Hello,
I noticed that difference in behaviour between ShadeLtContour and
ShadeLtGtContour of shea_util.ncl.
CnLt is not included in
ShadeLtGtContour(plot,CnLt,FillPatLt,CnGt,FillPatGt)
but is included in
ShadeLtContour(plot,CnLt,FillPatLt)
when FillPatLt is applied.
The source code is consistent with the results.
ShadeLtContour: -- note .le. in the condition
do i=0,N-1 ; Fill contour levels
depending on
if(cnlvls(i).le.CnLt) then ; different criteria.
patterns(i) = FillPatLt ; see above URL
end if
end do
ShadeLtGtContour: -- note .lt. in the condition
do i=0,N-1 ; Fill contour levels
depending on
if(cnlvls(i).lt.CnLt) then ; different criteria.
patterns(i) = FillPatLt ; see above URL
end if
It appears that ShadeLtContour is an exception.
".lt." would be consistent with the naming
but for me it would be easier if I could just shade the contour levels
equal to or smaller (greater) than CnLt (CnGt).
Confusingly CnGt in ShadeLtGtContour is somehow included in the shade.
If I need to shade x<=-6 and x>=6 (with a level spacing of 2), I have
to write
ShadeLtGtContour(plot, -5., 6, 6., 17); shades x<=-6 and x>=6
but not
ShadeLtGtContour(plot, -6., 6, 6., 17); shades x<=-4 and x>=6
or
ShadeLtGtContour(plot, -5., 6, 5., 17); shades x<=-6 and x>=4.
BTW, is it correct to compare float values with .le., .eq., .ne., and
.ge. in ncl?
Shouldn't I compare abs(x-y) with a very small value?
I would appreciate if I am doing something funny.
Happy holidays!
-----
Dr. Takeshi Enomoto
eno@jamstec.go.jp
Frontier Research System for Global Change
_______________________________________________
ncl-talk mailing list
ncl-talk@ucar.edu
http://mailman.ucar.edu/mailman/listinfo/ncl-talk
This archive was generated by hypermail 2b29 : Mon Dec 22 2003 - 07:42:09 MST