[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

20030811: IDV 2D Display only on Linux?

>From: "Todd Plessel" <address@hidden>
>Organization: EPA
>Keywords: 200308072228.h77MStLd012793

Hi Todd-

Sorry for the delay in responding.

re: 2D under Linux
>I assumed so - wouldn't it fail to JIT-compile/run otherwise? E.g.,
>./ucar/visad/display/AxisScalarMap.java:import javax.media.j3d.Transform3D;
>I'll investigate.

Are you using the same java path for running as building?  I've
made that mistake before, so just have to ask. ;-)

>> > Likewise, when a 1-layer grid is read in there are no Fields (e.g., 2D)
>> > listed so no displays can be made.
>>
>> Hmm, that is odd.  For the file I tried, it listed the 2D fields.
>
>Single-layer (Vis5D) grids work on IRIX.

But these are not listed under Linux for a Vis5D file?  Do you have
a sample Vis5D file that shows the problem?

>> Could you send a sample file?
>
>Attached is a zip file containing:
>a one-layer, three-time-step subset of MET_CROSS_2 (an M3IO VG Grid file)
>the current version of my M3IOVGGridConvention.java and
>modified Convention.java that uses my M3IOVGGridConvention.java.

It actually does show up in the display, if you zoom way in over
Illinois. ;-)  The problem is one of converting from km -> meters.
John Caron is out this week, so I'm not sure exactly what is going
on and will ask when he gets back.  In the mean time, I got it to
work by changing:

    CoordinateAxis axis =
      new CoordinateAxis1D( ds, "y", "double", dim.getName(), unit );

to: 

    CoordinateAxis axis =
      new CoordinateAxis1D( ds, "y", "double", dim.getName(), "km" );

in M3IOVGConvention.java in makeXCoordinateAxis and makeYCoordinateAxis.
Not sure if it will break other things.

One other note about the netCDf file.  The units for QV
(QV:units = "water/dry air(g)") is not a valid unit
specification.  If it's g/g, then that's what should be listed.
The "water/dry air" should be listed in another attribute.
It doesn't affect the display, but you can't change to other
units, or have the units propogate through formulas.  Since
we can't parse "water/dry air(g)", the units end up as null.

re: point data
>I've replaced our version with the attached one and I'll try it.

Let me know if that doesn't work for you.

>> For the profiler data, you would need to write a new adapter.
>
>Not just a new Convention class (since it's a NetCDF file)?

no, right now the Convention stuff really only works for grids.
The eventual goal is to be able to handle other types of
netCDF data (including point data), but we have no data model
for those yet.  The GeoGrid structure works pretty well for
gridded data, but not all (as evidenced by your stacked blocks
example).

Don
****************************************************************************
Unidata User Support                                    UCAR Unidata Program
303 497 8643                                                  P.O. Box 3000
address@hidden                                   Boulder, CO 80307
----------------------------------------------------------------------------
Unidata WWW Service              http://my.unidata.ucar.edu/content/support
****************************************************************************