"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
