Re: [Gretl-users] Bug in Gretl-Gnuplot interaction under Linux
On Tue, 29 Dec 2009, Jaromir Baxa wrote:
Here is the dataset - the testdata.gdt. Imported from OECD, than
from OpenOffice. The file TestPlot1.png shows two plots.
First, there is a plot of an output gap. Cursor was placed to
the highest point at the peak before 1990. The position
identified in the left-low corner of the plot shows it is in the
second quarter of 1989 although the data show it was in the
fourth quarter of 1989.
The read-back of the cursor position is only a rough approximation
unless you're using gnuplot 4.3 or higher. Perhaps we should not
do this at all unless we have a new enough gnuplot.
There's no problem like this in Windows on the same machine.
No, there wouldn't be, since the gretl package for Windows
includes gnuplot 4.5.
Allin Cottrell