Re: fill vs shade

[Ansley Manke <ansley.b.manke@xxxxxxxx>]

Hi Julien,
This question just came back to me when I was doing something else.  
There are
cases where we just cannot do a fill plot (or contour, which uses the 
same algorighm).
Contouring employs curve-fiitting algorithms to make smooth curves 
between grid
locations.  See in the Users Guide Chapter 6, Sec 8.2 PPLUS Contour commands
for a discussion of the related settings. (Navigate in the index to 
"CONTOUR, pplus
controls".)

Because of the contouring process, there may not be enough data in some 
regions
to do a Fill plot. As the contour algorithm moves through the grid, it 
does not always
have enough information to compute the data value.   So, I wouldn't 
necessarily call
this a bug, but more of a limitation.  For instance, here is a somewhat 
silly example:

  yes? use coads_climatology
  yes? fill/hlim=300:301/vlim=0:1/L=1 sst        ! blank plot
  yes? shade/hlim=300:301/vlim=0:1/L=1 sst     ! fill plot, one color.

I've talked to others who have done the same as you - make a SHADE plot, and
overlay it with the FILL plot you really want.

Ansley


Julien Demaria wrote:

> Hi,
>
> I don't know why, but sometimes the fill command let some blank hole 
> on top-left and bottom-left of my window, and if I test the same 
> parameters with the shade command, there aren't any hole.
> It is as if the 2 top-left and bottom-left last filled polygons aren't 
> well computed and so I see some blank regions...
> For the moment I use a bad hack : I shade my data, and then fill with 
> /overlay...
> Is it a known bug ?
>
> I use ferret 5.51 on Win2000 Pro.
>
> Thanks in advance for help,
>
> Julien