bob@loihi.hig.hawaii.EDU (Bob Cunningham) (11/25/87)
For those of you using DISSPLA (aka CA-DISSPLA), you should be aware
of a potentially-serious bug with THKCRV and THKRND calls that---with
closely-spaced data---causes "spikes" to appear on plots. Under
some circumstances, those spikes can cause serious distortion of plots.
To demonstrate, try this program. Then, try it again after
uncommenting line 5. Choice of device doesn't matter, this bug is
device-independent.
parameter (npts=4000)
dimension x(npts), y(npts)
integer*4 jjj
call ln03tk ! or your favorite device
c call thkcrv(0.1)
call page(10.43,7.5)
call area2d(8.0,6.0)
call xname('X VALUES$',100)
call yname('Y VALUES$',100)
jjj=57387 ! seed for RAN()
do i=1,npts
x(i) = float(i)
y(i) = ran(jjj)*ran(jjj) ! between 0.0 & 1.0
end do
call graf(0.0,'SCALE',x(npts),0.0,'SCALE',1.0)
call curve(x,y,npts,0)
call endpl(1)
call donepl
end
CA Support knows all about this (ref problem #1117), but seem to be a
bit slow in coming out with a fix.
Bob Cunningham
(bob@loihi.hig.hawaii.edu)