Do you understand the following?
sum =: +/
sum1=: +/1
sum1 is defined to apply to the rank 1 cells of
the argument. Now:
foo=: 3 : 0
blah blah blah
)
foo1=: 3 : 0 1
blah blah blah
)
The pattern regarding rank is the same between
sum/sum1 and foo/foo1 . What is different is
that for (m :
Instead of a PowerPoint presentation, I suggest
running a lab using a screen projector. That way,
questions from the audience can be answered using
live examples executed on the system.
The following labs look suitable:
A J Introduction (by Ken Iverson myself)
A Taste of J (1)
A Taste of J
a. The monad { computes a generalized Cartesian
product. For example:
{0 1 2;3 4 5 6
+---+---+---+---+
|0 3|0 4|0 5|0 6|
+---+---+---+---+
|1 3|1 4|1 5|1 6|
+---+---+---+---+
|2 3|2 4|2 5|2 6|
+---+---+---+---+
b. When using the dyad u/ to compute a function
table, it can be helpful to
the rank of the verb determins the cells.
This is not correct. For example, it make sense
to talk about the 1-cells of a matrix (the rows
of a matrix) without reference to any verb.
- Original Message -
From: Dan Bron [EMAIL PROTECTED]
Date: Friday, September 22, 2006 10:10 am
is the left rank of u .
- Original Message -
From: June Kim [EMAIL PROTECTED]
Date: Friday, September 22, 2006 10:42 am
Subject: Re: Cell (was Re: [Jgeneral] I expect a table of ordered pairs for
this)
2006/9/23, Roger Hui [EMAIL PROTECTED]:
the rank of the verb determins the cells
See also the paper Rank and Uniformity
http://www.jsoftware.com/papers/rank.htm
--
For information about J forums see http://www.jsoftware.com/forums.htm
PROTECTED]
Date: Friday, September 22, 2006 1:27 pm
Subject: RE: Cell (was Re: [Jgeneral] I expect a table of ordered pairs forthis)
On Behalf Of Roger Hui
the rank of the verb determins the cells.
This is not correct. For example, it make sense
to talk about the 1-cells of a matrix
There is a rough but not direct relationship
between a primitive and a test script. For example,
there is a test script for each of + and +. and +: ,
with the script containing tests for the monad and
the dyad. A test script avails itself of the full
power of J so that a test which is quite
J grammar is specified by section II of the J
Dictionary, in particular section IIe, Parsing and
Execution.
http://www.jsoftware.com/help/dictionary/dicte.htm
A sentence is evaluated in two distinct steps:
- word formation (tokenization) as in the monad ;:
- parsing as in section IIe
-
I have only a sketchy knowledge of MATLAB (and
I have written 0 MATLAB programs). If I were
comparing J and MATLAB, the following are areas
I would investigate:
- consistent treatment of arrays, especially
arrays with rank 2
- adverbs and conjunctions (higher-order functions)
- non-numeric data
Ken preferred descriptions in English over
descriptions in J. But that is just my guess for
the reason for the change; I don't remember the exact reason nor remember
whether I knew the
exact reason.
- Original Message -
From: bill lam [EMAIL PROTECTED]
Date: Monday, October 2, 2006
Thanks for the feedback re the J dictionary.
I have removed mention of conjunctions wrt trains.
I have removed the sentence in section II G
that contains the word operator. I will think
of a place to put the description that operator
denotes adverb or conjunction.
I have removed mention
boolean / each boxed_list in APL should work
as that is parsed as (boolean/) each boxed_list .
It is more telling if you have to select
differently in each box. Then you can not do
(x0;x1;x2)/each y0;y1;y2 because it is parsed
as ((x0;x1;x2)/) each y0;y1;y2 . With # this is no
problem:
Alternative solutions:
X=: (_2 _1) |: O {2 F
X -: O 4 : 'x}y'1 22 F
1
X -: (O ,.1 i.{:$O) {2 F
1
- Original Message -
From: Roger Hui [EMAIL PROTECTED]
Date: Friday, November 10, 2006 3:14 am
Subject: Re: [Jgeneral] selection expression in J
O=: 3 5 [EMAIL PROTECTED] 7
These are not bugs.
