Let's say you make a mistake:
rms=: 2 :'%:+/*:y'
You test your defined word and notice your mistake:
rms 2 3 5
|value error: rms
So you fix it, and test again:
rms=: 3 :'%:+/*:y'
rms 2 3 5
6.16441
But verbs can only take nouns as arguments.
Thanks,
--
Raul
On Thu, Nov 20, 2014 a
Those are all very pretty.
Henry Rich
On 11/19/2014 10:29 PM, Marshall Lochbaum wrote:
None of these beat Roger's by your criteria, but here are a few:
(,,.)&.>/ i.3 3
<"_1|:,: i.3 3
(0 1,:3 1) <;.3 i.3 3
The second features a rare specimen--monadic |: on an array of rank
greater than 2. O
=: is special because the lvalue (address) of the left argument is
used for evaluation.
Чт, 20 ноя 2014, Sergey Kamenev написал(а):
> 20.11.2014 09:05, Jan-Pieter Jacobs пишет:
> >To give you a more direct reply:
> >
> >conjunctions, like & and @ work on either verbs or nouns.
> >=: and =. are ne
And because the right argument might not be a noun (and might not be a verb).
--
Raul
On Thu, Nov 20, 2014 at 9:40 AM, bill lam wrote:
> =: is special because the lvalue (address) of the left argument is
> used for evaluation.
>
> Чт, 20 ноя 2014, Sergey Kamenev написал(а):
> > 20.11.2014 09:05
> This violates universality.
Universality is not necessarily a good principle for language design.
1.2 3.4 $ 5.6
|domain error
| 1.2 3.4$5.6
'abc' $ 3 4 5
|domain error
| 'abc'$3 4 5
On Wed, Nov 19, 2014 at 11:46 PM, Sergey Kamenev
wrote:
> 20.11.2014 09:05, Jan-Pieter J
If I'm allowed to jump in here, I'm using ;. a lot to subdivide images into 2x2
blocks (think wavelet), and I guess it's okay to do it like:
]z =: (2 2,:2 2) < ;. _3 i.4 6
+-+-+-+
|0 1 |2 3 | 4 5|
|6 7 |8 9 |10 11|
+-+-+-+
|12 13|14 15|16 17|
|18 19|20 21|22 23|
+-
Likely this comment should be in chat, but ...
https://www.cs.purdue.edu/homes/dec/essay.criticize.html
On 2014/11/20 07:12 , Roger Hui wrote:
This violates universality.
Universality is not necessarily a good principle for language design.
1.2 3.4 $ 5.6
|domain error
| 1.2 3.4$
Perhaps
,"1/,"2/,"3/>zz
Thanks,
--
Raul
On Thu, Nov 20, 2014 at 10:42 AM, Ben Gorte - CITG
wrote:
> If I'm allowed to jump in here, I'm using ;. a lot to subdivide images into
> 2x2 blocks (think wavelet), and I guess it's okay to do it like:
>
>]z =: (2 2,:2 2) < ;. _3 i.4 6
> +-+
Wow !
I understand them when I see them, but it seems I cannot 'forward engineer'
them.
Well, not always.
Thanks a lot!
Ben
From: programming-boun...@forums.jsoftware.com
[programming-boun...@forums.jsoftware.com] on behalf of Raul Miller
[rauldmil..
Hi -
I was just burned by this apparent anomaly:
1 e. 0~: i. 2 2
0
but
1 e. , 0~: i. 2 2
1
Has anyone else noticed this? Am I right in thinking this is incorrect
behavior in the first case?
Regards,
Devon
--
Devon McCormick, CFA
-
Both expressions give the correct answer.
The right argument of membership determines the target of the items to be
sought, just like the left argument of index-of determines the targets. In
my opinion, the APL\360 team made a mistake by making x∊y the same as
x∊,y. We are struggling with the co
I see, so
0 1 e. i.2 2
1
On Thu, Nov 20, 2014 at 1:27 PM, Roger Hui
wrote:
> Both expressions give the correct answer.
>
> The right argument of membership determines the target of the items to be
> sought, just like the left argument of index-of determines the targets. In
> my opinion, th
has to do with the definition of item
0 1 e. i. 2 2
1
1 3 e."0 1 i. 2 2
1 1
1 e."1 i. 2 2
1 0
- Original Message -
From: Devon McCormick
To: J-programming forum
Cc:
Sent: Thursday, November 20, 2014 1:19 PM
Subject: [Jprogramming] Correct behavior of "e." ?
Hi -
I was
Correct.
The following example may be more convincing:
x=: ];._1 ' Devon Roger Ken Morten'
y=: ];._1 ' Roger Sextus'
x
Devon
Roger
Ken
Morten
y
Roger
Sextus
y e. x
1 0
One way to define membership is y e. x iff (#x)>x i. y. Defining the "set
functions" this way makes even more s
Aha! Raul's solution can be rewritten as
,"_2/^:3 >zz
(negative rank is quite a useful concept). Since ,. is (exactly) ,"_1
the two-dimensional solution is
,"_1/^:2 >z
Extending the pattern, we can reduce a blocked array using
merge =: 3 : 0
r =. #$y
,"(1-r)/^:r >y
)
This works on block array
I'm new to J and am struggling to figure out why this works.
'';1 just looks like a noun:
'';1
┌┬─┐
││1│
└┴─┘
<;.1 i.3 3
┌─┐
│0 1 2│
│3 4 5│
│6 7 8│
└─┘
thanks for any help,
Ryan
On 20 Nov 2014, at 9:27, Roger Hui wrote:
> ('';1) <;.1 i.3 3
> ┌─┬─┬─┐
> │0│1│2│
> │3│4│5│
> │6│7│8│
> └─
Yes, '';1 is a noun, but having a noun changes <;.1 completely:
<;.1 i.3 3
executes the MONAD <;.1, which behaves according to its rules.
('';1) <;.1 i.3 3
executes the DYAD <;.1 which behaves according to its rules, which are
totally different (though somewhat similar in this case).
Every
Perhaps also worth noting that while the dictionary provides
documentation -- at http://www.jsoftware.com/help/dictionary/d331.htm
-- for boxed left arguments for u;.1, NuVoc -- at
http://www.jsoftware.com/jwiki/Vocabulary/semidot1#dyadic -- does not.
Thanks,
--
Raul
On Thu, Nov 20, 2014 at 9:4
18 matches
Mail list logo
