"Cell" is unavoidable and useful when talking about the rank operator, and
what is the shape of a cell?  Cell shape.


On Tue, Jan 19, 2016 at 3:39 PM, robert therriault <[email protected]>
wrote:

> Thanks Roger,
>
> When you get a minute, could you give some examples of why cell shape is
> better? My preference for the inner/outer shape description is that the
> inner shape is the component that describes the shape of the cell and the
> outer shape is the component that describes the shape of the frame.
> Together they make up the entire shape and the division between the two is
> the rank of the verb that is applied. There is a connection implied between
> the terms inner and outer that is not present between frame and cell (or
> outer and cell).
>
> You have a much deeper understanding of the language and have seen
> complications that I have not. My ignorance of how cell shape is a better
> term is an opportunity to learn.
>
> Cheers, bob
>
> > On Jan 19, 2016, at 11:59 AM, Roger Hui <[email protected]>
> wrote:
> >
> > The alternative to frame / cell shape being discussed is outer shape /
> cell
> > shape, not outer shape / inner shape.  I don't believe anyone has used
> > "inner shape" before (and it's not as good as "cell shape").
> >
> > On Tue, Jan 19, 2016 at 2:14 AM, Matthew Baulch <[email protected]>
> > wrote:
> >
> >> Outer/inner makes perfect sense. Seems unlikely to lead anyone astray.
> >> To play devil's advocate, it might seem silly but maybe a newbie could
> >> guess that inner/outer shape relates to boxing. Is this paranoia? I
> don't
> >> know.
> >>
> >> The important question is: who is the terminology intended to serve? The
> >> answer is, of course, everyone. But in what proportions? There are
> inherent
> >> trade-offs. As noted by Roger and others,
> >>
> >> Outer and inner shape
> >> Pros: fit together like a pair of gloves, suggest a dependency of some
> sort
> >> (hopefully on rank!), is being adopted by Dyalog in a similar form
> (maybe a
> >> pro?).
> >> Cons: has an unfortunate though slight suggestion of boxing.
> >>
> >> Frame:
> >> Pros: cells-in-frame concept makes some intuitive (and pictorial) sense,
> >> frame is (?) unused for terminology anywhere else in J so unlikely to be
> >> confused.
> >> Cons: has historical context around 'empty', tempts us to use the word
> >> empty, cells-in-frame concept breaks down when frame is empty (even
> though
> >> cells may still exist), no natural pairing with cells.
> >>
> >> Cells
> >> Pros: sort of makes sense...?
> >> Cons: has various meanings depending on the context, doesn't imply that
> >> shape or rank are at all important, no natural pairing with frame.
> >>
> >> I'm sure I've missed something. Anyway, I think there's a strong case
> for
> >> inner/outer shapes.
> >>
> >> On Tue, Jan 19, 2016 at 5:43 AM, Henry Rich <[email protected]>
> wrote:
> >>
> >>> I really like this suggestion.  "frame" makes sense for result: the
> frame
> >>> is held fixed while the cell-results are coerced into the same shape,
> and
> >>> then assembled using the frame.  For the arguments, "outer shape" shows
> >> the
> >>> dependence on the argument shape and (implicitly) the verb rank.
> >>>
> >>> I wonder whether we should try to move the documentation in this
> >>> direction.  There would need to be a general consensus in favor.
> >>>
> >>> Henry Rich
> >>>
> >>>
> >>> On 1/18/2016 11:52 AM, Roger Hui wrote:
> >>>
> >>>> The terminology originated in SHARP APL in the 1980s.  "Frame" was at
> >>>> times
> >>>> called "outer shape".  In some situations, "outer shape" may be a
> >> better,
> >>>> more easily understood term.  You know, cell shape and outer shape;
> >> outer
> >>>> shape is part of the shape; etc.
> >>>>
> >>>>
> >>>>
> >>>>
> >>>> On Mon, Jan 18, 2016 at 7:19 AM, Jose Mario Quintana <
> >>>> [email protected]> wrote:
> >>>>
> >>>> I would not be the one arguing for empty frame vs zero frame
> terminology
> >>>>> :)
> >>>>>  (thanks for providing the context).
> >>>>>
> >>>>> Regarding frame, I meant it in the sense that Ken Chakahwata did: "to
> >>>>> have
> >>>>> a J definition of that fictitious primitive."