- Original Message -
From: Dan Bron [EMAIL PROTECTED]
Date: Friday, December 1, 2006 7:30 am
Subject: [Jgeneral] Bug? Sin of pi.
Shouldn't the sin of pi be zero?
1 o. 1p1
1.22461e_16
This isn't a tolerance issue, because 1.22461e_16
Because
1p1 - 3.141592653589793
0
- Original Message -
From: Dan Bron [EMAIL PROTECTED]
Date: Friday, December 1, 2006 7:58 am
Subject: Re: [Jgeneral] Bug? Sin of pi.
DB 0 = 1 o. 1p1
DB 0
RH These are not bugs.
Why not?
Changes in the C library? Vagaries in the CPU?
Phase of the moon?
- Original Message -
From: Joey K Tuttle [EMAIL PROTECTED]
Date: Friday, December 1, 2006 10:55 am
Subject: Re: [Jgeneral] Bug? Sin of pi.
At 10:32 -0800 2006/12/01, Roger Hui wrote:
Because
1p1
epsilon to know when to stop.
- Original Message -
From: John Randall [EMAIL PROTECTED]
Date: Monday, December 4, 2006 4:02 am
Subject: Re: [Jgeneral] Bug? Sin of pi.
Roger Hui wrote:
Do you have a good estimate for how many terms
are needed in the exponential series
If you only knew the stories I can tell you about
the C compilers and C library functions ...
- Original Message -
From: Joey K Tuttle [EMAIL PROTECTED]
Date: Tuesday, December 5, 2006 11:26 am
Subject: Re: [Jgeneral] Bug? Sin of pi.
...
Still, I believe some numerical earthquake took
I was told by a long-time user whose opinions I
respect that it would be more useful to be able
compute to some specified precision than to have symbolic manipulations. So,
some day, Joey will be
able to say in J:
y=: 40 x: 2.5
y
2.5d40
^ y
12.18249396070347343807017595116796618318
60 decimal digits of accuracy come in handy when
doing computations based on the angular difference
between different sides of a proton on the opposite
end of the universe. :-)
- Original Message -
From: Ralph G Selfridge [EMAIL PROTECTED]
Date: Thursday, December 7, 2006 7:23 am
Personally, I think that adding multiple-precision calculations to J
is harmless and possibly useful idea. Roger's description suggested
that to the user it would appear as just another data type,
although I
imagine it would be a lot of work at his end. Moving in the direction
of symbolic
Is there a typo in your description?
g(x,y) is not defined. Do you mean g(n+1)=f(n,n-1)?
- Original Message -
From: John Randall [EMAIL PROTECTED]
Date: Friday, December 8, 2006 9:56 am
Subject: Re: [Jgeneral] exp(y). sin(y) and accuracy.
While we are talking about numerical
.
Roger Hui wrote:
Is there a typo in your description?
g(x,y) is not defined. Do you mean g(n+1)=f(n,n-1)?
Sorry, yes.
Let f(x,y)=108-(815-1500%y)%x and let g(n+1)=f(n,n-1), with g(0)=4,
g(1)=4.25. Find lim g(n) as n goes to infinity
f=: 4 : '108-(815-1500%y)%x'
g=: 3 : 'if. y e. 0 1 do. y{4 4.25 else. (g y-1) f g y-2 end.'
h=: 3 : 'if. y e. 0 1 do. y{4 17r4 else. (h y-1) f h y-2 end.'
g uses floating point and the analoguous h uses
rational numbers. Thus:
0j18 : (g ,. h)0 i.20
4.00
See also the essay in the J Wiki
http://www.jsoftware.com/jwiki/Essays/Logarithm_and_Exponential_(Extended_Precision)
- Original Message -
From: John Randall [EMAIL PROTECTED]
Date: Tuesday, December 12, 2006 8:49 am
Subject: Re: [Jgeneral] exp(y). sin(y) and accuracy.
