I can't believe what a relief this is:
iroots
[: ,. _1 ^ [: +: i. % ]
iroots ttem
3 : 0
r0=. i. y
q0=. +: r0 % y
,. (_1) ^ q0
)
irootse=: 13 :',.(_1)^+:(i.y)%y'
irootse
[: ,. _1 ^ [: +: i. % ]
Linda
-Original Message-
From: programming-boun...@forums.jsoftware.c
Or maybe imaginary roots:
Iroots=: 13 :',._1^+:(i.y)%y'
iroots
[: ,. _1 ^ [: +: i. % ]
iroots 3
1
_0.5j0.866025
_0.5j_0.866025
Linda
-Original Message-
From: programming-boun...@forums.jsoftware.com
[mailto:programming-boun...@forums.jsoftware.com] On Behal
wrote: "My roots-of-unity verb was designed to get
higher accuracy than simpler approaches."
The verb
PoU=:_1^+: NB. Power of Unity
is not contaminated by transcendental numbers such as (2p1) or transcendental
functions such as monadic (^). If (y) is rational then (PoU y) is algebraic, so
@Linda I simply keep it in my ~user folder (my own version, that is).
It can be loaded automatically by putting the appropriate load
statement in ~user/config/startup.ijs . But I don't do that: I load it
at need. I have things set up to do that conveniently.
You could of course put its name and lo
Where do you put the file 'tte.ijs' so it will be accessed?
It is a very impressive script even without all the ideas you have to
augment it.
Linda
-Original Message-
From: programming-boun...@forums.jsoftware.com
[mailto:programming-boun...@forums.jsoftware.com] On Behalf Of Ian Clark
I don't like this because usually I don't want to pad each item. If they
had the same dimensions I wouldn't box 'em in the first place.
On Tue, Jul 17, 2012 at 3:23 PM, Linda Alvord wrote:
> How does this rank?
>
> <"1(>x),"1 0>y
> ┌───┬───┬───┬───┐
> │CogitoA│, B│ergo
How does this rank?
<"1(>x),"1 0>y
┌───┬───┬───┬───┐
│CogitoA│, B│ergo C│sum. D│
└───┴───┴───┴───┘
Linda
-Original Message-
From: programming-boun...@forums.jsoftware.com
[mailto:programming-boun...@forums.jsoftware.com] On Behalf Of Roger Hui
S
, each i. 10
would put a list in each box created.
] each i. 10
would put a scalar in each box
Cheers, bob
On 2012-07-17, at 2:14 PM, Dan Bron wrote:
> It's not abuse - that's perfectly valid. Note, however:
>
>
> ]A=:1 <\ i. 10
> +-+-+-+-+-+-+-+-+-+-+
> |0|1|2|3|4|5|6
Nice.
With <"0 sometimes the 0 interacts with the argument and you'd need some
kind of separate. ;/ doesn't not suffer from this.
<"0 ]1 2 3
┌─┬─┬─┐
│1│2│3│
└─┴─┴─┘
;/1 2 3
┌─┬─┬─┐
│1│2│3│
└─┴─┴─┘
On Tue, Jul 17, 2012 at 2:33 PM, chris burke wrote:
> Also:
>
> ;/i.10
>
> (argument sho
Also:
;/i.10
(argument should be non-empty)
On Tue, Jul 17, 2012 at 10:11 AM, Devon McCormick wrote:
> <"0 i.10 or ]&.> i.10
>
> On Tue, Jul 17, 2012 at 12:49 PM, Kim Kuen Tang
> wrote:
> > Hi everyone,
> >
> > is there a better way to box each element ?
> >
> > 1 <\i.10
> >
> > Thx for any
It's not abuse - that's perfectly valid. Note, however:
]A=:1 <\ i. 10
+-+-+-+-+-+-+-+-+-+-+
|0|1|2|3|4|5|6|7|8|9|
+-+-+-+-+-+-+-+-+-+-+
]B=:<"0 i. 10
+-+-+-+-+-+-+-+-+-+-+
|0|1|2|3|4|5|6|7|8|9|
+-+-+-+-+-+-+-+-+-+-+
Yes, plus each boxed element is a vector rather than a scalar.
$&.> <"0 i.10
┌┬┐
│││
└┴┘
$&.> 1 <\i.10
┌─┬─┬─┬─┬─┬─┬─┬─┬─┬─┐
│1│1│1│1│1│1│1│1│1│1│
└─┴─┴─┴─┴─┴─┴─┴─┴─┴─┘
On Tue, Jul 17, 2012 at 1:35 PM, Kim Kuen Tang wrote:
>
> 1<\i.10
>
>
> is for me not ideal be
1<\i.10
is for me not ideal because you are abusing the adverb \ .
