Proposed NetCDF Conventions

John Sheldon (jps@GFDL.GOV)
Thu, 3 Jul 1997 19:51:49 -0400 (EDT)


Ron Stouffer passed your mail on to me regarding your proposed netCDF
conventions (even before it appeared in the netcdfgroup), as I am the
netCDF sponsor here at GFDL.  That doesn't necessarily make me an
expert, but I have been pretty immersed in netCDF for the past couple
of years and conventions have been an issue for far longer than that.

I must first back up and apologize for taking so long to respond to
your extraordinarily thorough effort, which deserves an equally thorough
evaluation by the rest of us in the netCDF community.  I am just
finishing up my N-th pass through it (the information density is quite
high!), and I will send you some more detailed comments next week.  In
the meantime, I at least owe you some overall impressions...

 1. Your ideas about "bounds_", "subcell", and "contraction" offer
     some very valuable ideas about how to handle averages, a subject
     we have been wrestling with for some time.  I still have some
     questions about just how to apply these, but this could well be
     the most valuable addition (conceptually) that we will extract
     from your proposal.

 2. A "calendar" attribute is something we adopted early on and is,
     obviously, a crucial piece of metadata (though there is still some
     debate about exactly how to use it, "unitime" notwithstanding).

 3. I was surprised at the absence of any CDL examples in your
     document.  A few CDL snippets (say, 20-30!) would have helped

 4. In general, I am opposed to the use of external tables.  While they
     can be handy for centers which exchange much the same data all the 
     time, it is problematic for researchers who can sometimes get
     quite "imaginative" and end up with things that aren't in the
     "official" table.  Complicating things is the fact that there
     often tends to be more than one "official" table.

 5. I heartily agree with your distinction between "missing" and
     "invalid" values. This is a definite increase in information
     content that has not been exploited to date.

I will reserve any other comments until I finish this one last pass
through the proposal.  I join the others who thank you for putting out
what was, obviously, a lot of work.  I am somewhat surprised at the
lack of responses to your proposal in the netcdfgroup, though I suppose
some people have responded outside the group (or maybe they are still
working their way through it, too!).  I hope your ideas eventually make
it into some kind of updated set of conventions.

While I share some of Brian Doty's concern for the proliferation of
conventions, I am fairly confident that a workable methodology can be
worked out in which a discipline-independent set of core conventions
could co-exist with various extensions agreed upon by individual
groups.  NUWG?


John P. Sheldon
Geophysical Fluid Dynamics Laboratory/NOAA 
Princeton University/Forrestal Campus/Rte. 1 
P.O. Box 308 
Princeton, NJ, USA  08542

(609) 987-5053 office
(609) 987-5063 fax
  "Calm down. It's only ones and zeros."