Anyone
The following URL is less likely to be mangled by
e-mail readers:
http://www.jsoftware.com/jwiki/Essays/Logarithm_and_Exponential_%28Extended_Precision%29
- Original Message -
From: Roger Hui [EMAIL PROTECTED]
Date: Tuesday, December 12, 2006 10:20 am
Subject: Re: [Jgeneral] exp(y
That is not what the discussion is about.
The e-mail forwarded by Devon McCormick says that
if sin squared plus cos squared is not quite equal
to 1, then that is evidence that the floating
point unit ... was malfunctioning. The example
I gave demonstrated that you can have perfect
values
Please see http://www.jsoftware.com/help/release/iamend.htm
- Original Message -
From: Jacobs, Jan [EMAIL PROTECTED]
Date: Tuesday, December 19, 2006 6:23 am
Subject: [Jgeneral] selection
Ls,
How do I define a tacit function for the expression
(stateNew*-.accept)+state*accept, where
It would not just be ./c as the indices can be negative. And I think you
overestimate the time
cost for allocating and populating new storage.
- Original Message -
From: Miller, Raul D [EMAIL PROTECTED]
Date: Tuesday, December 19, 2006 9:05 am
Subject: RE: [Jgeneral] selection
Roger
One problem with your expressions is that integers
in J64 are 8 bytes per number.
- Original Message -
From: Anssi Seppälä [EMAIL PROTECTED]
Date: Friday, December 29, 2006 4:28 pm
Subject: [Jgeneral] J64 mapped files
Testing allocating large mapped files with J64:
createjmf_jmf_
From http://www.jsoftware.com/help/dictionary/d310n.htm
3. ... A name assigned by =: is global, except
that global assignment to a local name is not
permitted. ...
In testLocalGlobal, var is local, and the assignment
var=: localGlobal ''
would have been global and is therefore not
Proof sketch:
0. Working with G is equivalent to working with
a matrix of indices into the ravelled values.
For example:
G=: 5 5 [EMAIL PROTECTED] 2e9
U=: , G
H=: i.5 5
G -: H { U
1
H
0 1 2 3 4
5 6 7 8 9
10 11 12 13 14
15 16 17 18 19
20 21 22 23 24
1. Since G is
Yes, also:
cd=: /..(|.1_)
- Original Message -
From: Miller, Raul D [EMAIL PROTECTED]
Date: Thursday, January 4, 2007 12:14 pm
Subject: RE: [Jgeneral] Array-oriented proof?
I'm still digesting this (using indices to make induction work seems
like a great idea).
That said, if I
cd=: (|. +/ ])@[EMAIL PROTECTED] (/., /: {./.,[EMAIL PROTECTED]) ]
cd=: /..(|.1_)
I should have called them diagonals instead of
counter diagonals. It's /.y that gets the
counter diagonals (aka oblique lines).
- Original Message -
From: Roger Hui [EMAIL PROTECTED]
Date: Thursday
symmetric:
(-: |:) K
0
(-: rot) K
0
- Original Message -
From: Roger Hui [EMAIL PROTECTED]
Date: Thursday, January 4, 2007 11:42 am
Subject: Re: [Jgeneral] Array-oriented proof?
Proof sketch:
0. Working with G is equivalent to working with
a matrix of indices into the ravelled
Zach Reiter, whom you met Tuesday evening at the
NYCJUG, implemented a key component that made the
improved grid facility possible.
