Thank you Raul for response. This optimization can be implemented as a
customized table
with computational time saving of a factor of 2. Does the pattern occur
often enough to be worthwhile?
Is a factor of 2 sufficient to trouble with?
I forget Roger's rule, he took care to change the source i
Rather than a new table adverb, I would prefer a conjunction that
asserted properties of a verb/noun: commutative, left-associative, and
right-associative come to mind for verbs, Hermitian and antiHermitian
for nouns, and there may be others. Then the existing adverbs like [x]
u/\ y and x u/ y
IIRC I was able to reproduce the crash, it should be the change in J engine
that affected cd call. I'll take a look when I have spare time.
On Thu, Jul 5, 2018, 1:20 AM J. Patrick Harrington
wrote:
> Just installed beta-g
>
> JVERSION
> Engine: j807/j64/linux
> Beta-g: commercial/2018-07-02T
Just installed beta-g
JVERSION
Engine: j807/j64/linux
Beta-g: commercial/2018-07-02T12:43:58
Library: 8.07.16
Qt IDE: 1.7.1/5.5.1
Platform: Linux 64
Installer: J807 install
Most things I've tried seem OK.
But the Fast Fourier Transform lab still crashes as before.
Does anyone know how to fix
The functionality of scalar&#: can be replaced with the scalar&| . The
problem that some may be using scalar&#: would need to fix their code, but
it is not necessary to go to elaborate means to make sure the scalar is
converted to a one element vector. Simply use scalar&| .
On Wed, Jul 4, 2018 at
No proposed change to #. - further, the only change to #: would be to require
the left argument be a vector IF one wants to match all current results.
Really just a scalar extension in the spirit of -
1 + i.3
> On 2018Jul 4, at 08:36, Skip Cave wrote:
>
> I use #. quite a bit, to convert
I use #. quite a bit, to convert strings of digits to integers:
Find all the three-digit integers that can be formed from the digits 2, 4,
& 6, with no duplication (permutations).
10#. 2 4 6 {~ (i.@! A.i.) 3
246 264 426 462 624 642
Woould this functionality stay the same?
Skip
Skip Cave
Cav
I agree with Joey Tuttle's idea. I can't think of any instances when
I've wanted scalar x to be x|
8 (| -: #:)j./~i:20
1
Furthermore j's antibase returns 1 quarter, a half dime, a nickel, and 3
pennies when asked to make change for 38 cents. My version gives 1
quarter, 1 dime, and 3 pen
In at least one of Arthur Whitney's languages scalar encode x does what
Joey is proposing (returns a result with the number of required digits with
no leading zeros).
--
For information about J forums see http://www.jsoftware.c
In cases I needed only one number to encode I used residue.
5#:123
3
5|123
3
On Wed, Jul 4, 2018, 7:45 AM Jan-Pieter Jacobs
wrote:
> I'm also in favor of #: working as #.^:_1 .
>
> I have no clue whether much would break, but I can imagine if anybody does
> need it they can as well use {
I'm also in favor of #: working as #.^:_1 .
I have no clue whether much would break, but I can imagine if anybody does
need it they can as well use {:@:#: . If this would be a problem
performance-wise, maybe the old logic could be applied in this case as
special code?
Just my 2 cents.
Best regar
I agree with you about the desirability of changing the definition but
boy! what code it might break! We have to decide whether the change is
worth the trouble.
On second thought, maybe not so much would break. I know that dozens of
times I have written (atom #: list) and gotten the result I
This update fixed the issue I had.
Thanks
On 7/3/2018 14:21, Eric Iverson wrote:
Jd (Jdatabase) new release available with fix for a bug in csvscan that can
result in a crash in the subsequent csvrd.
--
For information about J f
Beta g fixed the issue I had.
Thanks.
On 7/2/2018 21:01, Eric Iverson wrote:
J807 beta-g is now available for window/linux/macos.
This is a relatively small release that fixes one particular problem that
was discovered by beta users. Thanks for the feedback. Please have another
go.
---
14 matches
Mail list logo
