Re: limits of netCDF library

• To: langdon@xxxxxxxxxxxx (B. Langdon)
• Subject: Re: limits of netCDF library
• From: Russ Rew <russ@xxxxxxxxxxxxxxxx>
• Date: Thu, 02 May 1996 23:13:29 -0600

Bruce,

> It would be helpful to fans of netCDF in other branches of physics if netCDF
> were to include complex*4 and complex*8 as a data type.  Complex variables
> have been very useful for a couple of centuries in applied math..

The complex data type was not originally included because C has no complex
data type, and netCDF was intended to permit access of array-oriented data
from either C or Fortran.  If complex data types were added, accessing the
data from C would require use of a non-primitive complex type.  

Also, adding a complex data type would cause problems for generic netCDF
programs that assume any numeric value from a netCDF file can be stored in a
double for arithmetic operations and comparisons.  Operations requiring
ordered values, such as finding the maximum value in an array, would not
be meaningful for complex values.

Currently an array of complex values can be represented as an array of
floating-point values with an extra dimension of length 2 for the real and
imaginary components.  A convention for the name of a dimension that would
only be used to represent real and imaginary components of complex values
might suffice, since only one such dimension would be needed.  Another
conventional dimension could be used for polar representations ...

Besides, how would complex values be used for the really important
problems, like representing time :-) ?
______________________________________________________________________________

Russ Rew                                           UCAR Unidata Program
russ@xxxxxxxxxxxxxxxx                              http://www.unidata.ucar.edu