- Original Message -
From: Dan Bron [EMAIL PROTECTED]
Date: Thursday, January 11, 2007 12:26 pm
Subject: [Jgeneral] Thanks for Grid
To whomever is
Regarding the limit on %. in J: Perhaps you should
get a bigger machine? I tried the following just
now on this machine, a run-of-the-mill Celeron
(2.8 GHz, 512 MB):
x=: 100 100 [EMAIL PROTECTED] 0 NB. random real matrix
6!:2 'y=: %. x' NB. invert it
0.0141401
There is a minor time and space saving from using
x [EMAIL PROTECTED] y compared to ?x$y :
ts '? 1000 1000 $ 100'
0.0448292 8.38957e6
ts '1000 1000 [EMAIL PROTECTED] 100'
0.0292618 4.19526e6
More so for y which is a small power of 2, esp.
for y=2 :
ts '? 1000 1000 $ 2'
0.0143119
In this particular cases rational numbers provide
an escape:
hilbert 10x
1 1r2 1r3 1r4 1r5 1r6 1r7 1r8 1r9 1r10
1r2 1r3 1r4 1r5 1r6 1r7 1r8 1r9 1r10 1r11
1r3 1r4 1r5 1r6 1r7 1r8 1r9 1r10 1r11 1r12
1r4 1r5 1r6 1r7 1r8 1r9 1r10 1r11 1r12 1r13
1r5 1r6 1r7 1r8
Inverting a 50 by 50 random matrix, aka the JKT
benchmark, has long been used in J. An example
from my notebooks:
no. 5968433 filed 17.35.28 thu 13 feb 1992
from eem
to kei rhui
cc dhs jkt
The timing David mentioned was for Joey's benchmark
on the NeXT computer. The timing I reported
It has not yet been done because the style of
working in J is different from the style in APL
in the old old days. Also because of the
dustbin phenomenon previously mentioned.
Anyway the following is how you can do it.
To save everything, done in 3 lines:
nl_z_=: 4!:1
allnames=: ; @((3 : 'nl__y
When the workspace is 250 MB almost certainly most
of it is data (nouns). One variant could be to save
just the verbs adverbs and conjunctions. You can
get this by using nl__y 1 2 3 instead of
nl__y i.4 .
- Original Message -
From: Devon McCormick [EMAIL PROTECTED]
Date: Sunday,
See also:
http://www.jsoftware.com/jwiki/EditingGuidelines/Character_Palette_--_Box_Drawing
http://www.jsoftware.com/jwiki/EditingGuidelines/Character_Palette_--_Box_Drawing/Issues
- Original Message -
From: John Randall [EMAIL PROTECTED]
Date: Wednesday, January 31, 2007 12:04 pm
Yes, for arrays with the wrong size the
virtual memory used can be up to 2 times the
theoretical minimum. There are no plans to
change this behaviour.
- Original Message -
From: bill lam [EMAIL PROTECTED]
Date: Monday, February 5, 2007 7:40 am
Subject: [Jgeneral] large space overhead
[It] appears that this ARGB array is not
wrong size,
There is more than one way for an array to be
the wrong size. See:
http://www.jsoftware.com/jwiki/Essays/Boolean_Array_Size
- Original Message -
From: bill lam [EMAIL PROTECTED]
Date: Monday, February 5, 2007 11:29 pm
Subject:
Using ∑i.5 to display +/i.5 obscures the power
of adverbs. +/x and */x tell you that the two operations are related. Not so
Sigma x and
Pi x. And what do you do for *./x (for example)?
Big LCM?
- Original Message -
From: Don Guinn [EMAIL PROTECTED]
Date: Tuesday, February 6, 2007
The q: algorithm in J is not the best known.
(I am not a factoring expert.) It uses
elliptic curves together with Miller-Rabin to
test for primes. You may want to try
http://www.alpertron.com.ar/ecm.htm
- Original Message -
From: Devon McCormick [EMAIL PROTECTED]
Date: Wednesday,
instantaneously.
- Original Message -
From: Roger Hui [EMAIL PROTECTED]
Date: Wednesday, February 7, 2007 9:28 am
Subject: Re: [Jgeneral] speeding up J
The q: algorithm in J is not the best known.