Am 17.07.2012 21:15, schrieb Linda Alvord:
<"0 i.10
--T-T-T-T-T-T-T-T-T-┐
│0│1│2│3│4│5│6│7│8│9│
L-+-+-+-+-+-+-+-+-+--
I'm not sure what you will call better.
Linda
-Original Message-
From: programming-boun...@forums
What could be better is, if you are not just boxing each element but doing
something to each element that may involve boxing it. In such a case f&.>
aka f each may be better. For example:
x=: ;: 'Cogito, ergo sum.'
x
┌──┬─┬┬┐
│Cogito│,│ergo│sum.│
└──┴─┴┴┘
y=: 'AB
<"0 i.10
--T-T-T-T-T-T-T-T-T-┐
│0│1│2│3│4│5│6│7│8│9│
L-+-+-+-+-+-+-+-+-+--
I'm not sure what you will call better.
Linda
-Original Message-
From: programming-boun...@forums.jsoftware.com
[mailto:programming-boun...@forums.jsoftware.com] On Behalf Of Kim Kuen Tang
Sent: Tuesday, July
On Tue, Jul 17, 2012 at 2:07 PM, wrote:
> My roots-of-unity verb was designed to get higher accuracy than simpler
> approaches. If you just take numbers from 0 to pi you in effect lose an ULP
> or two of significance. So, I calculated over a smaller & more accurate
> interval & used reflecti
My roots-of-unity verb was designed to get higher accuracy than simpler
approaches. If you just take numbers from 0 to pi you in effect lose an ULP or
two of significance. So, I calculated over a smaller & more accurate interval
& used reflection and sign-change. The results of FFT using the
Very cool,
many thx.
Kim
Am 17.07.2012 19:11, schrieb Devon McCormick:
<"0 i.10 or ]&.> i.10
On Tue, Jul 17, 2012 at 12:49 PM, Kim Kuen Tang wrote:
Hi everyone,
is there a better way to box each element ?
1<\i.10
Thx for any comments or solutions.
Kim
--
<"0 i.10 or ]&.> i.10
On Tue, Jul 17, 2012 at 12:49 PM, Kim Kuen Tang wrote:
> Hi everyone,
>
> is there a better way to box each element ?
>
> 1 <\i.10
>
> Thx for any comments or solutions.
>
> Kim
> --
> For information abo
Another option might be:
] each i. 10
where 'each' is already defined in J as &.>
Although this looks closer to a direct english translation, it is actually
using the functionality of 'each' to do the boxing.
Cheers, bob
On 2012-07-17, at 9:55 AM, Joey K Tuttle wrote:
> <"0]i.10
>
> On
<"0]i.10
On 2012/07/17 09:49 , Kim Kuen Tang wrote:
Hi everyone,
is there a better way to box each element ?
1 <\i.10
Thx for any comments or solutions.
Kim
--
For information about J forums see http://www.jsoftware.com
Hi everyone,
is there a better way to box each element ?
1 <\i.10
Thx for any comments or solutions.
Kim
--
For information about J forums see http://www.jsoftware.com/forums.htm
Thanks, Linda. I often feel I come out with things that everybody
knows about except me.
I'm delighted you share my enthusiasm for Ambrus's tte (ttem/ttes).
The writeup says it's unfinished, and suggests it may fail on hooks. I
use it with a slight modification to run it in its own locale, which
m
On Tue, Jul 17, 2012 at 10:52 AM, Istvan Kadar wrote:
> Dear Forum,
>
> load 'viewmat'
>a=: ?20 20$2
>b=:viewmat a
>$b
> 0 0
>#b
> 0
The result from viewmat is nothing -- no rows, and no columns of nothing.
>c=:(viewmat^:_1) b
> |domain error
viewmat is not designed to displ
On Tue, Jul 17, 2012 at 5:12 AM, Bo Jacoby wrote:
> Yes, the roots-of-unity verb rou from the Essay/FFT is only to be used for
> computing this fft.
Ok... but I do not see any really good reasons for pushing that -:
into the definition of rou.
You can just as easily use rou@-:
--
Raul
---
Yes, the roots-of-unity verb rou from the Essay/FFT is only to be used for
computing this fft.
A perfectly symmetric transform is nice
FT=:([:(%[:%:#)+floop&.cube rou@#) :.FT NB. Fourier Transform
(=FT^:2)i.16 NB. test
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
but then the convolution is trickier
26 matches
Mail list logo
