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

Reply via email to