(I am not a factoring expert.) It uses
elliptic curves together with Miller-Rabin to
test
How is the APL {lamp} symbol an improvement over
NB. as far as being more understandable/readable?
- Original Message -
From: Markus Triska [EMAIL PROTECTED]
Date: Wednesday, February 7, 2007 11:17 am
Subject: Re: [Jgeneral] Displaying J with Unicode Special Symbols
Joey K Tuttle
}.
- Original Message -
From: Markus Triska [EMAIL PROTECTED]
Date: Wednesday, February 7, 2007 1:59 pm
Subject: Re: [Jgeneral] Displaying J with Unicode Special Symbols
Roger Hui [EMAIL PROTECTED] writes:
How is the APL {lamp} symbol an improvement over
NB. as far as being more understandable
APL has the advantage of expressing many terms
in one symbol, and it uses familiar intuitive
symbols instead of new ones that need to be
learnt.
So tell me what is so familiar and intuitive about:
{rho} {log} {base} {encode} {grade} etc. etc.
I think you have to admit at least that they
: Re: [Jgeneral] Displaying J with Unicode Special Symbols
Deny my existence at your peril, Roger :-)
- Original Message -
From: Roger Hui [EMAIL PROTECTED]
To: General forum general@jsoftware.com
Sent: Wednesday, February 07, 2007 6:21 PM
Subject: Re: [Jgeneral] Displaying J
A result shape of 5 1 in that case is consistent
with the specifications in the dictionary but
one of ,5 is not.
Also:
y=: : i.5
$@(q:y). 1 2 3 4 5 6
+---+---+---+---+---+---+
|5 1|5 2|5 3|5 4|5 5|5 6|
+---+---+---+---+---+---+
Making the first one ,5 instead of 5 1 would be
anomalous.
it.
The results that don't into integer magnitude domain
should probably conform with usual J rules. I.e. the
user is responsible to ensure that the result will
stay integer.
--- Roger Hui [EMAIL PROTECTED] wrote:
How would [EMAIL PROTECTED] be different from just * ? What
would
.
But at the expense of speed, which should be
compensated by special code.
--- Roger Hui [EMAIL PROTECTED] wrote:
A quick benchmark:
x=: 1e6 [EMAIL PROTECTED] 0
y=: 1e6 [EMAIL PROTECTED] 1000
t=: x*y
ts=: 6!:2 , 7!:[EMAIL PROTECTED]
ts '.t'
0.201745 1.25836e7
I have looked further into the matter.
y=: 1000 * 1e6 [EMAIL PROTECTED] 0
7!:2 '.y'
12583552
One expects that it should take 4 bytes per
element but it takes more than 8. The problem
is that .y uses an intermediate result which
is float and then converts it into integers
if possible,
Message -
From: Roger Hui [EMAIL PROTECTED]
Date: Monday, February 12, 2007 10:00 pm
Subject: Re: [Jgeneral] large space overhead
I have looked further into the matter.
y=: 1000 * 1e6 [EMAIL PROTECTED] 0
7!:2 '.y'
12583552
One expects that it should take 4 bytes per
element
: [Jgeneral] large space overhead
What processor are you running Roger?
I cannot get anything like that speed on a celeron 2 GHz, J601c
y=: 1 * _0.5 + 1e6 [EMAIL PROTECTED] 0
ts=: 6!:2 , 7!:[EMAIL PROTECTED]
ts '.y'
1.00667 1.25835e7
From: Roger Hui [EMAIL PROTECTED]
Reply
bericht-
Van: [EMAIL PROTECTED] [mailto:general-
[EMAIL PROTECTED] Roger Hui
Verzonden: dinsdag 13 februari 2007 0:28
Aan: General forum
Onderwerp: Re: [Jgeneral] Shape of 1q:
A result shape of 5 1 in that case is consistent
with the specifications in the dictionary but
one of ,5
http://www.ddj.com/dept/architect/196900217
- Original Message -
From: Eric Iverson [EMAIL PROTECTED]
Date: Tuesday, February 20, 2007 8:07 am
Subject: [Jgeneral] Feb 2007 issue of Dr Dobbs
There is a backhanded compliment (longevity) to APL and J in the
Feb
2007 issue of Dr.
