At 12:20 -0400 2009/06/16, Cliff Reiter wrote:
>Another observation:
>Running the "15" the density plot is sometimes (I think) correct when
>it is created, but subsequent invocations of "v" give different
>changing near infinity values (but don't crash).
>
> load 'plot'
> sinc=:(sin%])`1:@.(=&0)"0
> x=:0.1*i:200
> g=:+&.:*:"0
> y=:g/~ 0.1*i:200
> 'density' plot y
> 2 2{.v=:sinc y=:g/~ 0.1*i:15
>0.401755 0.432062
>0.432062 0.463396
> >./,y
>2.12132
> >./,v NB. OK
>1
> 'density' plot v
> v
>4.24399e_314 9.90108e_321 9.33678e_313 4.25248e_311 3.45885e_312
>4.24399e_314 9.90108e_321 1.35808e_312 4.25248e_311 3.24665e_312
>4.24399e_314 9.90108e_321 1.95224e_312 4.25248e_311 2.94957e_312 2
>
>and on one occasion, it changed half "v" on a small "5"
> ........
Those extremely small values certainly look "suspicious" - the crash
would seem to be a v which is "damaged" so the magnitudes may be an
artifact of the damage.
I too reduced it to a script:
rr =: 3 : 0
sin =: 1&o.
sinc=:(sin%])`1:@.(=&0)"0
g=:+&.:*:"0
2 2{.v=:sinc z=:g/~ 0.1*i:y
2 2{. v
)
and get the following results (and like you, occasionally the case
that usually works crashes instead).
In 32 bit Linux -
version ''
sh: sw_vers: command not found
Installer: j602a_linux32.sh
Engine: j602/2008-03-03/16:45
Library: 6.02.023java version "1.4.2" gij (GNU libgcj) version 4.1.2
20070626 (Red Hat 4.1.2-13) Copyright (C) 2006
rr 22
0.00974463 0.0328942
0.0328942 0.0575454
rr 23
Segmentation fault
In Darwin (Mac OS X - PowerPC G5)
version''
Installer: j602a_mac_powerpc.dmg
Engine: j602/2008-03-03/16:45
Library: 6.02.023ProductName: Mac OS X
ProductVersion: 10.5.7
BuildVersion: 9J61
java version "1.5.0_16" Java(TM) 2 Runtime Environment, Standard
Edition (build 1.5.0_16-b06-284) Java HotSpot(TM) Client VM (build
1.5.0_16-133, mixed mode, sharing)
rr 15
0.401755 0.432062
0.432062 0.463396
rr 16
Segmentation fault
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
