> Onderwerp: Re: [Jprogramming] odds and nths
> >>>>>
> >>>>> One character shorter:
> >>>>>
> >>>>> 3(]#~0=(|#\))i.8
> >>>>> 2 5
> >>>>>
> >>>>> Thanks,
> >>>>>
> >>>>> --
> >>>>> Raul
> >>>>>
> >>>>>> On Thu, Oct 31, 2019 at 12:43 PM Henry Rich <[email protected]>
> >>> wrote:
> >>>>>> 3 ({:;._3~ ,:~)~ i. 8
> >>>>>> 2 5
> >>>>>>
> >>>>>> Henry Rich
> >>>>>>
> >>>>>>> On 10/31/2019 12:33 PM, Lorenz Köhl wrote:
> >>>>>>> NB. get items with odd index
> >>>>>>> odd=. ] #~ 0 1 $~ #
> >>>>>>>
> >>>>>>> Generalizing this I got
> >>>>>>>
> >>>>>>> NB. get the nth items
> >>>>>>> nth=. ] #~ (1 ,~ 0 $~ 1 -~ [) $~ [: # ]
> >>>>>>>
> >>>>>>> I bet this is not the shortest version!
> >>>>>>>
> >>>>>>> I tried infix but it includes fret/fill
> >>>>>>>
> >>>>>>> wrong_odd=. [: {: [: |: _2 ]\ ]
> >>>>>>>
> >>>>>>> happy halloween
> >>>>>>>
> >>>>>>> Lo
> >>>>>>> -------------------------------------------------------------------
> >>>>>>> -
> >>>>>>> -- 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
> >> ----------------------------------------------------------------------
> >> 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