It depends on what is meant by garbage collection.
There is definitely no memory compaction, moving
stuff around so that free blocks are collected
together into one contiguous block. Memory
compaction does not make sense in modern virtual
memory systems.
As far as memory management is
to Donald Knuth and Roger Hui. The final N,B and S
solution uses just 9 tokens.
Is it really 19 years since I wrote it? It was lots of fun.
Eugene
--
For information about J forums see http://www.jsoftware.com/forums.htm
-
From: ramacd [EMAIL PROTECTED]
Date: Friday, March 9, 2007 11:48 am
Subject: Re: [Jgeneral] Feb 2007 issue of Dr Dobbs
Hi Roger;
Sorry if this is a STFW but what font do I need to install to see
all the
APL?
- Original Message -
From: Roger Hui [EMAIL PROTECTED]
To: General forum
I would appreciate it if you can apply the same
process to
http://www.jsoftware.com/jwiki/EditingGuidelines/Character_Palette_--_APL
and tell me what problems you run into.
What is UnicAPL?
- Original Message -
From: Markus Triska [EMAIL PROTECTED]
Date: Friday, March 9, 2007 2:41 pm
: [Jgeneral] Feb 2007 issue of Dr Dobbs
On 8 Mar 2007 at 21:08, Roger Hui said:
The text of Life: Nasty, Brutish, and Short can be
found at http://www.jsoftware.com/jwiki/Doc/Life
I rather like the penultimate clause of the footnote [Ho51]
explaining
the source of the title
I started using {rho}0 to denote the empty list
in APL because (a) it is shorter than 0{rho}0
(b) it does not depend on {quad}io like {iota}0
does (c) it has right type where '' does not.
From APL to J {rho}0 got changed to $0.
Of course strictly speaking it should be an empty
list rather
?
: ...
- Original Message -
From: Roger Hui [EMAIL PROTECTED]
To: General forum general@jsoftware.com
Sent: Friday, March 16, 2007 12:49 PM
Subject: Re: [Jgeneral] definition of a:
I started using {rho}0 to denote the empty list
in APL because (a) it is shorter than 0
, the global []IO takes over if the localized one is
undefined
and []-[]EX'IO' gives 0.
Roger Hui [EMAIL PROTECTED]
Sent by: [EMAIL PROTECTED]
03/16/2007 02:41 PM
Please respond to
General forum general@jsoftware.com
To
General forum general@jsoftware.com
cc
Subject
Re: [Jgeneral
Backus had a reciprocal relationship with APL.
He admired it but didn't think it had gone far
enough. In his FP he described an insert
function which was later adopted by APL, as a
better way to describe what was then called
reduction in APL.
I don't think this description of how
a. It is true that the J Wikipedia page does not
tell you directly what the functional forms are.
However, the quicksort examples shows off
functional programming pretty well.
b. Since it is the Wikipedia, if you don't like
something, why don't you do something about it?
- Original
Yes, there should be a period between x and If.
This is now fixed. Thanks for finding and reporting
this error.
