Re: NCL compared to Ferret

From: Sylvia Murphy (murphys AT XXXXXX)
Date: Tue Aug 28 2001 - 08:05:31 MDT

  • Next message: Karsten Bolding: "Curvilinear coordinates and ncl."

    s,

    I don't know all the detail of Ferret because I have not used it since grad
    school, but I do know that in many ways it is tailored to oceanographic data.
    That is what my degree was in. Ferret does produce nice graphics, but I would
    like to point out some of the present advantages of NCL, particularly for
    atmospheric data:

    1) NCL is based upon a netCDF data model, so that when data is read into
    NCL it will carry with it all the attributes and coordinate variables of that
    data (if they exist). This creates a robust data model.

    2) The interface to read in netCDF, GRIB, HDF data is the same and takes just
    two lines of code. IO is usually the beast of any computer language and NCL
    has beat them all.

    3) NCL also comes with a huge data processing capability that Ferret can not
    beat. There are statistical functions, special atmospheric calculations,
    regridding functions etc. You can peruse the list at:
    http://ngwww.ucar.edu/ngdoc/ng/ref/ncl/NclFuncsAndProcs.html

    4) The Graphics for NCL is ever evolving. There are several full time people
    who develope graphics capabilities and examples. They are also there to provide
    technical assistance and support. This is something you will not get from the
    Ferret developers. For instance, just yesterday a user came by my office
    because they needed to tweak a time series plot that they had. Within a day
    that capability was added to the graphics package. I maintain an on line
    graphical reference tutorial and library that is always growing. You can see
    some of the topics at:
    http://www.cgd.ucar.edu/csm/support/CSM_Graphics/
     
    5) WE also host NCL workshops periodically to get people spun up on NCL. YOu
    can access information about those at:
     http://www.cgd.ucar.edu/csm/support/workshop.shtml
     
    6) On the downside, if you only have acsii or binary data, and you want to
    take advantage of NCL's data model, you will have to manually fill in the
    attributes and coordinate variables. This is not necessary, but we find it
    particularly interesting.

    7) if you have some of your data, I would be happy to create a couple of demo
    plots to see if NCL will work for you. Perhaps the biggest advantage is our
    technical support team.

    cheers,

    sylvia

    ******************************************************
    Sylvia Murphy EML: murphys@ucar.edu
    NCAR CGD/CAS PHN: 303-497-1720
    1850 Table Mesa Dr FAX: 303-497-1333
    Boulder, CO 80305
                                    
    WEB: http://www.cgd.ucar.edu/csm/support/
         http://www.cgd.ucar.edu/csm/support/CSM_Graphics/
              
    ******************************************************



    This archive was generated by hypermail 2b29 : Tue Aug 28 2001 - 08:38:38 MDT