some comments

["Sang-ki Lee" <Sang-ki.Lee@xxxxxxxx>]

Steve and Ansley,

Great. Both suggestions worked for me. One mistery solved!

I have been trying so many different graphic packages so far,
including IDL, SPLUS(R), Surfer (Grapher) and NCL(NCAR).
Although Mathlab is the most popular software in where I work
(We have only two ferret users here), the ferret is the best
software for me. In fact, I cannot live a day with it.
However, I have two major complains:

(1) The following is the typical viewport statement that I use.
As you can see, ylim for the first viewport is ylim=0.651, 0.975,
and ylim=0.528, 0.852 for the second one. I came up with these
values by trial and error to get rid of space between the two
plots. As you can see, the magic number is (0.201 = 0.852 - 0.651).
But, if I change the text size, this magic number changes.....
To make a long story short, it would be very nice if I can use
something like ylim=0.500, 1.000, and ylim=0.000, 0.500 instead.

define viewport/xlim=0.0000,0.6025/ylim=0.651, 0.975/text=0.6 v01
define viewport/xlim=0.0000,0.6025/ylim=0.528, 0.852/text=0.6 v02
define viewport/xlim=0.0000,0.6025/ylim=0.405, 0.729/text=0.6 v03
define viewport/xlim=0.0000,0.6025/ylim=0.282, 0.606/text=0.6 v04
define viewport/xlim=0.0000,0.6025/ylim=0.159, 0.483/text=0.6 v05
define viewport/xlim=0.0000,0.6025/ylim=0.036, 0.360/text=0.6 v06
define viewport/xlim=0.3976,1.0000/ylim=0.651, 0.975/text=0.6 v07
define viewport/xlim=0.3976,1.0000/ylim=0.528, 0.852/text=0.6 v08
define viewport/xlim=0.3976,1.0000/ylim=0.405, 0.729/text=0.6 v09
define viewport/xlim=0.3976,1.0000/ylim=0.282, 0.606/text=0.6 v10
define viewport/xlim=0.3976,1.0000/ylim=0.159, 0.483/text=0.6 v11
define viewport/xlim=0.3976,1.0000/ylim=0.036, 0.360/text=0.6 v12

(2) I use the following combination of the commands to get
my legend for xyplot. What these do is to plot legend lines,
then put symbols right at the middle of each legend line. I do
spend too much time in order to put the symbols at the right
locations. It would be very nice if you invent a new legline
like script that does all these things automatically.

go legline_nu  0.00 0.05 0.084  1 0.1 "Q_C_S_R"
go legline_nu  0.20 0.05 0.084  2 0.1 "Q_C_R_F"
go legline_nu  0.40 0.05 0.084  3 0.1 "Q_L_A_T_(_T_)"
go legline_nu  0.60 0.05 0.084  4 0.1 "Q_L_A_T_(_W_)"
label/nouser   0.45, 0.13, 0,0,0.08,"@pm01"
label/nouser   1.57, 0.13, 0,0,0.08,"@pm03"
label/nouser   2.70, 0.13, 0,0,0.08,"@pm17"
label/nouser   3.80, 0.13, 0,0,0.08,"@pm19"
label/nouser   3.0,  -0.5, 0,0,0.10,"month"

Thanks a lot!

Sang-ki

Steve Hankin wrote:
> Hi Sang-ki,
>
> The FILL command is an interpolation operation (it has to find a contour 
> path between grid cells).  The SHADE command is not (it simply treats 
> each cell independently).  Normally Ferret makes a small adjustment in 
> the extent of  the plotting region to accomodate the 1/2 grid cell 
> variation that exists between these two methods of rendering the graphics.
>
> To see this difference look at the X axis limits used by the PLOT+ 
> engine ("ppl list xaxis") in this output:
>
>     yes? use coads_climatology
>     yes? shade/x=130e:80w/y=20s:20n/l=1 sst
>     yes? ppl list xaxis
>              XAXIS QUALITIES
>              *LO= 1.30E+02 HI= 2.80E+02* TIC= 2.00E+01 LEN= 8.00E+00
>     yes? fill/x=130e:80w/y=20s:20n/l=1 sst
>     yes? ppl list xaxis
>              XAXIS QUALITIES
>              *LO= 1.31E+02 HI= 2.79E+02* TIC= 2.00E+01 LEN= 8.00E+00
>
>  >From what you've described I presume that you are using /OVERLAY or 
> /HLIMITS and thereby (accidentally?) taking control of the extent of the 
> plotting region.  The following command illustrates how you can create 
> white space at the edges this way:
>
>     yes? fill/x=130e:80w/y=20s:20n/l=1/hlimits=130e:80w sst
>
> Solution: As long as you set the data region ("x=130e:80w" above) to be 
> at least a grid cell larger on each side than the plotting region 
> ("hlimits=130e:80w" above) you can avoid this problem.
>
>     - steve
>
> ============================
>
> Ansley Manke wrote:
>
> > Hi Sang-ki
> > Yes, we've seen this.  Isyour data on an irregularly spaced
> > grid? Depending on the data and the grid it's on, the algorithm
> > that does the fill plot can't always fill the entire region.  A
> > workaround is to do a SHADE plot and then overlay it with
> > your FILL command, using the same contour levels.
> >
> > Ansley
> >
> > Sang-ki Lee wrote:
> >
> > > Hi. My problem is that when "fill" command is used for colored contour,
> > > sometimes I get white space at top or side edge of the colored contour
> > > plot. This problem come and go, and it is nothing to do with missing
> > > data point. Has anyone noticed this problem?
> > >
> > > I am using v5.60 on Linux.
> > >
> > > Thank you.
> > >
> > > Sang-ki
> --
> º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤ø,¸¸,ø¤º°`°º¤º¤ø,¸¸,ø¤º
>
> Steve Hankin, NOAA/PMEL -- Steven.C.Hankin@noaa.gov
> 7600 Sand Point Way NE, Seattle, WA 98115-6349
> ph. (206) 526-6080, FAX (206) 526-6744
>  

--
----------------------------------------------------------------------
Dr. Sang-ki Lee
RSMAS/CIMAS
University of Miami
4600 Rickenbacker Causeway
Miami, FL 33149 USA
Tel) 305-361-4521
Fax) 305-361-4412
E-mail) sang-ki.lee@noaa.gov
----------------------------------------------------------------------
Mailing-address:
Dr. Sang-ki Lee
AOML/PHOD
4301 Rickenbacker Causeway
Miami, FL, 33149 USA
----------------------------------------------------------------------