Indexing is a function and like all functions the
arguments are arrays. The left argument is an array
of integers in the range from -n to strictly less
than n; the right argument is
For ordinary splitting, you should start with ;.
rather than ;: . For example:
;._1 ' dazlious eleemosynary deipnosophist aba'
+++-++---+
|dazlious|eleemosynary|deipnosophist||aba|
+++-++---+
You will not find an actual
There are optimizations but they require that you
express the computation differently, in this case as
+/@: . The optimizations are listed in Appendix B
of the dictionary.
http://www.jsoftware.com/help/dictionary/special.htm
You can detect that special code is being employed
by doing some
Your expressions all produce a 64-bit IEEE
floating point number first, and then converted
it into extended precision. To avoid that,
append an x to the string before executing it:
. 'x',~ 50#'1'
11
. 'x',~ 50#'2'
Some languages and at least one (old) version of
APL identify atoms with 1-element vectors. This
treatment has consequences such as muddling the
treatment of unit axes. An example demonstrates what I mean:
If i is an array of integers and m is a matrix,
then i{m has shape ($i),(}.$m).
m=:
I think you misunderstand the second example.
It says:
The four 5x6 2-cells are arranged as a vector
of 2-cells; this is the meaning of their frame.
Each of the 2-cells (rank 2) cells has shape 5 6.
The frame is ,4 . The frame catenated with the
cell shape is (,4) catenated with 5 6, or 4 5
a. In J noun is a synonym for array.
A statement such as every noun is an array is
misleading as it leaves open the possibility that
there are some arrays which are not nouns, like
saying every man is an animal.
b. The proposed use of dimension to denote the
rank of an array and also a
It doesn't take a scholarly knowledge of J to know
that two different arrays can have the same display:
2+3 4 5
5 6 7
'5 6 7'
5 6 7
,:'5 6 7'
5 6 7
- Original Message -
From: Terrence Brannon [EMAIL PROTECTED]
Date: Wednesday, April 4, 2007 2:30 pm
Subject: Re: [Jgeneral] J
Some non-mathematician friends once asked me
why 1 isn't a prime. The answer I gave, off the
top of my head, is that it makes many statements
shorter, else you'd saying lots of times, for
every prime 1 or for every prime other than 1.
It is now years later, and I have had lots of
time to think
help\dictionary\dsco.htm:
Symbols are a new data type...
I guess that was true at one time, but they're not
new any more.
OK.
- Original Message -
From: Henry Rich [EMAIL PROTECTED]
Date: Wednesday, April 4, 2007 4:07 pm
Subject: [Jgeneral] A couple of doc errors
While I was
In what ways are atoms degenerate or improper?
- Original Message -
From: Oleg Kobchenko [EMAIL PROTECTED]
Date: Wednesday, April 4, 2007 4:22 pm
Subject: Re: [Jgeneral] Need a word for a non-atomic array
There is no such thing as two-demensional atomic
arrays. All arrays with
You can finesse it by focussing on items instead of
saying list of items. Talking about list of items
has other potential problems when the number of
items is 0.
The dictionary says this in section II a
http://www.jsoftware.com/help/dictionary/dicta.htm
A cell of rank one less than the
/4/07, Roger Hui [EMAIL PROTECTED] wrote:
In what ways are atoms degenerate or improper?
--
For information about J forums see http://www.jsoftware.com/forums.htm
The J dictionary uses atom and scalar
synonymously, much like verb and function, and
does not distinguish atom and scalar in the way
described in the appended msg. Section II A
http://www.jsoftware.com/help/dictionary/dicta.htm
says:
Arrays of ranks 0, 1, and 2 are also called
_atom_,
There is an executable model (algorithmic
description) of the rank conjunction in my APL95
paper Rank and Uniformity.
http://www.jsoftware.com/papers/rank.htm
- Original Message -
From: Terrence Brannon [EMAIL PROTECTED]
Date: Thursday, April 5, 2007 8:42 am
Subject: [Jgeneral]
Rank is used in APL/J to denote the number of
atoms in the shape of a noun. It made its first
published appearance in Ken Iverson's A Programming
Language, 1962. Section 1.20, Levels of Structure,
says:
Vectors and matrices are arrays which exhibit one
level and two levels of structure,
cells array of boxed cells
etc.
