Correction:

"An argument must always have one outer cell." -- except in the degenerate
case. Similarly, "An argument has [zero or more] inner cells".

On Wed, Jan 20, 2016 at 4:36 PM, Matthew Baulch <[email protected]>
wrote:

> Sorry. Thanks for that correction. I quite like Robert's suggestion,
> though I took it on board without adequate attribution. Wasn't trying to be
> sneaky, though I see it might appear that way.
>
> Pairing 'outer shape' and 'cell shape' does seem a bit unnatural. I think
> we need a phrase like
> (i) the argument's 'frame of cells', or
> (ii) the argument's 'inner and outer shapes'.
>
> The decision is not mine, but I can't see how adopting 'outer' on its own
> is wise.
>
> Here's an idea:
>
> An argument has inner cells, each with shape 'inner shape'. Any argument
> must always have one outer cell. It has shape 'outer shape'.
>
> Any thoughts?
> 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

Reply via email to