20040326: confusion over GOES resolution in GINI data

Tom-Tom-

Tom Whittaker wrote:
> Good to hear from you.  Thanks for looking at this and giving me more 
> details. As you might have guessed, people were using the IDV which 
> hides all the dataset and adde request details behind a GUI.

As does Tom's McGUI unless you expand the frame to change the resolution,
etc. ;-)  In fact, the resolution that is returned is similar to
what the defaults give you in McGUI.  The idea was that the defaults
would give you the entire image at a resolution that would fit in
the screen.  The defaults are stored in a properties file
ImageSelector.defaults which is in the ucar/unidata/ui/imagery package.
You could change this as needed and make a new local release if you
want.  However, that is not an efficient way to do things. ;-(

> In any event, I took your lead, and requested a 1km vis GINI _centered_ 
> at a given lat/long, and...whalha!  I got a nice 1km image.  I think the 
> problem was that people were not specifying a centerpoint, and the GUI 
> must default the resolution to something other than 1km.

If you don't center, the IDV selector uses defaults that were taken
from ADDEIMAGE.CORE.  In many cases this means that there is a
blow-down factor applied to big images.  If you have another strategy
for defaults for displaying imagery, please bring them up as part
of the IDV Steering Committee discussions.

> Don?  Is that the case?

If you specify a center point, it uses the native resolution of the
image and a 700x700 size (or image size if smaller) for
the image, not what's in ImageSelector.defaults.  I added this feature
because I wanted to see 1km VIS imagery over CO.

As Tom Y has pointed out before, the image selector in the
IDV is not adequate.  There are a few things I would like
to do:

1) change ImageSelector.defaults to be an IDV resource file like
other IDV files.  The current setup does not allow for easy
local configuration like other IDV resources.  This is all left
over from the original ImageViewer application I wrote back in
1999.

2) Allow the user to set up preferences for particular images.
I guess this is like what you have in your image selector which
would override the defaults.  We already have support for this
in the IDV resources (modelled after the CORE, SITE, USER stuff
in McIDAS), so it's just a matter of the UI that would support
this editing (beyond vi).

3) Enhance the selection GUI.  I would like to avoid a McIDAS'y
interface (like yours) because non-McIDAS users would not know
anything about line/element magnification.  Also, as you've also noted,
it doesn't handle multi-banded datasets.  I would really like to
have a selector that would allow the user to see the region that
the image covers and rubberband in on the area they want to
load.  At one point, the SSEC McIDAS GUI had something like this.
I haven't run it in a long time, so don't know if that's still
there or not.

I think a discussion of the image selection requirements by
the IDV Steering Committee would be a good thing.  Having multiple
choosers (i.e. a McIDAS'y chooser for SSEC folks, an a simple interface
for non-SSEC folks, and another for ....) would be feasible.
We'd just have to decide on the features for each one.  It's just
a matter of priorities and time.

Don
*************************************************************
Don Murray                               UCAR Unidata Program
address@hidden                        P.O. Box 3000
(303) 497-8628                              Boulder, CO 80307
http://www.unidata.ucar.edu/staff/donm
"There's someone in my head, but it's not me"    Roger Waters
*************************************************************