> >>>>>
> >>>>> Your executable model can, of course, readily address Ken's question
> >> and
> >>>>> other similar questions for specific instances (pointing out, albeit
> >>>>> rather
> >>>>> tacitly, that such J definition already existed, was my main reason
> for
> >>>>> mentioning your article):
> >>>>>
> >>>>>    rk    =. #@$
> >>>>>    er    =. (0:>.(+rk))`(<.rk) @. (0:<:[)
> >>>>>    fr    =. -@er }. $@]
> >>>>>    cs    =. -@er {. $@]
> >>>>>
> >>>>>    (Y=. i.2 3 4)
> >>>>>  0  1  2  3
> >>>>>  4  5  6  7
> >>>>>  8  9 10 11
> >>>>>
> >>>>> 12 13 14 15
> >>>>> 16 17 18 19
> >>>>> 20 21 22 23
> >>>>>
> >>>>>    3 (er;fr;cs) Y    NB. effective rank; frame; cell shape
> >>>>> ┌─┬┬─────┐
> >>>>> │3││2 3 4│
> >>>>> └─┴┴─────┘
> >>>>>
> >>>>>    2 (er;fr;cs) Y    NB. effective rank; frame; cell shape
> >>>>> ┌─┬─┬───┐
> >>>>> │2│2│3 4│
> >>>>> └─┴─┴───┘
> >>>>>   _1 (er;fr;cs) Y    NB. effective rank; frame; cell shape
> >>>>> ┌─┬─┬───┐
> >>>>> │2│2│3 4│
> >>>>> └─┴─┴───┘
> >>>>>
> >>>>>
> >>>>>
> >>>>> On Sun, Jan 17, 2016 at 11:33 PM, Roger Hui <
> [email protected]
> >>>
> >>>>> wrote:
> >>>>>
> >>>>> I did not define them; Roland Pesch did: Empty Frames in SHARP APL
> >>>>>> <http://www.jsoftware.com/papers/EmptyFrames.htm>, 1986.  I did
> >> rename
> >>>>>> them
> >>>>>> to "zero frames".  Read the 1986 paper and you can decide for
> yourself
> >>>>>> whether "empty frame" or "zero frame" is the better name.
> >>>>>>
> >>>>>>
> >>>>>>
> >>>>>> On Sun, Jan 17, 2016 at 5:28 PM, Jose Mario Quintana <
> >>>>>> [email protected]> wrote:
> >>>>>>
> >>>>>> The verb (frame) as well as the Zero Frame concept are defined in
> [0]
> >>>>>>>
> >>>>>> by
> >>>>>
> >>>>>> Roger.
> >>>>>>>
> >>>>>>> [0] Rank and Uniformity
> >>>>>>>     http://www.jsoftware.com/papers/rank.htm
> >>>>>>>
> >>>>>>> On Sun, Jan 17, 2016 at 7:11 PM, Ken Chakahwata <
> >>>>>>> [email protected]> wrote:
> >>>>>>>
> >>>>>>> My guess is that it would help if we could imagine that we had a
> >>>>>>>>
> >>>>>>> primitive
> >>>>>>>
> >>>>>>>> called 'frame' in the same way as we have one called 'shape' i.e.
> $
> >>>>>>>> Then one way to get to the precise meaning of frame is to have a J
> >>>>>>>> definition of that ficticious primitive. At a guess, this
> primitive
> >>>>>>>> requires the 'rank' of the cells in order to then return the
> >>>>>>>>
> >>>>>>> appropriate
> >>>>>>
> >>>>>>> frame.
> >>>>>>>> If we have an array of shape (x,y,z), and we stipulate cells of
> rank
> >>>>>>>>
> >>>>>>> 3,
> >>>>>
> >>>>>> then the frame is presumably empty? Not sure of this... but anyhow,
> >>>>>>>>
> >>>>>>> just
> >>>>>>
> >>>>>>> a
> >>>>>>>
> >>>>>>>> thought...
> >>>>>>>>
> >>>>>>>> Enjoy
> >>>>>>>> ken
> >>>>>>>>
> >>>>>>>> -----Original Message-----
> >>>>>>>> From: Programming [mailto:
> [email protected]]
> >>>>>>>>
> >>>>>>> On
> >>>>>
> >>>>>> Behalf Of Henry Rich
> >>>>>>>> Sent: 17 January 2016 23:59
> >>>>>>>> To: [email protected]
> >>>>>>>> Subject: Re: [Jprogramming] Definition: Frame of an argument
> >>>>>>>>
> >>>>>>>> The terminology I use is an (x by y by z) array of cells, or an
> >> array
> >>>>>>>>
> >>>>>>> of
> >>>>>>
> >>>>>>> cells with frame (x,y,z), emphasizing that the frame is a (part of
> >>>>>>>>
> >>>>>>> the)
> >>>>>
> >>>>>> shape rather than an array.
> >>>>>>>>
> >>>>>>>> Henry Rich
> >>>>>>>>
> >>>>>>>> On 1/17/2016 6:16 PM, Raul Miller wrote:
> >>>>>>>>
> >>>>>>>>> Hmm... ok, reviewing
> >>>>>>>>> http://www.jsoftware.com/help/primer/frame_and_cell.htm 'frame'
> >>>>>>>>>
> >>>>>>>> does
> >>>>>
> >>>>>> get used that way.
> >>>>>>>>>
> >>>>>>>>> I was thinking of the frame as having a shape rather than being
> the
> >>>>>>>>>
> >>>>>>>> shape.
> >>>>>>>>
> >>>>>>>>> Then again, since you can think of an array as being (for
> example)
> >>>>>>>>>
> >>>>>>>> an
> >>>>>
> >>>>>> (x,y,z) frame of cells, I do not think that my interpretation was
> >>>>>>>>> entirely incorrect, either. So I suppose I have gotten myself
> into
> >>>>>>>>>
> >>>>>>>> a
> >>>>>
> >>>>>> "much ado about nothing" sort of issue.
> >>>>>>>>>
> >>>>>>>>> Thanks,
> >>>>>>>>>
> >>>>>>>>>
> >>>>>>>>
> >> ----------------------------------------------------------------------
> >>>>>
> >>>>>> 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
> >>
> > ----------------------------------------------------------------------
> > 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