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