- Original Message -
From: Raul Miller [EMAIL PROTECTED]
Date: Thursday, April 5, 2007 10:59 am
Subject: Re: [Jgeneral] fleshing out the operation of negative verb rank
On 4/5/07, Roger Hui [EMAIL PROTECTED] wrote:
There is an executable model
x=: 1 1 1$5
x
5
$ x
1 1 1
It can show the shape of an array just fine.
- Original Message -
From: bill lam [EMAIL PROTECTED]
Date: Thursday, April 5, 2007 9:30 pm
Subject: Re: [Jgeneral] Need a word for a non-atomic array
Mark D. Niemiec wrote:
'x' and 42 and a: are all
If you are interested in the history of ideas in J,
you can read some of papers in
http://www.jsoftware.com/jwiki/Essays/Bibliography
In particular,
Falkoff and Iverson, The Design of APL, 1973.
Falkoff and Iverson, The Evolution of APL, 1978.
Iverson, A Personal View of APL, 1991.
The last
For examples of f^:(n), see:
http://www.jsoftware.com/jwiki/Essays/Text_Formatting
http://www.jsoftware.com/jwiki/Essays/Non-Overlapping_Substrings
http://www.jsoftware.com/jwiki/Essays/Collatz_Conjecture
http://www.jsoftware.com/jwiki/Essays/Continued_Fractions
- Original Message -
It is not some deep mysterious concept.
For example:
sumvector=: 3 : 0
z=. 0
for_i. i.#y do. z=. z+i{y end.
z
)
I have written verb sumvector that finds the
sum of a vector:
sumvector 1 2 3 4 5
15
Now suppose I want to find the sums of a bunch
of vectors:
x=: 2 3 7 [EMAIL PROTECTED]
Judging by
http://www.jsoftware.com/help/dictionary/partsofspeech.htm
EVOKE on the vocabulary page should be in
mixed-case bold italic.
- Original Message -
From: Henry Rich [EMAIL PROTECTED]
Date: Sunday, April 15, 2007 10:33 am
Subject: RE: [Jgeneral] Is EVOKE (~) an adverb ?
~ is
What are some reasons for your assertion?
- Original Message -
From: bill lam [EMAIL PROTECTED]
Date: Tuesday, April 17, 2007 3:21 am
Subject: Re: [Jgeneral] J for C Programmers - asymmetry / control /data
discussed without definition?
Roger Hui wrote:
However, if a verb
I basically agree with this point. On these grounds, ! is the wrong
way round: you are more likely to calculate x ! y for fixed y than
fixed x.
What are some examples of fixing an argument of ! ? The only non-trivial one I
am aware of is
2! = +/@i.
- Original Message -
From:
Bill Lam's analysis was correct. For a dyad,
once you have figured out the frame and cells,
it remains to apply something to corresponding
cells, and with respect to that dyad that
something is a black box.
You have figured out that the cells are
1 0 1 0 0 0 and 3 1 4 1 5 9. (Since the left
There was a debate in I.P. Sharp Associates
on the left rank of { , in 1987, prior to the
existence of J. The outcome hinged on the fact
(expressed by Ken Iverson) that left rank 0
permits more sensible index arrays (arrays of
indices). With left rank 1 an index array would
have an extra
, apply the existing minus_zero
eradicator to every array.
- Original Message -
From: Dan Bron [EMAIL PROTECTED]
Date: Monday, April 23, 2007 10:29 am
Subject: Re: [Jgeneral] zero
RM = Raul Miller, DB = me, RH = Roger Hui
DB %floating_zero
DB_
RM
The interpreter knows what arrays it has;
The interpreter can not do anything without
allocating memory;
The interpreter allocates memory in a single place;
In the single place, apply the existing minus_zero
eradicator to every array.
a. It doesn't just eradicate minus_zero in memory
that
What can you do in that time? In J, nothing.
That time is not the time from one user entry to
the next, not even the time from the invocation of
one user-defined entity to the next, but from one
allocation of memory to the next.
I can also use part of my one hour to disable all
15!:x foreigns.
1 - 100 of 331 matches
Mail list logo
