Thank you Brian, using 'type symbol' does avoid the length error, e.g.
with -
iMi7:~ jkt$ j64
load 'plot'
pd 'type symbol'
pd 2;3
pd 'show'
pd 'type line'
pd i.5
pd 'show'
but puts a tiny red numeral 2 at (approximately) 2.2,3 on the canvas,
then shows the line i.5 as expected. But, if in the other order -
iMi7:~ jkt$ j64
load 'plot'
pd i.5
pd 'show'
NB. line looks as expected
pd 'type symbol'
pd 2;3
pd 'show'
The i.5 line disappears from the canvas (and the symbol 2) is the only
remaining thing, again at coordinates 2.2,3
If instead of (pd 2;3), I plot two symbols e.g. (pd 2;3 1) then the
canvas becomes completely blank - this doesn't happen with 'type point',
but the i.5 plot does change to points... By the way, a slightly simpler
workaround for the scalar 'type point' is
pd 2;,3
which avoids the length error in pdf_circle ... BTW, this is with all
current libraries in place (and results are same in 32 and 64 bit
versions) -
iMi7:~ jkt$ j64
load 'pacman'
'status' jpkg ''
Local JAL information was last updated: 06 Dec 2012 08:49:37
All available packages are installed and up to date.
~~~
When my friend and I stumbled across these things, we were actually
trying to use the NLLS verbs described in
http://www.jsoftware.com/jwiki/NYCJUG/2010-11-09/Levenberg-MarquardtAlgorithm?highlight=%28NLLS%29
which raised several other issues that I want to address in a separate
thread.
On 2012/12/06 04:44 , Brian Schott wrote:
Joey,
Below is a copy of the script in j60232 that produces both lines and
symbols on the same plot. It is taken from the demo named
Gallery|Curve Fitting. Perhaps type symbol would work more
predictably?
y=. 3 2 5 7 6 5 7 4
x=. i.#y
dat=. x,:y
m=. cubicspline dat
x2=. steps 0,(<:#y),50
y2=. m interspline x2
y3=. (3 lsfit dat) p. x2
y4=. (5 lsfit dat) p. x2
pd 'title Spline vs Cubic vs Quintic'
pd 'backcolor lightgray'
pd 'type symbol'
pd x;y
pd 'type line'
pd x2;y2,y3,:y4
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
