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

Re: netCDF 3.0?

Hi Brian,

> In a recent post about the NC_UNLIM dimension and sub_records, you
> mentioned that netCDF 3 is planned.  You mentioned packed data types, what
> does this mean?

It means you will be able to store arrays of n-bit data compactly, and you
will be able to request the type of data you want without knowing how it is
stored.  For example, you will be able to request data from a variable as an
array of floating-point numbers without knowing or caring that the data was
stored as ints, longs, doubles, or 21-bit packed and scaled integers.

Here's some more info, but these are planning documents that are still in
draft form (we've already changed the naming convention for the new
functions from what's in these documents), so please don't start programming
to these interfaces yet! (-:

  http://www.unidata.ucar.edu/staff/hld/iita/packing.html
  http://www.unidata.ucar.edu/packages/netcdf/iita/summary-1996-04-30.html
  http://www.unidata.ucar.edu/packages/netcdf/iita/plans-index.html  

The last URL is a presentation that reveals some early plans for netCDF 3,
4, and 4+, but we're getting ready to refine these plans and get some
feedback on a second iteration at a meeting with some interested users on
May 17.  After that, things may be ready to post some of these plans to
the netcdfgroup mailing list to get other users reactions, but we had to
first present this stuff to a group that is helping fund the development.

--Russ